5 5 32 4

Celerra
Network Server
Version 5.5.32.4
Release Notes
P/N 300-002-692
Rev A22
January 14, 2007
The EMC Celerra Network Server is a network-attached storage (NAS)
server. It enables clients on a network to store and retrieve files over the
network. These release notes contain information about version 5.5 of
Celerra Network Server.
New features and changes ...................................................................... 2
Known problems and limitations ........................................................ 48
Problem prevention................................................................................ 82
Technical notes ........................................................................................ 88
Environment and system requirements ............................................ 107
Documentation ..................................................................................... 109
Installation ............................................................................................. 124
Bug fixes................................................................................................. 167
Version 5.5.32.4 ............................................................................. 168
Version 5.5.31.6 ............................................................................. 202
Version 5.5.30.4 ............................................................................. 227
Version 5.5.29.1 ............................................................................. 248
Version 5.5.28.1 ............................................................................. 268
Version 5.5.27.5 ............................................................................. 283
Version 5.5.26.4 ............................................................................. 323
Version 5.5.25.2 ............................................................................. 335
Version 5.5.24.2 ............................................................................. 356
Version 5.5.23.2 ............................................................................. 380
Version 5.5.22.2 ............................................................................. 394
Version 5.5.21.4 ............................................................................. 417
Version 5.5.20.1 ............................................................................. 440
Troubleshooting and getting help ...................................................... 451
Please check Powerlink for updates to this document. Updates are made
when a new GA version comes out, when a patch is released, or when
new information is discovered.
New features and changes

The following new features are supported as of Celerra Network
Server version 5.5.32.4.
2008 Daylight
Saving Time change
for Australia
Starting in 2008, the Daylight Saving Time (DST) dates have changed
for the Australia Eastern (New South Wales, Victoria, Australia
Capital Territory, and Tasmania), Central (South Australia), and Lord
Howe Island time zones. DST will end on the first Sunday in October
and begin on the first Sunday in April for all states.
Version 5.5.32.4 updates Control Station components to support the
new DST dates for Australia. To update Data Movers for the
Australia DST changes, login to the Control Station as nasadmin and
follow the instructions below.
1. To check the current time zone of the Data Movers, use the
following syntax:
server_date ALL timezone
If the local time zone is set to CST (Adelaide, Broken Hill), EST
(Currie, Hobart, Melbourne, Sydney), or LHST (Lord Howe
Island), continue with step 2. If the time zone is GMT (the
default) or any other time zone outside of eastern and central
Australia, you do not need to do anything to the Data Movers for
this DST change.
2. To set the time zone or change the DST transition dates, use the
following syntax:
server_date <movername> timezone <timezone_string>
Where <movername> is the Data Mover name and

<timezone_string> is one of the POSIX time zone settings
shown below for the Australia Central, Eastern, and Lord Howe
time zones:
Adelaide, Broken Hill: CST-9:30CST,M10.1.0/2,M4.1.0/3
Currie, Hobart, Melbourne, Sydney:
EST-10EST,M10.1.0/2,M4.1.0/3
Lord Howe Island: LHST-10:30LHST-11,M10.1.0/2,M4.1.0/2.
CSA version check
CSA version check prevents running the current CSA version on an

incompatible NAS version. CSA 5.5.32 is compatible with NAS
5.5.31.6 and NAS 5.5.32.4.
Celerra Network Server Release Notes
Custom Connect
EMC Modem
Settings
The procedures to disable dial tone detection and set the dial method
have changed in the ConnectHome feature as of this release. Please
familiarize yourself with the version-specific procedures to make
sure any custom settings are preserved during an upgrade.
Modify/Create the /nas/opt/connectemc/modem.cfg file:
To disable dial tone detection, add the following line to the
modem.cfg file:
set dial ignore-dialtone on
To use tone dialing, add the following line to the modem.cfg file::
set dial method tone
To use pulse dialing, add the following line to the modem.cfg file:
set dial method pulse
Each setting should reside on its own line. For example:

set dial method pulse
set dial ignore-dialtone on
Table 1
RAID 6 support on
CLARiiON
Impact of modem settings during upgrades
Upgrading from Upgrading to
Custom modem
settings?
5.5.30.5 and
earlier
5.5.31.6
Are NOT
preserved.
Settings are not preserved,

therefore, later upgrades to
5.5.32.x cannot resurrect the
settings.
5.5.30.5 and
earlier
5.5.32.4 or
later
Are preserved.
The settings must be in the

/nas/sys/callhome.modem file.
5.5.31.6
5.5.32.4 or
later
Are preserved

/nas/opt/connectemc/login.cmds
file.
5.5.32.4
Future
releases
Are preserved

/nas/opt/connectemc/modem.cfg
file
Notes
Celerra Network Server introduced support for binding disks as

RAID 6 on CLARiiON systems running FLARE release 26 or later.
There are 4 additional system-defined storage pools added for access
to RAID 6 storage. These are clar_r6, clarata_r6, cm_r6, and
cmata_r6. Celerra supports 4+2, 6+2, and 12+2 RAID 6
configurations.
Requirements
FLARE R26 or later is required for CLARiiON to support RAID 6.
RAID 6 is only supported on CLARiiON with backend security
enabled and properly setup on the Celerra (see EMC Knowledge base
emc168569 for more information). EMC Solutions Enabler version
V6.4.2 is required to communicate with FLARE R26 from the Celerra.
Since CLARiiON only support N-1 backward compatibility if you
are running Flare R22 you will need to upgrade to R24 or R26. If you
don't upgrade you will get a pre-upgrade health check error if you
are running a release earlier than Flare R24 and you should upgrade
Flare prior to upgrading NAS.
CX200/400/600 are end of life at Flare R19 and remain supported
with this version.
Version 5.5.31.6
features
ConnectHome
This section describes the features introduced in version 5.5.31.6.

The Celerra CallHome process was replaced with the Celerra
ConnectHome feature.
Celerra ConnectHome automatically notifies the EMC service center,
or your authorized service provider, if the Celerra system detects a
serious problem. ConnectHome sends messages using Email, FTP for
ESRS Gateway, or the Celerra modem and a customer-supplied
analog telephone line.
Celerra Manager
Use the Celerra Manager interface to configure the ConnectHome
feature. The Celerra Manager on-line help also contains information
about ConnectHome.
Email transport
EMC highly recommends using the Email transport as the primary
transport and keeping the default EMC Recipient address,
emailalert@emc.com, unless otherwise specified by your authorized
service provider.
Configure the Email transports sender address, using the
nas_connecthome command, with the site administrators email
address to catch email relay errors.
There is no need to manually edit a file. Its all done through the
nas_connecthome modify serial_number
<manual_serial_number> command.
If the serial number is automatically detected, the following message

is returned:
[root@csdev-16 ~]# /nas/sbin/nas_connecthome -modify
-serial_number 'BOGUSSERIAL'
Serial number automatically detected
Ok
[root@csdev-16 ~]# /nas/sbin/nas_connecthome -info | grep
Serial
Serial number = APM00072901930
If the serial number is not detected, it gets updated accordingly and

displays a message similar to:
[root@silver ~]# /nas/sbin/nas_connecthome -modify
-serial_number "MYSERIAL"
Ok
[root@silver ~]# /nas/sbin/nas_connecthome -info | grep
Serial
Serial number = MYSERIAL
Complete details about the nas_connecthome command can be

found in the Command Reference Manual which is available on the
Documentation CD and on Powerlink.
DMX-4 with ATA
This version of Celerra Network Server introduced support for the

new Symmetrix hardware offering, the DMX-4.
DMX-4 extends the Symmetrix tiered storage functionality with the
addition of the lower cost SATA drives. The SATA drives will be
added in a Stiletto enclosure.
Disk types and storage pools
Celerra has added the following 6 new disk types to support SATA
drives from Symmetrix storage: ATA, R1ATA, R2ATA, BCVA, R1BCA
and R2BCA. Also, 6 new system-defined storage pools were
introduced. These are symm_ata, symm_ata_bcv, symm_ata_rdf_src,
symm_ata_rdf_tgt, symm_ata_bcv_src, and symm_ata_bcv_tgt. At
this time, there is no differentiation based on RAID type of devices.
TimeFinder
TimeFinder snapshots will by default use the same storage type (FC
or ATA) as the production file system, although this can be changed
by using the disktype option.
Requirements
Enginuity version 5772 is required to support DMX4 platform. EMC
Solutions Enabler version V6.4.2 is required to query for ATA
information on Celerra.
The EMC Symmetrix DMX-4 and Enginuity 5772 Overview - Applied
Technology white paper contains more detailed information and can
be found on Powerlink.
Proxy ARP
Version 5.5.30.5
features
2007 Daylight Saving
Time change for New
Zealand
The network topology used to access the public IP addresses of the

storage processors in the NS20 and NS40 systems has transitioned
from NAT to Proxy ARP. With Proxy ARP enabled, the backend
array storage processors have their IP addresses on the same subnet
as the Control Station's external network interface, while still wired to
the Control Station's private network. The Control Station acts as a
network bridge between the public network and the backend array
SPs, allowing the clients and hosts see the backend array SPs as
though they were connected directly to the public network.
for New Zealand. It is now observed from the last Sunday in
September to the first Sunday in April, adding three weeks to the
total DST period. Information about the 2007 Daylight Saving Time
change for New Zealand is detailed in the EMC Celerra Network
Server New Zealand 2007 Daylight Saving Time Change technical
note, available on Powerlink.
Versions 5.5.29.1 and later update the Control Station components to
support the new DST dates for New Zealand. To update Data
Movers for the New Zealand DST changes, login to the Control
Station as nasadmin and follow the instructions below.
1. To check the current time zone of the Data Movers, use the
following syntax: server_date ALL timezone
If the local time zone is set to NZ (Auckland) or CHA (Chatham),
continue with step 2. If the time zone is GMT (the default) or any
other time zone outside of New Zealand, you do not need to do
anything to the Data Movers for this DST change.
2. To set the time zone or change the DST transition dates, use the
following syntax:
server_date <movername> timezone <timezone_string>
Where <movername> is the Data Mover name and
<timezone_string> is one of the POSIX time zone settings shown
below for the New Zealand time zones:
Pacific/Auckland: NZST-12NZDT,M9.5.0/2,M4.1.0/3
Pacific/Chatham:
CHAST-12:45CHADT,M9.5.0/2:45,M4.1.0/3:45
Celerra Startup
Assistant
The Celerra Startup Assistant (CSA) tool simplifies the installation of

Celerra NS20/NS20FC/NS40/ NS40FC systems. The CSA is a
stand-alone application that runs on a client computer on the same
network as the Celerra system.
The CSA is used to:
Verify cabling and connectivity are working properly prior to

software configuration
Configure Control Station and blades
Input site-specific information and configure call-home
Create NFS Export, CIF shares, or iSCSI LUNs
Register for services
Refer to the licensing terms section on page on page 127 for copyright
and licensing information pertaining to the open source modules
used by the EMC Registration Wizard and CSA included on the Apps
and Tools CD for Celerra release 5.5.30.4.
iSCSI
replication-based LUN
cloning
The server_iscsi command supports the following new option:

-lun -modify <lun_number> -target <target_alias_name>
-readonly {yes [-Force]|no}
This option changes the specified LUN to read-only (-readonly yes)

or read/write (-readonly no).
This option enables iSCSI LUN replication to be used to create a copy
of an iSCSI LUN. You can change a destination LUN from read-only
to read/write, thereby creating a replication-based clone of the source
production LUN. Like the original LUN, the clone LUN supports
snapshots, replication, and LUN extension.
This option also enables a read/write LUN to be changed to
read-only. This feature can be used with reverse replication to copy
changes from a replication-based clone LUN back to the original
source LUN (which is made a read-only destination).
The following requirements and restrictions apply:
You cannot use the -modify option with a promoted snapshot.
Before changing a LUN to read/write or read-only, you must

delete any replication session that specifies the LUN as a source
or destination.
Changing a LUN to read-only while it is being accessed by a host

can interrupt data transfer operations that are pending or in
progress, possibly causing data corruption.
If the LUNs mask grants access to an iSCSI initiator and that

initiator is logged in to the target that contains the LUN, the
-readonly yes option fails unless the -Force option is included.
Before using the -Force option, flush any cached or unwritten
data to the LUN by unmounting and then remounting the file
system that contains the LUN.
Examples
Do not create a replication-based clone of a LUN that is in an

inconsistent state (for example, a destination LUN after the
baseline copy operation is aborted).
To use a clone LUN with Windows Server 2003, you must also
remove the VSS read-only and hidden bits from the LUN using
the diskpart utility or the script RmMountVolume.bat provided
with Replication Manager.
To change LUN 24 (a read-only LUN) on target t1 on server_2 into a

read/write LUN, type:
$ server_iscsi server_2 -lun -modify 24 -target t1
-readonly no
To change LUN 11 (a read/write LUN) on target t1 on server_2 into

a read-only LUN (suitable as a destination for replication), type:
$ server_iscsi server_2 -lun -modify 11 -target t1
-readonly yes
NS20, NS20FC and

NS40FC
The Celerra NS series includes the NS20, the NS20FC, and the
NS40FC.
Integrated NS20
The Integrated NS20 consists of a complete Celerra Network Server
and a CLARiiON CX3 storage system in a single, compact system.
The NS20 supports a single Control Station and is available with
single or dual blades.
Integrated NS20FC
The Integrated NS20FC consists of a complete Celerra Network
Server and a CLARiiON CX3 storage system in a single, compact
system.
The NS40 supports a single Control Station and is available with
single or dual blades. There are 8 FC ports to permit additional
Windows and/or UNIX hosts to share storage in the captive
storage array.
Integrated NS40FC
The Integrated NS40FC consists of a complete Celerra Network
Server and a CLARiiON CX3 storage system in a single, compact
cabinet (referred to as an integrated array).
The NS40FC supports a single Control Station and is available with
single or dual blades. The Fibre Channel (FC) option allows an
additional host to be connected to the storage array.
NS20FC and NS40FC installation
NS20 and NS40 systems are installed using NAT for communication
between the storage processors and external applications, such as
Navisphere Service Taskbar. For NS20FC and NS40FC systems, after
installation and before adding other hosts, a script must be executed
to convert the NAT implementation to Proxy ARP. Refer to PRIMUS
Solution emc165977 for more information.
Version 5.5.29.1
features

Symmetrix RAID 6 support Version 5.5.29.1 supports EnginuityTM
release 5772 along with Symmetrix RAID 6. Symmetrix RAID 6 and
RAID 5 will appear very similar to most Celerra users. Although the
RAID levels have different performance characteristics, these
differences will often be masked due to the read-intensive nature of
most NAS workloads. Outside of performance issues, both RAID
types will be recognized as STD type volumes and will be treated
similarly by Celerra.
Recommendations
The existing best practices for optimizing Symmetrix RAID 5 with a
Celerra Network Server also apply to RAID 6. In addition, note the
following recommendations for RAID 6:
Balance LUNs across disk adapters/directors.
Control Volumes must be created on mirrored devices, not on

parity RAID, RAID 5, or RAID 6.
Do not use multiple disk volumes (dvols) built from the same
physical spindles in a single file system.
Do not build a file system on a single Symmetrix metavolume.
Do not mix multiple RAID types (mirrors/RAID 1, RAID 5, parity

RAID, RAID 6) within a single file system.
AVM is not RAID1, RAID5, RAID6, RAID-S aware on

Symmetrix.
The system-defined AVM profiles for Symmetrix will attempt

to stripe together 8 dvols. If you intend to use an AVM
system-defined pool for Symmetrix (such as symm_std),
create sufficient LUNs/dvols so that sets of 8 consecutive
same-sized dvols will all reside on different physical spindles
and are all of the same RAID type.
Consider using AVM user-defined pools along with Manual
Volume Management to create file systems for highly
performance-sensitive environments, especially if the file
system will ultimately reside on Symmetrix RAID 6, RAID 5,
or parity RAID hypervolumes. If you have a mix of RAID
protection types, segment them by creating appropriate
stripes of like type dvols and putting the stripes in separate
user defined pools. Use AVM system-defined pools for all file
systems ultimately residing on Symmetrix RAID 1
hypervolumes.
If you manually create stripe volumes on Symmetrix based disk

volumes, including those based on RAID 6, use the following
guidelines to determine the appropriate Celerra stripe element
size:
NFS and/or CIFS:
32 KB
MPFS:
256 KB
iSCSI:
xx 32 KB
RAID 6 performance is more sensitive to write intensive workloads

than RAID 5. For LUNs created on identical numbers of physical
spindles, write performance may be slower with RAID 6. Read
performance for RAID 6 should be essentially the same as for RAID 5.
Version 5.5.28.1
features
CX3-10C support

Version 5.5.28.1 introduced Celerra Gateway support of the CX3-10C
disk storage array. The CX3-10C facilitates low cost all ATA drive
configurations (including vault and control LUN drives) as well as
conventional all FC drive and mixed FC-ATA drive configurations.
CX3-10C features iSCSI ports for use by non-Celerra hosts in addition
to the FC ports used for Celerra Gateway blade connection. CX3-10C
is similar to CX3-20C except that the CX3-10C is limited to 60 FC
and/or ATA drives, CX3-10C can be configured with only ATA
drives, and each CX3-10C SP uses 1 GB of memory rather than 2 GB.
10
IP Alias with IP
Replication
Microsoft Windows
Vista
Celerra Network Server versions 5.5.28.1 and later support IP Alias

with IP Replication. All restrictions on Control Station failover also
apply to IP Alias with IP Replication configurations. The following
guidelines apply to this feature:
When using IP Replication for the first time, or on new systems,

configure IP Alias first, then use IP Alias in the -ip <ipaddr>
option of nas_cel -create command.
For existing systems with existing IP Replication sessions, the

current slot_0 IP address (primary Control Station IP address)
must be used. For example: nas_config -IPalias -create 0 .
Do you want to slot_0 IP address as your alias [yes or no] yes.
If the Control Station fails over while IP Replication is running,

the IP Replication command (fs_replicate) may need to be
re-executed manually. Please check logs (/nas/log/cmd_log*,
server_log. command output etc.) to determine how to proceed.
Keep the fs_replicate command output for resync, suspend,
restart, failover, and reverse options in case of failure, then
execute the rest of steps based on the instruction in command
output.
When IP Alias is deleted using the nas_config -IPalias

-delete command, the IP address of the primary or the
secondary Control Station is not changed. Changes to the IP
address of the primary or the secondary Control Station must be
done separately. IP Replication depends on communication
between the source Control Station and the destination Control
Station. When IP Alias is used for IP Replication, deleting the IP
Alias breaks the communication. The IP address which was used
as the IP Alias must be restored on the primary Control Station to
restore the communication.
While performing a NAS code upgrade using ssh or telnet, do not

use an IP Alias in the SSH or telnet to logon to the Control Station.
Releases 5.5.28.1 and later support Microsoft Vista client in Windows

2000 and Windows 2003 domains.
11
Version 5.5.27.5
features
Checking CIFS
components and their
dependencies
Version 5.5.27.5 introduced a new command, server_checkup,

which enables you to check the state of all CIFS components and their
dependencies. When this command is executed, it also checks the
validity of the configuration parameters and displays the items that
may cause unexpected behaviors. Additionally, the server_checkup
command can be used to perform an initial analysis of a problem, or
to verify a newly configured CIFS server.
This command doesnt fix errors detected, and does not modify any
parameters.
More information about how to use the server_checkup command

can be found in the Configuring CIFS on Celerra technical module, and
in the server_checkup man page.
Configurable login
banner
Celerra Manager was enhanced to allow you to create a configurable

banner message and message of the day (MOTD).
The login banner displays when a Celerra is accessed, but before the
login prompt. The banner is visible from each management interface
including Celerra Manager, secure shell sessions (SSH), and console.
The MOTD appears in a pop up in Celerra Manager after the user
logs in successfully. From SSH or console, the text displays in the
shell after successful login.
As root user, you can configure the banner and the message of the
day (MOTD) on the Celerras > Control Station Properties
tab. You can also configure the banner message from the CLI using
your favorite text editor to edit the /etc/issue or /etc/motd files.
12
Creating automated
checkpoint schedules
using the command
line interface
In addition to creating automated checkpoint schedules using the

Celerra Manager, you can now create and manage these schedules
using the nas_ckpt_schedule command. All automated checkpoint
schedules created using the Celerra Manager can be managed using
the CLI, and all checkpoint schedules created using the CLI can be
managed through the Celerra Manager.
Detection and
recovery of corrupted
XLT files
This feature provides a mechanism to detect XLT (translation) file

corruptions each time XLT files are updated and accessed, and
recover those corrupted files automatically from the Control Station.
If file corruptions are detected, a log event is generated and sent to
the Control Station, which records the event in the system log and
triggers a call-home event as well as an auto-recovery event.
Internet Explorer 7.0
Celerra Manager fully supports Internet Explorer 7.0 in version

5.5.27.5 and later.
iSCSI snapshots for

Linux hosts
The CBMCLI package provides commands that can be run on a Linux

host to create and manage snapshots of iSCSI LUNs on a Celerra
Network Server and to manage remote replication of those snapshots.
The CBMCLI package does not support hardware iSCSI initiators for
Linux.
NDMP Integrated
Checkpoint support
for Read-Only file
systems
As of Celerra Network Server version 5.5.27.5, NDMP backup with

integrated checkpoint supports a read-only file system as the backup
root. The only exception is if the file system itself is a checkpoint.
Network services: port

management
You can list the current state of some nonessential network services
(and associated communications ports and protocols) on the Control
Station and Data Movers. You can enable, disable, and monitor these
services. Some services require a reboot. You have the option of
rebooting immediately or delaying the reboot to a later time.
You can manage the following network services with Celerra
Manager (select Celerras > [Celerra_name] > Security > Network
Services):
FTP data transfer, FTP control
NDMP
NETBIOS Name, NETBIOS Datagram, NETBIOS Session, CIFS
PAX
RIP
rquotad
SNMP
statd, lockd
TFTP
You can find more detailed information about network services in the
Celerra Manager on-line help.
NFS hostnames
lookup
Celerra Network Server version 5.5.27.5 was enhanced to be case

insensitive for both hostnames lookup and net groups lookup in NFS
Exports.
Proactive Health
Check (PAHC)
The PAHC uses the nas_checkup command to run a series of

successive checks to allow Celerra Administrators and EMC Service
Providers to assess and monitor the health of the Celerra. Refer to the
nas_checkup man page or the Celerra Network Server Command
Reference Manual for a list of successive checks.
13
By default, a scheduled run is every two weeks. During these

scheduled runs, nas_checkup:
Does not display output on the screen
Writes an entry in the system log
Generates nas_checkup and nas_checkup callhome logs

If a scheduled run finds an issue, nas_checkup:
Displays an alert on the Celerra Manager
When necessary, sends nas_checkup callhome logs to EMC

Service Providers
Sends an email to the email notification list.

Secure NaviCLI
Secure NaviCLI provides a mechanism to gather and store

CLARiiON authentication information, and use it for subsequent
CLARiiON requests. You can use the CLI nas_storage command or
Celerra Manager (select Celerras > [Celerra_name] > Storage >
Systems > CLARiiON storage system.) to create or modify the Celerra
cached version of CLARiiON credentials (username/password).
CLARiiON credentials should be managed through Navisphere
(Navisphere Manager GUI or Navisphere CLI naviseccli command).
Secure NaviCLI and MirrorView/Synchronous
Secure NaviCLI has changed the format of the security file on
Celerra. If you previously configured MirrorView/S for disaster
recovery, rerun /nas/sbin/nas_mview -init after upgrading to
re-create the security file.
Note: MirrorView/S is currently only available through the RPQ (Request for
Price Quotation) process.
Session timeout
To enhance the Celerra Network Servers system security, Celerra

provides the means for setting a system-wide session timeout value
by introducing a -sessiontimeout option to the /nas/sbin/nas_config
command. The -sessiontimeout option allows you to configure
Control Station shell sessions to automatically time out after a
specified period of inactivity.
The Control Station supports three shells:
bash
ksh
tcsh
14
Each of these shells supports a session timeout feature. The

-sessiontimeout option sets the session timeout value across the
system, automatically updating the appropriate values in the
/etc/profile for the bash and ksh shells and the auto logout variable
in /etc/csh.cshrc for the tcsh shell.
Session timeout is enabled by default. The default value is 60
minutes. Possible values are 5 to 240 minutes.
You can change the session timeout value for individual users by
setting the relevant variable in the users shell configuration file (for
example, ~/.bashrc). Values are not restricted if you edit the
configuration file directly.
To change the session timeout value, type

/nas/sbin/nas_config -sessiontimeout <minutes>
To disable session timeout, type
/nas/sbin/nas_config -sessiontimeout 0 (to indicate zero minutes)
or
/nas/sbin/nas_config -sessiontimeout off
The Celerra Manager provides a similar session timeout feature. You

can configure a Celerra Manager session to automatically time out
after a specified period of inactivity. By default, a Celerra Manager
session will time out 480 minutes (8 hours) after a user logs in,
regardless of activity.
Configuring a session timeout value using nas_config
-sessiontimeout only sets this value for a Control Station shell
session. Session timeout for a Manager session must be set in the
Manager.
The nas_config man page and the Celerra Network Server Command
Reference guide provide an explanation of the command syntax and
usage.
XML API v2
The Celerra Network Server XML API v2 provides active and passive
management for the following areas:
iSCSI management -- The XML API v2 allows clients to query and

manage an iSCSI target available on Celerra by means of iSCSI.
Automatic file system extension management -- XML API v2

automatic file system extension allows clients to extend file
systems without user intervention and retrieve information
related to an automatically extended file system.
15
Version 5.5.26.4
features
CLARiiON backend
software (FLARE)
Celerra storage configuration and statistics management for

Symmetrix and CLARiiON -- The XML API v2 Celerra storage
configuration and statistics feature allows clients to obtain
configuration and statistics information from an attached
CLARiiON or Symmetrix backend storage system.

FLARE 24
Celerra Network Server versions 5.5.26.2 and later are compatible
with release 24 of the CLARiiON Backend software (FLARE).
LCC cable issue
With release 24 of CLARiiON FLARE software, a missing or damaged

LCC cable might cause the CLARiiON storage array to report the
disks as removed or empty. This will result in hardware errors and
disk totals to be less than expected, but the I/O to the disks will not
be interrupted. Be sure to check the CLARiiON storage array cables if
the disks are reported as removed or missing and the disks are still
functioning properly.
CX3-20F and CX3-40F
support
CLARiiON added an I/O card to the CX3-20 and CX3-40 arrays. This
I/O card expands the number of Fibre Channel ports on the CX3-20
Storage Processor to seven ports and the CX3-40 Storage Processor to
eight ports. These arrays ship with this new I/O card by default and
are differentiated from the previous shipping arrays by new model
numbers, CX3-20F and CX3-40F. As of version 5.5.26.2, the CX3-20F
and CX3-40F are supported arrays for connection to Celerra Gateway
systems.
RAID 5 support
Version 5.5.26.4 and later support both 4+1 RAID-5 and 6+1 RAID-5
on CX3 SATA-II/Low Cost Fibre Channel (LCFC) drives via both
manual volume management and AVM. AVM support for both
RAID-5 configurations has been added to the clarata_archive AVM
pool and new templates are provided to bind various shelf
configurations of 4+1 RAID-5 and 6+1 RAID-5. In addition, the
clarata_archive storage pool stripe size has been increased from 8KB
to 64KB.
Note of caution: The clarata_archive storage pool will now contain both 4+1
RAID-5 and 6+1 RAID-5 LUNs; as a result, 4+1 RAID-5 and 6+1 RAID-5
LUNs may be combined in a single file system. This is expected behavior.
16
Version 5.5.25.2
features
nas_halt command

Nas_halt is a command that can be used to perform an immediate
halt of the Celerra Network Server.

To halt the Celerra server, type:
# /nasmcd/sbin/nas_halt now
ARE YOU SURE YOU WANT TO CONTINUE? [ yes or no ] :
# yes
Sending the halt signal to the Master Control Daemon...:
Done
.
.
Halting system...
flushing ide devices: hda hdd
Power down.
It can take as long as 20 minutes to halt the server, depending on the number
of blades and the size of the database. Wait until the command completes
before continuing.
NDMP SnapSure
compatibility
To use a checkpoint for backup, set the NDMP environment variable

SNAPSURE to y on the NDMP client software. Celerra creates and
employs a checkpoint when the NDMP backup is initiated. If the

SNAPSURE variable is not configured or supported by the DMA
software, you can set the NDMP snapsure parameter to use SnapSure.
The NDMP variable SNAPSURE always overrides the NDMP
snapsure parameter.
The following table shows how the NDMP variable SNAPSURE and
the NDMP.snapsure parameter work together.
Table 2
NDMP SnapSure
NDMP variable
SNAPSURE (set in the
DMA product)
NDMP snapsure
parameter
(set at the Celerra)
Checkpoint created
Yes
Yes
No
No
Yes
No
17
Version 5.5.24.2
features
Automatic log
collection and transfer
using the Celerra
Manager
Celerra Manager provides an interface for the Automatic Log

Collection and Transfer feature. This interface lets you automatically
transfer collected data to a preconfigured FTP server. When
Automatic Transfer is enabled, you can initiate on-demand log
collection and transfer the logs from the Celerra Control Station to a
preconfigured FTP server.
When you discover a problem with Celerra Manager or your Celerra
Network Server, before calling your service provider or EMC for
help, collect support information to help correct the problem by
running the automatic collection script on the Control Station.
EMC recommends that you run this script as soon as possible after
you discover a problem. The sooner you run this script, the more
likely you are to capture information that can help correct the
problem.
To run the script to collect support information into a file:
1. In Celerra Manager, select Tools > SSH Shell and log in to the
Control Station as root.This opens a separate window containing
a secure shell in which you can run the script on the Control
Station.
2. In the secure shell window, enter this complete path to run the
script:
/nas/tools/automaticcollection -getlogs
The script compiles Celerra Network Server support information
into a compressed tar file and saves the file in the /nas/var/log
directory. The tar file is given this name:
support_materials_<serial number>.<yymmdd_hhss>.tar.gz
Where <serial number> is the serial number of the Celerra and
<yymmdd_hhss> is the time the script was run.
After you create the support materials file, contact your service
provider or the EMC Support Center for directions on how to deliver
the file for analysis.
The Problem Resolution Roadmap for Celerra technical module describes
how you can set up the automatic transfer of the support materials
file to EMC.
18
Automatic collection
and transfer using
command line
interface
The Automatic Collection and Transfer tool works in conjunction

with the collect_support_materials script as an aid in
troubleshooting and resolving system errors and other problems with
the Celerra Network Server.
The tool, which you run from the command line interface (CLI),
produces zipped files of the logs needed to analyze issues. If the
transfer feature is enabled, it automatically transfers the collected
data in a zipped file via FTP to EMC. You can also disable the transfer
capability and send the files manually to EMC. If the transfer feature
is disabled, it creates dumps in /nas/var/dump.
Complete documentation of the Automatic Collection and Transfer is
located in the Celerra Network Server System Operations technical
module.
CX disk arrays
NS80
Celerra Network Server supports CX3-80, CX3-40 and CX3-20

CLARiiON disk arrays including Fibre Channel, Low Cost Fibre
Channel, and NorthStar SATA-II drives.
The Celerra NS series includes the NS80.
Integrated NS80
The Integrated NS80 has two blade enclosures and can be configured
with two, three, or four blades and one or two Control Stations. The
NS80 is available with two types of blades. The X-Blade 60 has eight
ports that support six 10/100/1000 copper Ethernet ports and two
optical Gigabit Ethernet ports. The X-Blade 65 has the same eight
ports plus an additional optical 10 Gigabit Ethernet port located in an
extra module at the bottom of the enclosure.
This system comes with an AUX integrated array and one DAE that
holds 6 to 15 disks. You can add additional DAEs to increase the
available storage.
server_cifs -audit
The usage for -audit option of server_cifs has changed to:

server_cifs { <movername> | ALL } [options]
'options' can be one of the following:
| -option { audit [[client=<machine_name>] | [user=<user_name> ]]
| homedir[=NO] }
Previously, both the client and user options could be entered on the
same command line. With the new usage, only one option, client or
user, can be entered per command line.
19
Version 5.5.23.2
features
Expanded device
number support

The major and minor device number has been extended from 16-bit
to 32-bit. This enhancement enables NFS and CIFS clients to create
special device files with either 16 or 32 bit major and minor device
numbers.
Major and minor numbers are associated with the device special files.
The major number is used by the operating system to identify which
device driver is used to access a particular device. The minor number
identifies special characteristics of the device being accessed.
Low-cost Fibre
Channel drives
Multilevel NFS exports
Celerra supports CLARiiON storage systems with low-cost Fibre

Channel (LCFC) drives.
The Network File System (NFS) protocol supports the multilevel
export of file systems, that is, the export of any part of a file system
independent of existing exports. This feature is supported by all
versions (2, 3, and 4) of the NFS protocol.
In previous releases, parts of a file system could be exported to NFS
clients but there was no provision for exporting a subdirectory under
an already exported path. Multilevel exports support the export of
multiple levels of a file systems directories. Users can export file
systems using the server_export command or the Celerra Manager
NFS Exports option.
Behavior
Access permissions, the level of read and write access available to the
NFS clients of each exported file system, may be defined differently
for each export. Conflicting access permissions are handled as
follows:
For NFS v2 and v3, clients mounting an exported path get the
same access permissions to the entire path even if a subdirectory
of that path is exported with some other permissions. Clients get
the access permissions of the exported subdirectory only if they
mount the subdirectory and access that mount point.
The way NFS v2 and v3 clients handle conflicting access permissions
deviates from the NFS v2 and v3 protocol standard.
20
For NFS v4, access permissions are calculated every time a client
crosses a directory.
For example, if a NFS v2 and v3 client is given read-only access to

/ufs1 and read-write access to /ufs1/dir1 and that client
mounts /ufs1, all data in /ufs1 is read-only, including
/ufs1/dir1. If the client mounts /ufs1/dir1, all data in
/ufs1/dir1 is read-write.
An NFS v4 client always accesses the pseudo-root of the server and
then navigates until it locates the file system it is trying to mount.
NFS v4 does not have real mounts and access permissions are
verified every time a directory is crossed.
Restrictions
Multilevel NFS exports are not supported across file system
boundaries.
Multilevel NFS exports do not work for file systems mounted on

a nested mount file system (NMFS).
Examples
For additional information about multilevel exports, refer to the
server_export man page.
NS80G
The Gateway NS80G has two blade enclosures and can be configured
with two, three, or four blades and one or two Control Stations. The
NS80G is available with two types of blades. The X-Blade 60 has eight
ports that support six 10/100/1000 copper Ethernet ports and two
optical 1 Gigabit Ethernet ports.
The X-Blade 65 has the same eight ports plus an additional optical 10
Gigabit Ethernet port located in an extra module at the bottom of the
enclosure. They are always fabric-connected to a CLARiiON or
Symmetrix array
Version 5.5.22.2
features

NS40
The Celerra NS series now includes the NS40.

Integrated NS40
The Integrated NS40 consists of a complete Celerra Network Server
and a CLARiiON CX storage system in a single, compact cabinet
(referred to as an integrated array). The NS40 supports a single
Control Station and is available with a single or dual Data Movers.
21
Gateway NS40
The Gateway NS40 consists of a Data Mover enclosure and a single
Control Station connected to CLARiiON CX series or Symmetrix
storage, using either a direct connection or a Fibre Channel-switched
SAN connection. The NS40G is available with a single or dual Data
Movers.
Version 5.5.21.4
features
FLARE 22

FLARE 22 is no longer supported as of this release. If you are running
release 22 or earlier of the CLARiiON backend software FLARE , you
will need to upgrade to FLARE release 24 and should do so prior to
upgrading NAS. You will receive a warning during the upgrade if
your CLARiiON software is not a supported version.
Version 5.5.19.4
features
Automatic File System

Extension
Automatic File System Extension is a configurable Celerra file system

feature that automatically extends file systems created with
Automatic Volume Management (AVM) when they reach their
specified high water mark (HWM), without manual intervention.
Automatic File System Extension can be configured at the time of file
system creation or on an existing file system.
You can use both system-defined and user-defined storage pools to
create or extend the file system and enable Automatic File System
Extension. As long as there is space in the storage pool, when the file
system reaches the HWM, Automatic File System Extension
guarantees the percentage used for the file system to be below the
HWM.
If Automatic File System Extension is enabled, you can set the
maximum size to which the file system can grow. You can also enable
virtual provisioning, which permits the file system to slowly grow on
demand as data is written. With virtual provisioning, the space
presented to the user or application is the maximum space available,
while only a portion of that is actually allocated to the file system.
For more detailed information about auto extension, refer to the
Managing Celerra Volumes and File Systems with Automatic Volume
Management technical module.
22
NFSv4
The NFS version 4 protocol [RFC3530] is a further revision of the NFS

protocol defined by versions 2 [RFC1094] and 3 [RFC1813]. It retains
the essential characteristics of previous versions (a design
independent of transport protocols, operating systems, and file
systems) and integrates file locking, strong security, operation
coalescing, and delegation capabilities to enhance client performance.
Access control lists
NFSv4 adds support for access control lists (ACLs), which provide
finer-grained user access control to file system objects than traditional
NFS mode bits. The Celerra Network Server has a single
implementation of ACLs, which provides consistent access control no
matter which protocol (NFS or CIFS) is used to access data.
Client contexts
Unlike NFSv2 and v3, which are stateless protocols, the NFSv4
protocol maintains context or states for the client systems accessing
the file server and for the locks being held, allowing client recovery of
locking state after a server or client failure or network partition.
Delegation
When using NFSv4, the Celerra Network Server can delegate specific
actions on a file to a client, specifically more aggressive client caching
of data, metadata, and locking. Delegation improves network
performance by allowing NFS clients to locally buffer file data and
metadata and perform operations on that data and metadata before
sending it to the server.
File locking
NFSv4 file locking is similar to the adjunct Network Lock Manager
(NLM) protocol used with NFSv2 and v3, but is now integrated into
the NFSv4 protocol, which provides better support for different
operating system semantics and error recovery.
I18N
NFSv4 uses UTF-8 for character encoding. By default, a Celerra
system is configured for ASCII Latin-1. Before accessing data through
NFSv4 clients, change the default configuration in
/nas/site/locale/xlt.cfg. For more information, refer to the
Using International Character Sets with Celerra technical module.
User authentication
NFSv4, in addition to NFSv3, supports secure NFS. Secure NFS
provides Kerberos-based user and data authentication, data integrity,
and data privacy, and can use either a UNIX, Linux, or Windows
Kerberos KDC. All NFS versions support UNIX security and a PC
authentication daemon.
23
For more information about NFS, refer to the Configuring NFS on

Celerra technical module.
Support for LDAP
signing and
encryption
The Celerra Network Server provides signing or encryption for the

LDAP messages sent between the Data Mover and the Active
Directory. During an LDAP BIND process, the domain controller
negotiates one of the following security levels with a Data Mover:
Signed: Checks the integrity of each LDAP packet on the network

to ensure an intermediate party did not tamper with its contents,
but does not provide encryption.
Encryption: Prevents the data in the LDAP packets from being

sent in clear text, but does not check their integrity.
None: Does not sign or encrypt any LDAP messages.
By default, when a domain controller proposes a security layer of

signing or encryption to the Data Mover, it responds with signing
(integrity protection without encryption). You can modify this
behavior using the CIFS ldap.SecurityLayer parameter.
Windows server 2003
access-based
enumeration (ABE)
The Celerra Network Server supports ABE, which was introduced in

the Windows Server 2003 Service Pack 1. ABE filters shared folders
visible to a user based on that user's access rights. This prevents the
display of folders and files of which the user does not have full-read
access.
ABE functionality does the following:
Increases folder-level security.
Improves administrator productivity by reliably streamlining

large directory structures for less-technically savvy users.
Provides a more seamless migration experience for end users

migrating to Windows servers.
Requirements
ABE requires the following:
Windows 2003 Server Service Pack 1

User be a member of a Windows Administrators Group or
Account Operators Group.
Exceptions
ABE has the following exceptions:
ABE does not affect a user that has backup privileges on a CIFS
server.
ABE affects only on the last path component of a share request.

Parent directories are not affected.
24
Enabling and disabling ABE

In Microsoft Windows, you can enable and disable ABE through a
graphical user interface (GUI), a command-line executable tool, and
an advanced programming interface (API). Additionally, EMC offers
its own command-line tool called EMCABE to enable, disable, and
display the status of ABE shares on a CIFS server. Once enabled,
EMCABE runs without user intervention until disabled.
For more information on access-based enumeration, refer to the
Microsoft Windows Server 2003 white paper on Access-based
Enumeration. The Using EMC Utilities for the CIFS Environment
technical module provides information about using the EMCABE
tool.
Enhancements to
existing features
ACL Checking
The ACL checking utility was enhanced to remove duplicate ACL

records stored on a specified file system. When enabled, the
-aclchkonly option of nas_fsck removes duplications from the
ACL database. This change will reclaim slots in the ACL database
when aclchk is run. Full ACL databases can cause CIFS users to be
unable to access shares, files, or directories.
The -aclchkonly option of nas_fsck can only be used on a file
system that is mounted read/write, but is not exported.
Celerra Events and

Notifications
As of version 5.5.30.4, SNMP community names used when

specifying SNMP trap address for any of the four types of
notifications can contain the following special characters ~ ! @ $ % ^
* + = { } : ? _ #.
Characters A-Z, a-z, 0-9, . and - were already allowed. (Note: These
are shown with a spaces for readability, but no spaces are allowed.)
The hostname/IP restrictions didn't change, hostnames can only be
A-Z, a-z, 0-9, - (dash or minus sign), and . (period) as a delimiter in a
domain-style name. The following special characters are still not
accepted due to restrictions with Linux and with the notification
implementation: & ( ) | [ ] \ " ' ; < > , / space.
Refer to the Configuring Celerra Events and Notifications technical
module for more information.
Celerra FileMover
Celerra FileMover was enhanced to allow access to secondary storage

systems using HTTP V1.1 protocol. Users can create, modify, query,
and delete HTTP connections via the command line interface (CLI).
25
File system scanning performance

EMC has enhanced Celerra FileMover to improve file system
scanning performance for policy and migration software. The
metadata attributes retrieval operation allows FileMover policy and
migration software to scan a file system and locate files matching
criteria defined by the policy and migration software. A variety of
search criteria can be specified in the scan request, such as timestamp,
file size, and file type.
When matching files are found, the metadataor file attributesthat
is associated with those files is written to an output file located on a
Celerra file system. Because the content of output files may be
sensitive in nature, EMC strongly recommends creating a file system
specifically to hold output files and be accessible only to the policy
and migration software that initiated the metadata retrieval request.
Under normal circumstances, the policy and migration software
deletes output files that are no longer required. Occasionally, the
administrator may have to clean up output files if the policy and
migration software fails to do this.
The Using Celerra FileMover technical module contains detailed
information about automated migration of files across a hierarchy of
storage platforms.
Celerra Manager support information
To generate a compressed tar file that collects Celerra Manager
support materials (logs, configuration details, versions) that are ready
for FTP transfer, type: nas/tools/webui_support_materials.
The resultant file will be placed in the /tmp directory (or in $TMPDIR if
you have that environmental variable defined) under the name
webui_support_logs.<yymmdd_hhss> .tar.gz, where
<yymmdd_hhss> is the timestamp when the script ran.
Collecting logging information
You can use the script /nas/tools/collect_support_materials to
gather substantial logging information.
Celerra Manager
26
GNU screen support

The 'screen" package has been added to the Control Station. Screen
allows virtual shell sessions that are persistent across logins to run on
the Control Station. Screen can thus be used to perform procedures
remotely that need to survive a network or modem disconnect. More
information about this package can be found in the screen(1) man
page.
Celerra Manager performance

The following changes were made to the Celerra Manager in version
5.5.30.4 to improve performance:
inUse filter
The default volume type which is selected on the 'Volume New Page'
has been changed from 'Slice' to 'Stripe'. Version 5.5.30.4 introduces a
new filter 'inUse' to query unused volumes. This change takes effect
on the 'Volume New' screen for 'Meta'/'Stripe' type of volumes. The
'inUse' filter, does not require querying 'clients' property. This
improves the page performance.
NAS queries
The 'nas_fs' query returns all file systems, including checkpoint file
systems, when no filter is specified. APL was making 'nas_fs' query
without specifying any filter. APL code has been modified to use
filters to query only the required type of file systems.
APL was using nas_query to populate 'slice table' internally. APL
code has been changed to make use of CORE code to get this data to
improve the performance.
nas_pool
The 'nas_pool' command and 'nas_pool' query were both used to

populate storage pool details. Code has been modified to use
'nas_pool' command output only when required.
The above changes improve performance for the following screens:
Volume New, Storage Pool New, Storage Pool List , Volume List, and
Volume Properties.
CIFS server registry
A value has been added to the Celerra CIFS server registry that can
be changed to make the referrals contain the Fully Qualified Domain
Name (FQDN) rather than NetBIOS.
HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Services\Dfs
Value Name: DfsDnsConfig
Data Type: REG_DWORD
Value Data: 0 or 1
If you set the data value to 1, all roots added to the DFS tree use an
FQDN. A value of 0 (the default) specifies that NetBIOS names are
used.
Refer to the Managing Celerra for the Windows Environment technical
module for information on the CIFS server registry.
27
Increased SnapSure
and Replication limits
Celerra supports a maximum of 96 checkpoints per PFS, whether the

PFS is replicated or not, for all Data Movers except model 510 (which
permits 32 checkpoints with PFS replication and 64 without).
SnapSure and Replication limits have been increased dramatically
due to improvements in paging SavVol data to disk and reducing
memory overhead for mounted file systems.
The maximum number of file systems involved in a replication
session has been increased to the following: 64 file systems for NSX
and NS80 Data Movers, 32 file systems for all other Data Movers.
Table 3
SnapSure and Replication limits

Maximum number
of file systems
involved in a
replication session
Maximum
number of
checkpoints if
PFS is replicated
Maximum number of
checkpoints if PFS is
not replicated
NSX/NS80
64
96
96
Any Data Mover

except NSX and
510
32
96
96
510 Data Movers
32
32
64
Data Movers
Operational environment and limits

The following operational environment and limits apply:
For 510 Data Movers, SnapSure supports 32 checkpoints with PFS

replication and 64 without.
In version 5.5, Celerra supports the same number of checkpoints

per file system whether replication is running or not.
Checkpoints created/used for replication are counted against the

total number of checkpoints supported.
The total number of supported file systems per Data Mover is

2048. The total number of supported file systems per Celerra
cabinet is 2048 for non-NSX platforms and 4096 for NSX
platforms.
The Using SnapSure on Celerra and Using Celerra Replicator technical

modules contain additional information on checkpoint and
replication session limits.
28
Increased stripe size

for automatic volume
based-file systems
In Version 5.5, the default stripe element size has increased from 8 KB
to 32 KB for the clar_r5_performance, clar_r5_economy, and clar_r1
AVM storage pools. This change improves backend performance and
allows you to enable MPFSi on file systems.
If space remains available in used disk volumes (existing pool
members) of the storage pool, the stripe size remains 8 KB. When you
allocate space from unused disk volumes (new pool members), the
stripe element size automatically defaults to 32 KB.
When a new file system is created or an existing file system is
extended, the stripe element size is determined by:
The available space remaining in the used disk volume
Availability of unused disk volumes
nas_pool -is_greedy=True is set.

In any case, there is no performance impact.
If you plan to create a file system manually or intend to create
user-defined storage pools, carefully consider the size of the stripe
volume you want. After the stripe volume is created, its size remains
fixed. Enter stripe depth in multiples of 8,192 bytes with a
recommended size of 32,768 bytes (default) for file systems running
in a CIFS or NFS environment with a Symmetrix or CLARiiON
storage system. Stripe size of 256 KB is recommended for MPFSi and
HighRoad environments; RAID 3 and 64 KB stripe size are
recommended for ATA-based file systems.
Integrated platforms
NS20, NS40, NS20FC,
or NS40FC installation
During a fresh install of an NS20, NS40, NS20FC or NS40FC, if you

choose not to setup an external network on the Control Station, on the
next reboot, the cable_check startup script will prevent the Control
Station from booting up all the way and wait for it to be configured
using the Celerra Startup Assistant (CSA). If you want the Control
Station to boot up all the way, delete the
/etc/rc3.d/S95cable_check file after the NAS installation.
Integrated NS80 and

NS20, NS40, NS20FC,
NS40FC installation
CLARiiON SPs may reboot to update the gateway addresses. New

SPA Gateway will be <Primary subnet>.104 (e.g. 192.168.1.104) and
new SPB Gateway will be <Secondary subnet>.104 (e.g.
192.168.2.104). This will facilitate SPA - SPB communication via the
Control Station (primary / secondary) required for Secure NaviCLI
support. This process may take a while (15-20 minutes).
29
AccessLogix will be enabled, a Storage Group for Celerra will be

created. Information about the Storage Group will be stored in the
/etc/be_sg_info file on the Control Station. Do not modify or delete
this file.
Secure NaviCLI will be used. The default username and password
used will be nasadmin/nasadmin. Commands like navicli,
nas_raid, setup_clariion2 should continue to work as before
(unless you change the CLARiiON username and/or password).
Upgrades
During an upgrade, you may get a message (on the screen and in the
upgrade log) to update Gateway addresses on CLARiiON SPs. You
are not required to do so, unless you want to use Secure NaviCLI. To
update the gateways, use the "/nas/sbin/navicli -h <SPA|SPB
networkadmin -set -gateway <gateway" command. Use <Primary
subnet>.104 for SPA gateway and <Secondary subnet.104 for SPB
gateway.
Integrated NS500,
NS600, NS700 and
NS350 installation
During a fresh install, Secure NaviCLI will be setup, if supported by

the FLARE release on the CLARiiON. The default username and
password used will be nasadmin/nasadmin. Commands like
navicli, nas_raid, setup_clariion2 should continue to work as
before (unless you change the CLARiiON username and/or
password).
Multi-Protocol
Directories
Earlier versions of Celerra Network Server stored CIFS names and

attributes separately from the primary directory information in an
auxiliary database called the shadow directory. Keeping this
information separate would sometimes result in data consistency
problems. Beginning with version 5.2, the Multi-Protocol Directory
(MPD) feature stores NFS and CIFS names directly in the file system
directory, improving overall CIFS performance.
MPD is a three-name directory format comprised of the NFS name,
the CIFS long name, and the CIFS short name. When a Celerra was
upgraded to versions 5.2, 5.3, or 5.4, one or more file systems could be
excluded from the MPD translation process. These file systems were
listed in /nas/site/fs_exclude.
In version 5.5, this exclusion is no longer supported, because the
COMPAT (single-protocol directory) type file system is no longer
supported. During an upgrade to version 5.5, if the exclude file is
detected, the upgrade script stops. The exclude file must be removed
before the upgrade can continue.
30
Space requirements
During MPD conversion, a temporary directory is created, which
consumes disk space. If there is insufficient disk space for the file
system upgrade to complete, the upgrade pauses and an event log is
generated. The upgrade periodically attempts to continue. When
sufficient disk space becomes available, the upgrade will complete.
Quotas
Disk quota limits can cause an upgrade to pause. If disk quotas are
used on the file server, upgrading can change the amount of space
charged to the user, group, or tree. In particular, translation from
single- to three-name entries will likely cause an increase in the
charged quota. Translation can exceed a soft quota limit, but will
never exceed a hard-quota limit.
If a quota limit pauses a file system upgrade, a log entry is written
and a log event generated, and the translation thread continues with
other users, groups, or trees. Afterward, the thread suspends until
more space is available in the quota. All quota trees in a file system
must be translated successfully before any directory not in a quota
tree can be translated.
Timestamp
Directories are rebuilt during MPD translation, a process that
involves writing to it. The atime and mtime timestamps are changed
to the moment when the translate thread commits the changes for
any particular directory.
Unicode
If Unicode is enabled and there are non-ASCII file names in a
shadowed COMPAT directory that have not been converted to the
proper Unicode format, the MPD translate thread is unable to
translate those files. You should run Unicode conversion on the Data
Mover to ensure that all files and directories have been converted to
the proper Unicode format before running the MPD translate
command again.
The Using International Character Sets with Celerra technical module
provides more information on how to convert file systems to
Unicode.
Querying for current MPD directory type
Use the following command to display MPD information about the
specified file system including directory type and file system
translation status.
31
Action
To query the current directory type and translation status, use the following syntax:
$ nas_fs -info <fs_name> -option mpd
Where:
<fs_name> = file system name you want to query
Example:
To query the translation status on file system ajm03, type:
$ nas_fs -info ajm03 -option mpd
Output
id
=
name
=
acl
=
in_use
=
type
=
volume
=
pool
=
member_of =
rw_servers=
ro_servers=
rw_vdms
=
ro_vdms
=
symm_devs =
401470-0016
disks
=
disk=d28
disk=d28
disk=d9
disk=d9
disk=d37
disk=d37
disk=d11
disk=d11
56
ajm03
0
True
uxfs
v224
clar_r5_performance
root_avm_fs_group_3
server_2
APM00034401470-0019,APM00034401470-0012,APM00034401470-002B,APM00034
d28,d9,d37,d11
symm_dev=APM00034401470-0019
symm_dev=APM00034401470-0019
symm_dev=APM00034401470-0012
symm_dev=APM00034401470-0012
symm_dev=APM00034401470-002B
symm_dev=APM00034401470-002B
symm_dev=APM00034401470-0016
symm_dev=APM00034401470-0016
addr=c16t1l9
addr=c0t1l9
addr=c0t1l2
addr=c16t1l2
addr=c16t2l11
addr=c0t2l11
addr=c0t1l6
addr=c16t1l6
Multi-Protocol Directory Information

Default_directory_type
= DIR3
Needs_translation
= True
Translation_state
= Running
Percent_nodes_scanned
= 7%
Has_translation_error
= False
Skipped_dir_inodes
= 0
Current_inode
= 460175
Entries_in_current_dir
= 13 of
81
32
server=server_2
server=server_2
server=server_2
server=server_2
server=server_2
server=server_2
server=server_2
server=server_2
Note
The output of this command consists of key-value pairs:
default_directory_type={DIR3 | COMPAT} The default directory type for the file system or
volume.
needs_translation={True|False} If true, the file system might contain more than one directory
type. If false, then all directories are of the file system default directory type.
percent_inodes_scanned={percent} The number of inodes examined during this translation
instance, expressed as a percentage of the total number of inodes in the file system (displayed
only if needs_translation is true).
translation_state={never|not requested|pending|queued|running|paused|completed| failed}
The current state of the translation thread. States have the following meanings:
never: Translation has never been requested on this file system. Scheduling translation
changes this state to pending if the file system is unmounted or queued if mounted.
not requested: Translation is not scheduled on this file system. Scheduling translation changes
this state to pending if the file system is unmounted or queued if mounted.
pending: Translation is scheduled, but the file system is not mounted. When the file system is
mounted, it creates a translation object and becomes state queued. If translation is turned off,
the state returns to not requested.
queued: Translation object exists and is ready to be processed, but is waiting for a thread to
become available for processing. When a thread becomes available, the state changes to
running.
running: Translation thread started and is now running.
paused: Translation is waiting for a resource (user quota, group quota, tree quota, disk space,
or inodes). The state becomes queued when the condition for waiting is met.
completed: Translation completed successfully and is not needed.
failed: Translation exited after skipping directories with errors. Translation is still scheduled, and
will restart on the next mount. It can be manually restarted from this state (translation does not
self-resume from this state).
reason_for_wait=STRING If the translation state is paused or queued, the reason is displayed.
has_translation_error={True|False} Indicates if the most recent translation encountered an
error.
translation_error_code=INTEGER The last error (if any) encountered during the last translation
(if translation state is pending) or the current translation (if translation state is paused, or running).
translation_error_message=STRING The error message associated with
translation_error_code.
skipped_dir_inodes=INTEGER The number of directories skipped due to either a hard error or
exceeded quota. Translation returns to any directories skipped due to quota errors, but if there are
other errors, translation will fail after completing this scan of the entire file system.
current_inode=INTEGER If translation state is running, pending, or queued, the current inode
examined is displayed.
entries_in_current_inode=<N> of <T> If translation state is running, pending, or queued, the
number of entries processed of the directory total is displayed.
33
Definitions
The following table defines the possible directory types and
translation states:
Table 4
Directory types and translation states

Value
Definition
Default_directory_type
The default directory type for the file system.

Available types are: DIR3 and COMPAT.
Needs_translation
If true, the file system might contain more than one

directory type. If false, all directories are of the file
system default directory type.
Translation_state
The current state of the translation thread. Available

states are: never, not requested,
pending, queued, running,
paused, completed, and failed.
Has_translation_error
Indicates if the most recent translation encountered

any errors.
Required action
The following table lists the action required, based on the directory
type and the translation state:
Table 5
Required action
If the
and the
Default_directory Needs_translation
_type is:
state is:
then, the file system:
DIR3
False
is MPD. No action required.
DIR3
True
requires translation or file system maintenance.

Contact EMC Customer Service.
COMPAT
False
is COMPAT and requires translation. Contact

EMC Customer Service.
COMPAT
True
requires translation. Contact EMC Customer

Service.
When Default_directory_type=DIR3 and

Needs_translation=False, this file system's directories are all in
MPD (Multi-Protocol Directory) format, and there are no directories
of the obsolete single-protocol format.
34
Any other combination of states, for example,

Needs_translation=True, indicates that the file system might
contain non-MPD directories, which may not be compatible with a
future release.
NDMP
When performing NDMP backups on Celerra, Celerra Network

Server Version 5.5 offers the following new features designed to
improve the process, performance, and accuracy of backups.
Celerra NDMP Volume Backup

File and directory filtering
Improved DAR results
Integrated checkpoints
Celerra NDMP Volume Backup

Celerra NDMP Volume Backup is an EMC-specific type of NDMP
backup mechanism that backs up data blocks at a volume level,
rather than at a file level. Celerra NDMP Volume Backup reads a set
of disk data blocks in an efficient manner compared to the method
used for traditional, file-based backups. Celerra NDMP Volume
Backup supports the following types of restore:
Full Destructive Restore When using the data you created with
the Celerra NDMP Volume Backup feature to perform a full
destructive restore, the restore, like the backup itself, occurs at the
volume level (volume-by-volume). This type of restore ensures
the best possible write performance.
File-Level Restore When using the data you created with the
Celerra NDMP Volume Backup feature to perform a file-level
(file-by-file) restore, you can selectively restore a subset of the
backup image.
File and Directory Filtering

The File and Directory Filtering for Backup Clients feature lets you
exclude specific directories and/or files within a file system from an
NDMP backup session. The filtering feature is implemented using
one or both of these environmental variables in the qualified
backup-vendor software: EMC_EDIR0[01-05] and
EMC_EFILE0[01-05]. These variables let you avoid backing up
noncritical data such as .temp directories and .mp3 files.
The benefits of filtering include:
Shorter backup windows
Efficient use of tape (that is, space is not consumed by useless

data)
More flexibility to structure a backup session

35
Improved DAR results

In Celerra Network Server Version 5.5 (and later), improved DAR
functionality, called directory direct access restore (DDAR), optimizes
the data recovery operation by allowing the NDMP client to directly
access backed-up data anywhere in a tape set without having to
traverse the tape set sequentially. This method can save significant
time in the restore operation, especially when restoring many small
files. DDAR is realized only when you use EMC-qualified backup
vendor software that fully supports DDAR.
Other improvements introduced with version 5.5 include
enhancements in the PAX DUMP format incremental backup
operation and the three-way DAR restore operation. With DUMP
incremental backups, only the changed files and the directories
leading up to them are backed up, which can save a lot of tape space
for a big file system. The three-way DAR performance enhancement
changes the way the data is required, and can dramatically improve
the performance regarding small files. No environmental variable
changes are required to realize these enhancements; the
improvements are part of the Celerra Network Server Version 5.5
(and above) software.
If you use a version 5.4 backup source to recover a file system on a Celerra
system running version 5.5, the files under a specified directory will be
recovered, but it may take longer than expected because the restore will be
non-DAR. A non-DAR restore means the entire tape is searched sequentially
for the data because no index is available. It is a time-consuming process.
The DDAR functionality does not support Celerra NDMP Volume

Backup (that is, DDAR does not occur when the VBB backup
type/variable is used).
CA BrightStor ARCserve backup software
CA BrightStor ARCserve backup software version 11.1 is qualified
with Celerra Network Server V5.4.19.4 and later. If you are using CA
BrightStor ARCserve 11.1 with any Celerra Network Server Version
earlier than V5.4.21.0, ensure that the Celerra NDMP parameter
v4OldTapeCompatible is set to 1 (the default value). However, if you
are using CA BrightStor ARCserve v11.1 with Celerra Network
Server Version 5.4.21.0 or later, ensure that the Celerra NDMP
parameter v4OldTapeCompatible is set to zero (0). The Celerra
Network Server Parameters Guide explains how to change parameter
settings.
36
CA BrightStor ARCserve backup software version 11.5 is qualified

with Celerra Network Server V5.4.21.4 and later. If you are using CA
BrightStor ARCserve 11.5 with Celerra Network Server Version
5.4.21.4 or later, ensure that the Celerra NDMP parameter
v4OldTapeCompatible is set to zero (0). The Celerra Parameters Guide
explains how to change parameter settings.
Dynamic Drive Sharing
The following vendors are EMC-qualified to support Dynamic Drive
Sharing:
CommVault Galaxy (Version 5.9 and later, requiring a SCSI-3

Reserve license)
CA BrightStor ARCserve (Version 11.1)
If you plan to use Dynamic Drive Sharing (DDS), set the Celerra
Network Server's NDMP scsiReserve parameter to zero (0) on the
Data Mover (running Celerra Network Server Version 5.1 and later)
where devices are connected.
Integrated checkpoints
A checkpoint serves as a read-only source file system for a backup,
allowing production activity on the file system to continue
undisrupted. Now, by setting an environmental variable
(SnapSure=y) in the EMC-qualified backup-vendor software or
setting parameter NDMP.snapsure=1, you can automatically integrate
a checkpoint into all the NDMP backup sessions you run. This means
that each time you run a backup session, a checkpoint is
automatically created, used for the backup, and then deleted when
the backup completes. Integrating a checkpoint into the backup
process eliminates the need to manually perform these checkpoint
management tasks each time, saving both time and resources.
Backups can be run more frequently and/or at times when manual
intervention may not be available.
SRDF and MirrorView/S
support for iSCSI
configurations
SRDF/A, SRDF/S, and MirrorView/S on Celerra support the remote

mirroring of a source iSCSI LUN configuration at a destination site
for disaster recovery (DR) purposes. With this change in support,
MirrorView/S no longer blocks iSCSI operations involving file
systems mounted on a Data Mover configured for MirrorView/S,
and no longer reports iSCSI configuration error messages. (SRDF did
not previously enforce a block of iSCSI configuration.)
Note: MirrorView/S is currently available as RPQ only.
37
The remote mirroring of source iSCSI LUNs using SRDF or

MirrorView is transparent to Celerra iSCSI operations. Each iSCSI
LUN is equivalent to a file within a file system on the Celerra;
therefore, to establish remote mirroring of iSCSI LUNs, the
appropriate file systems must be mounted on a source Data Mover
configured for SRDF or MirrorView/S (that is, an SRDF- or
MirrorView-protected Data Mover must provide the file system
representing the iSCSI LUN and the root file system of the Data
Mover managing LUN visibility).
As described in the SRDF and MirrorView documentation, both
SRDF and MirrorView/S on the Celerra rely on backend storage to
protect and mirror source production data at a remote destination. In
a disaster scenario where the source site becomes unavailable, a
failover ensures that the destination continues to provide the source
production environment. For SRDF/S or MirrorView/S, the
mirroring is synchronous to ensure identical copies of the
source-protected data at the destination; for SRDF/A, the mirroring
is asynchronous to provide time-delayed copies at the
source-protected data at the destination.
For iSCSI configurations employing SRDF or MirrorView/S for
remote mirroring and disaster recovery, iSCSI behavior can be
characterized as follows:
38
iSCSI adheres to the same restrictions that apply to SRDF and

MirrorView/S. (See the next section for more information.)
During normal, steady-state operation, the destination-site copy

of the production data is inaccessible, even read-only. Therefore,
before a failover is activated (for example, in response to a
disaster scenario), no iSCSI host can access the destination-site
copy of the production data.
When a failover is activated, the failover is cabinet-level, that is,

the destination Celerra now provides the entire DR-protected
environment of the source Celerra. You cannot activate a failover
for a single LUN, file system, or Data Mover.
Upon activation of the destination site, the iSCSI LUNs will be

crash-consistent, appearing as if the source disk went offline
without the host or application being able to shut down
gracefully. For SRDF/A, the destination will be crash-consistent
and there will be some amount of data loss, since the destination
is a period of time behind the source. (For an alternative form of
replication, you can consider using Celerra Replicator for iSCSI
instead.)
Since this implementation is for disaster recovery, not continuous

availability, regardless of whether a disaster affects the iSCSI host
directly, the iSCSI host will lose connectivity during the failover
activation, and it will need to reboot and recover services.
For example, if a disaster affects the source site (including an
iSCSI host logged in to an iSCSI target on the source-site
Celerra only), the destination site needs to be activated. An
iSCSI host at the source site should be expected to time out
and become inaccessible in the amount of time that it takes the
administrator to detect the failure of the source site and decide
to activate a failover to the destination site.
By default, a failover activation causes the destination site to
assume the same resources, IP addresses, MAC addresses, and
other metadata information as the source site.
The administrator can decide to use either an iSCSI host
located at the destination, or use the original iSCSI host at
the source site, if it was unaffected by the disaster.
Regardless of which iSCSI host is used, the iSCSI LUN on
disk is power-fail (or crash-consistent) and may need to be
recovered (using utilities such as fsck and chkdsk) before
being utilized. See Preventing possible Windows NT File
System inconsistency on page 40 for more information.
Limitations that apply to iSCSI configurations

The following limitations also apply to iSCSI configurations in the
backend replication environment:
Activation of the destination site means that the entire Celerra is

activated, not a single entity.
The same SRDF and MirrorView restrictions apply concerning

Control Station and GUI features not supported at the destination
after a failover activation.
Automatic File System Extension cannot be used for any file

system that is part of an SRDF or MirrorView/S configuration
(that is, file systems on Data Movers configured with a remote
standby).
The same SRDF and MirrorView restrictions apply concerning IP

replication and fs_copy; that is, Celerra Replicator and fs_copy
are not supported.
Celerra iSCSI configurations do not support TimeFinder/FS,

NearCopy, and FarCopy.
39
Preventing possible Windows NT File System inconsistency

If you use iSCSI with SRDF or MirrorView/S and plan to activate a
failover or perform a restore operation, try to use the following
procedure for your iSCSI hosts to prevent possible Windows NT File
System (NTFS) inconsistency:
1. Stop all Windows applications on the iSCSI hosts (that is, all
applications that use the Celerra iSCSI LUNs).
2. From the iSCSI Initiator Properties dialog box, navigate to the
target properties and log off from the iSCSI initiator session to
clear the session information.
You can now proceed to activate a failover or perform a restore.
If you use iSCSI with one of the backend replication products and
need to activate a failover in response to a disaster scenario, attempt
to perform these steps, then use the Windows chkdsk utility to check
NTFS-level file system consistency once the failover activation is
complete. Consult Microsoft for detailed information about
recovering data and the chkdsk utility.
Parameter updates
Parameters Guide
corrections
Note the following corrections to the Celerra Network Server Parameters

Guide.
filter.caseSensitive
The PAX facility parameter shown as file.caseSensitive is

incorrect. The actual parameter name is filter.caseSensitive.
tcpwindowlowat
For the RCP facility, the parameter tcpwindowlowat is all lowercase,
not tcpwindowLowat.
useDCLdapPing
The default value for the ds facility useDCLdapPing parameter is 1,

not 0.
New or changed
parameters
The following section lists the parameters that are new or have
changed in this release. Refer to the Celerra Network Server Parameters
Guide for a complete list of the documented NAS parameters.
CAUTION
You should not edit anything in /nas/sys. These settings are
overwritten with each Celerra Network Server code upgrade.
40
The following table lists the parameters introduced in version 5.5

Table 6
Parameters new to v5.5
Facility
Parameter
Values
Comments/description
cfs
showHiddenCkpt
0 or 1 (default)
Enables/disables the Celerra Virtual File System (CVFS) version 2

client access to checkpoints. By default, CVFS v2 is enabled.
Value 1 enables CVFS v2 so each mounted checkpoint of a
production file system is accessible to clients as a special virtual
subdirectory in the production file systems user directory.
Value 0 disables CVFS v2 and prevents accessing virtual
checkpoint subdirectories.
cifs
acl.archive
0, 1, or 3 (default)
Controls whether a change to the security descriptor (SD) of a file

or directory sets the DOS archive bit.
0 = Set the DOS archive bit when the SD of a file or directory
changes.
1 = Set the DOS archive bit when the SD of a file changes.
2 = Set the DOS archive bit when the SD of a directory changes.
3 = Do not set the DOS archive bit when the SD of a file or directory
changes.
Note: Changing the archive bit also changes the last modification
time of the parent directory since DOS attributes are stored in the
parent directory.
cifs
acl.checkAclCon 0 or 1 (default)
sistency
cifs
acl.extacl
0 (default) - 127
This parameter is intended to troubleshoot the inconsistent ACL.

When set to 1, the consistency of ACLs is checked during ACL
creation and on ACL access. A panic is generated when this check
finds an inconsistent ACL.
When set to 0, no check is done.
This parameter is a bit list that enables special capabilities for ACL
(access control list) management. There are two kinds of
capabilities:
Backup/restore specific UNIX attributes like access right, UNIX
mode, UNIX name and symbolic link, using a regular CIFS based
backup tool (like NTbackup)
View/change UNIX access rights from ACL management tool,
such as Windows Explorer
The bit list consists of 7 binary bits (bits 0 through 6, right to left).
Any combination of bits is allowed. Each bit is 1 when set;
otherwise 0.
41
Table 6
42
Parameters new to v5.5 (continued)
Facility
Parameter
Values
cifs
acl.extacl
(continued from previous Bit 0 set (0000001 or +1) Celerra presents the UNIX
page)
metadata associated with files and directories to CIFS backup
clients using a special ACE entry in the file or directory's ACL. This
ACE can take either of two forms. If bit 0 is not set, Celerra uses an
EMC ACE type (CIFS allows vendors to define their own ACE
types). If bit 0 is set, Celerra uses a standard ACE and encodes the
information in the SID associated with that ACE.
Bit 1 set (0000010 or +2) If bit 1 is set, the UNIX
permissions on files and directories on the Celerra can be viewed
and modified by Windows clients. The UNIX permissions are
presented as three additional ACEs in the ACL of each file and
directory. These ACEs can be viewed and modified by any CIFS
ACL management application such as Windows Explorer.
Bit 2 set (0000100 or +4) If bit 2 is set, Celerra presents the
UNIX permissions associated with files and directories in the ACL
of the files so that CIFS network backup applications can backup
and restore them from and to a Celerra file system.
Bit 3 set (0001000 or +8) With bit 3 set, Celerra presents
UNIX symbolic links as zero byte files with a special ACL that
captures the information associated with the symbolic link (for
example, its target). If bit 3 is not set, Celerra may follow symbolic
links on behalf of CIFS clients and hence a CIFS backup
application does not back up the symbolic links, but instead, the
files they point to. If this bit is set, Celerra follows symbolic links on
behalf of CIFS network backup clients. This means the CIFS
backup application backs up the symbolic links, not the files and
directories they point to.
Bit 4 set (0010000 or +16) Any file or directory on a Celerra
can have as many as three names in the file system: a UNIX name,
a long Windows or M256 name, and a DOS 8.3 name.
When a CIFS network backup client backs up a file, it can get the
long Windows and DOS 8.3 name using standard CIFS calls. If the
file has a UNIX name that differs from the long Windows name it is,
by default, not backed up or restored by CIFS network backup
applications.
If bit 4 is set, Celerra encodes the UNIX name of files and
directories in a special access control entry (ACE) in the access
control list (ACL) of the file so that CIFS network backup
applications can back up and restore all three names of files and
directories.
Table 6
Facility
Parameter
Values
cifs
acl.extacl
(continued from previous Bit 5 set (01000000 or +32) By default, there is no way for
page)
NFS v2 and v3 clients to view or modify the ACLs associated with
files and directories on the Celerra. The Celerra Applications and
Tools CD includes the tool emcsetsd , which allows NFS v2
and v3 clients to view and, if the user has permission to do so,
modify the ACLs associated with files and directories on the
Celerra. Bit 5 must be set for the emcsetsd client tool to work.
Bit 6 set (1000000 or +64) If set, bit 6 modifies the functions
enabled with bit 1.
If bit 6 is not set, UNIX rights applied to the file are the granted
rights plus the rights not denied by the discretionary ACL (DACL).
If bit 6 is set, UNIX rights applied are the granted rights less the
denied rights by the DACL. In addition, the request is rejected if one
of the 3 special ACE is inheritable. This is because when changing
rights on a directory, the client propagates rights down the tree to all
nodes (files and directories), which is typically not a desired
behavior. Setting this bit prevents it. In practice, this means that
ACLs for directories must be set using the Advanced panel in the
security properties within Windows Explorer.
Changes do not take effect until after the CIFS service is restarted.
Examples of bit string settings:
0000010 (bit 1 only = 2) Allows CIFS clients to view and
modify the UNIX permissions on files and directories using
Windows Explorer. All other settings use the default values.
1000010 (bit 1 + bit 6 = 66) Changes the way that the ACL is
translated into UNIX permissions on files and directories so that the
UNIX permissions applied to files and directories are the rights
granted by any grant ACE for the UNIX user/group/other less any
rights explicitly denied in any deny ACE for the UNIX
user/group/other.
1100010 (bit 1 + bit 6 + bit 5 = 98) Enables NFS v2 and v3
clients to view and modify the ACLs on files and directories using
the emcsetsd tool.
43
Table 6
Facility
Parameter
Values
cifs
acl.mappingErro 0 (default) 15
rAction
Defines the rules for unknown mapping between security, user, and
group identifiers (SID/UID/GID) on ACL settings. Two kinds of
errors might occur: the SID set in the ACL is unknown to the
domain controllers being used or the username is not yet mapped
to a UID/GID.
The bit list consists of 4 binary bits (bits 0 through 3, right to left).
Each bit is 1 when set; otherwise 0.
Bit 0 (0001 or +1): Store unknown SID.
Bit 1 (0010 or +2): Store SID with no UNIX mapping.
Bit 2 (0100 or +4): Enable debug traces.
Bit 3 (1000 or +8): Do lookup only in cache (secmap or globalSID
cache or per connection SID cache).
Refer to the cifs facility acl.retryAuthSid
parameter.
cifs
acl.sortAces
0 (default) or 1
Sort the (access control entries) ACEs in an ACL in the order

expected by Windows client.
Value 1 stores the ACE of an ACL set by CIFS or NFSV4 in the
following order:
First the DENY inherited ACE
then the ALLOW inherited ACE
then the DENY non inherited ACE
then the ALLOW non inherited ACE
Value 0 stores the ACE in the order they are received.
cifs
secmap.enable
0 or 1 (default)
Enables/disables CIFS Secure Mapping cache. Secmap stores all

the UID/GID mappings corresponding to SIDs to be sure they are
consistent. Each time a new mapping is used in the system,
Secmap adds it to its database. All mappings are then directly
retrieved from the Secmap cache without having to query any
external name resolution mechanisms (DC, NIS, Usermapper, ...).
Secmap is enabled only at CIFS start and disabled at CIFS stop.
EMC strongly recommends you not disable it.
44
Table 6
Facility
Parameter
Values
cifs
set_eas_ok
0 (default) or 1
CIFS allows applications to apply extended attributes to files. Some

applications use this mechanism to store proprietary information.
By default, Celerra rejects any request to set the extended
attributes of a file. This breaks some applications that rely on this
mechanism. When enabled, this parameter tells Celerra to accept
requests to set extended attributes on files. Note that this does not
mean that the Celerra will store any extended attributes for files,
simply that it will not send a request to set them. This allows backup
and migration applications to restore or migrate files with extended
attributes to the Celerra. The file data and standard attributes of
the files will be restored or migrated to the Celerra, but the
extended attributes will not be.
Value 0 prevents Celerra from accepting requests to set extended
attributes on files.
Value 1 enables Celerra to accept requests to set extended
attributes on files.
cifs
srvpwd.encryptA 0 or 1 (default)
ccountFile
cifs
srvpwd.maxhist
ory
Range: 1-10, (2 default)
Controls whether the account file is encrypted.

0 = Do not encrypt the account file.
1 = Encrypt the account file.
After a password for a CIFS server is changed, this parameter
indicates the maximum number of passwords to retain for
Kerberos authentication. The default value of 2 retains the most
recent password and the previous password.
Value 1 retains only the most recent password.
Value <2-10> retains the most recent password and n-1
(previous passwords).
ftpd
bounceAttackChk 0 or 1 (default)
FTP servers can allow an attacker to connect to arbitrary ports on

machines other than the FTP client (CVE-1999-0017).
Value 1 enables FTPD to verify the PORT command does not
contains an IP address different than the IP address of the client.
Value 0 disables this PORT command verification.
This parameter must be disabled if you want to transfer files
between two FTP servers.
ftpd
wildcharsInDir
0 (default) or 1
Enable (1) or disable (0) wild characters for directory names in the
commands dir or ls.
The wild characters are: [ * ?
45
Table 6
Facility
Parameter
Values
iscsi
AsyncEvent
0 or 1 (default)
Specifies whether Async Event Reporting or Async Logout

Message should be sent by target.
Value 0, target sends Async Logout Message.
Value 1, target sends Async Event Reporting.
ldap
SecurityLayer
0, 1, 2 (default), or 4
During an LDAP BIND session, this parameter sets the level of

security the Data Mover uses during negotiation with the domain
controller.
Value 0 No Security Layer: Responds with no security layer
regardless of what the domain controller proposes. In Windows
environments configured to enforce LDAP signing, the BIND
procedure will fail because of this value.
Value 1 Same as LDAP Server: Agrees with any security layer
proposed by the domain controller.
Value 2 Integrity Protection: Always proposes LDAP signing,
which checks the contents of the LDAP messages.
Value 4 Privacy Protection: Always proposes LDAP message
encryption, which prevents the data in the packets from being sent
in clear text.
Changes to this parameter take effect at the next LDAP Bind.
mount
allowNullCred
0 (default) or 1
Used to allow or disallow clients that send AUTH_NONE

credentials to access NFS exports. By default, mount requests from
clients that send AUTH_NONE credentials fail. If this parameter is
enabled, these clients can mount NFS file systems.
nbs
sparseTws
0 (default) or 1
Disables or enables virtually provisioned Temporary Writable Snap

(TWS) support.
Value 0 enables Celerra to create regular TWS if production LUN
is regular.
Value 1 forces Celerra to always create sparse TWS.
NDMP
forceRecursiveF 0 (default) or 1
orNonDAR
Controls whether to override the environment variable settings

(DIRECT and RECURSIVE) and force a recursive restore for
directories in non-DAR restoration operations.
0 = Do not override environment variable settings for recursive
restore.
1 = Force a recursive restore even if the DIRECT and RECURSIVE
environment variables are set (or default) to N.
NDMP
snapTimeout
Specifies the timeout (in minutes) for creation of the backup

SnapSure file system.
2 60
Default: 5
46
Table 6
Facility
Parameter
Values
nfsv4
32bitClient
0 or 1 (default)
Used to set the bit size of attribute values.

Value 1 requires the Celerra NFSv4 server to return attribute
values in 32-bits only.
Value 0 allows the Celerra NFSv4 server to return attribute values
in 64-bits only.
nfsv4
domain
Specifies the NFS version 4 domain name. The NFS v4 domain

A name in the form
name is necessary for the service to map users UIDs and group
abc.def.ghi.
Default is an empty string GIDs to UTF-8 encoded user and group names.
PAX
allowVLCRestoreT
oUFS
0 (default) or 1
Specifies whether to allow VLC backup to be restored to UFS file

system.
Value 1 allows VLC backup to be restored to UFS file system.
PAX
checkUtf8Filenam
es
0 or 1 (default)
Specifies whether to check if the filenames are in valid UTF8

format.
Value 0 disables the check.
Value 1 enables the check.
PAX
dump
0 (default) or 1
Specifies whether to force dump format for backup.

Value 0 disables the dump format.
Value 1 forces the dump format.
PAX
filter.numFileFi
lter
050, 5 (default)
Specifies the number of file filters that can be specified for a single
backup.
Increasing the number of file filters increases the number of
different file types you can exclude from a backup.
PAX
filter.numDirFil
ter
050, 5 (default)
Specifies the number of directory filters that can be specified for a

single backup.
Increasing the number of directory filters increases the number of
different types of directories you can exclude from a backup.
PAX
file.caseSensiti
ve
0 or 1 (default)
Specifies whether the directory and file filters used to exclude

directories and files from a backup are case sensitive.
Value 1 enables case sensitivity on directory and file filters.
Value 0 disables case sensitivity on directory and file filters.
PAX
noFileStreams
0 (default) or 1
Specifies whether to backup/restore CIFS file streams.

Value 0 enables CIFS file streams
Value 1 disables CIFS file streams.
PAX
scanOnRestore
0 or 1 (default)
Specifies whether to virus scan after a file is restored.

Value 0 disables virus scan on restore
Value 1 enables virus scan on restore.
47
Known problems and limitations
Table 6
Facility
Parameter
Values
svtl
dbLocation
/svtl (default)
Specifies the directory that contains the SVTL database.
svtl
discardTapeData 0 (default) or 1
ufs
syncwatchdog
6 (default)
Specifies the maximum time (in milliseconds) required to flush the

UFS log. This value should include the time required for the
backend to correct itself. If the UFS log fails to progress within this
time limit, the Data Mover is forced to panic.
ufs
xlateToInline
0 or 1 (default)
Enables or disables xlate creation of inline directories. If set to

nonzero, MPDTranslate attempts to create inline directories.
Determines whether SVTL discards data before writing to virtual

tape. This parameter must be set before opening the tape to take
full effect.
Value 0 specifies that the SVTL writes to tape.
Value 1 specifies that the data is discarded. This setting allows you
to test the backup speed while eliminating the backend storage
speed as a performance factor.

CFS-14 hardware
restrictions
The following hardware restrictions apply:

Data Mover model
Celerra version supported
510
Version 4.2, 5.1 or later (version 5.0 not supported)
514
Version 5.1 or later
514T
Version 5.3 or later
Data Mover model 510 is supported in Celerra Network Server

versions 4.2, 5.1 or later, excluding version 5.0. Data Mover model
510 supports only Fast Ethernet, Gigabit Ethernet, and LP9002
Fibre Channel adapters (LP9002 FC adapters use LC connectors).
Data Mover model 514 is supported in Celerra Network Server

version 5.1 or later.
Data Mover model 514T is supported in Celerra Network Server

version 5.3 or later.
You can verify memory size by using the server_sysconfig

<movername>-Platform command.
48
FC4700-2 installation
When connecting the storage processors (SPs) to the Fibre Channel

switch, connect SPA-0 to a lower port number than SPB-0. This
ensures that the QLogic HBA board BIOS does not create a phantom
LUN 0 on SPB, which causes the FC4700-2 installation script to fail.
Also, if the Fibre Channel cables are moved after a successful install,
the Control Station may fail to boot.
Spanning disks
CAUTION
EMC does not recommend spanning a file system across multiple
storage systems.
Although the Celerra Network Server can support a heterogeneous
backend set, it is not recommended for file systems (including
checkpoint file systems) to span more than one storage system.
This is primarily due to the high-availability concern, as one storage
system could fail while the other continues, making failover difficult,
because the targets might not be consistent in this case. In addition, a
spanned file system is subject to any performance and feature set
differences between storage systems.
ConnectHome
Control LUN
requirements
enforced
Following are the limitations of the ConnectHome feature:
Following an upgrade, if the Modem Dial Tone Detection was

disabled, it must be reconfigured. For configuration instructions,
refer to Appendix B of the Setup Guide which is available on the
Documentation CD and on Powerlink.
The Serial Number must be manually entered for CNS, CFS and
older NS systems. Otherwise it's automatically detected.
The Email transport's sender address can only be configured

through the nas_connecthome command.
There is strict enforcement of the control LUN size and HLU numbers
for fresh installations. Control LUNS 0 and 1 must be at least 11 GB
and control LUNS 2, 3, 4, and 5 must be at least 2 GB.
In the past, if you had six control LUNs and they were not assigned
the correct HLUs, the installation may have succeeded. Now, an exact
match between the LUN capacity and the correct HLU value is
strictly enforced. Mismatches will generate a Control LUN Check
Failed error message during installation. You must correct the
control LUNs before installation can complete successfully.
49
Automatic File
System Extension
and virtual
provisioning
You should be aware of the following restrictions and limitations

when using Automatic File System Extension and virtual
provisioning:
Automatic File System Extension works only with UXFS, MGFS,

and FileMover-enabled file systems created with AVM storage
pools.
Automatic File System Extension is not supported on file systems

created with manual volume management.
Automatic File System Extension is not supported on file systems

used with TimeFinder NearCopy or FarCopy.
While Automatic File System Extension is running, the Control

Station blocks all other commands that apply to this file system.
Once auto extension is complete, the Control Station allows the
commands to complete.
The Control Station must be on and operating properly for

Automatic File System Extension, or any other Celerra feature, to
work correctly.
Automatic File System Extension cannot be used for any file

system that is part of a Remote Data Facility (RDF) configuration.
Do not use the nas_fs command with the -auto_extend
option for file systems associated with RDF configurations. Doing
so generates the error message: Error 4121: operation not
supported for file systems of type SRDF.
The options associated with Automatic File System Extension can

be modified only on file systems mounted read/write. If the file
system is mounted read-only, you must remount the file system
as read/write before modifying the high water mark, maximum
size, or virtual provisioning options.
Enabling virtual provisioning with Automatic File System

Extension does not automatically reserve the space from the
storage pool for that file system. Administrators must ensure
adequate storage space exists for the automatic extension
operation to succeed.
When there is not enough storage available to extend the file
system to the size requested, Automatic File System Extension
extends the file system to utilize all the available storage. For
example, if Automatic File System Extension requires 6 GB but
only 3 GB is available, the file system automatically extends to 3
GB.
50
Although the file system was partially extended, an error

message appears indicating there was not enough space available
to perform automatic extension. When there is no available
storage, Automatic File System Extension fails. You must
manually extend the file system to recover from this issue.
Automatic File System Extension and virtual provisioning are

supported with Celerra Replicator. Enable Automatic File
System Extension and virtual provisioning only on the source file
system in a replication scenario. The destination file system
synchronizes with the source file system; it extends automatically.
Do not enable Automatic File System Extension or virtual
provisioning on the destination file system.
When virtual provisioning is enabled, NFS, CIFS, and FTP clients

see the actual size of the Replicator destination file system, while
they see the virtually provisioned maximum size of the source file
system.
You can create replicated copies of extendible file systems, but to

do so EMC recommends that you use slice volumes (slice=y).
In the replication initial setup procedure, if Automatic File
System Extension occurs while full fs_copy is running without
replication service, the extension operation only extends the
source file system. In this case, when replication service starts, it
fails because the source file system and the destination file system
sizes are different.
Replication service must be running when the file system is
extended. The Using Celerra Replicator technical module contains
information about extending the destination file system manually
to keep file system sizes identical.
Virtual provisioning is supported on the primary file system, but

not supported with primary file system checkpoints. NFS, CIFS,
and FTP clients cannot see the virtually provisioned maximum
size of a SnapSure checkpoint file system.
You cannot create iSCSI dense LUNs on file systems with

Automatic File System Extension enabled. You cannot enable
Automatic File System Extension on a file system if there is a
storage mode iSCSI LUN present on the file system. Doing so
generates the error message: Error 2216: <fs_name>: item is
currently in use by iscsi. However, iSCSI virtually
provisioned LUNs are supported on file systems with Automatic
File System Extension enabled.
51
Automatic File System Extension is not supported on the root file

system of a Data Mover or a Virtual Data Mover (VDM).
In a configuration with multiple backend storage systems, an

Automatic File System Extension operation calculates the
extension size based on the storage available on all backends
rather than just the free storage on the backend where the file
system resides. If the backend where the file system resides has
less free storage available than the calculated extension size, the
extension operation fails and generates the error message: Feb 4
16:43:35 2006 CFS:3:101 fs auto extension failed:
/nas/sbin/rootnas_fs -x autox_svrvdm3_fs size=1024M -o
slice=y Error 3006: CLSTD : volume(s) are not
available, fs name=autox_svrvdm3_fs
(Slot3:1139089600:) This is only an example of a sys_log
message, some information may vary such as: fs name, size, disk
type, etc. EMC recommends creating your file systems with
adequate space to accommodate the estimated file system usage.
If the Control Station cannot adequately extend the file system to
accommodate the usage, the automatic extension operation fails.
Alternatively, you can manually extend the file system.
CAUTION
Automatic File System Extension is interrupted during Celerra
software upgrades. If Automatic File System Extension is enabled,
the Control Station continues to capture high water mark events,
but actual file system extension does not start until the Celerra
upgrade process completes.
The Managing Celerra Volumes and File Systems with Automatic Volume
Management technical module contains detailed information about
Automatic File System Extension.
Celerra iSCSI
Deleting snapshot of a
production LUN
File system extension
52
Avoid deleting the most recent snapshot of a production LUN while

I/O is active. I/O timeout problems can occur if the latest snapshot is
very large, possibly resulting in data corruption. As a workaround,
create a new snapshot and then delete the previous one.
If you have a file system with one or more iSCSI LUNs and the file
system is full or nearly full (less than 100 megabytes available),
extend the file system by at least 100 megabytes before upgrading
from version 5.4 to version 5.5. If you do not extend the file system,
data loss will occur.
iSCSI LUN extension
When using a virtually provisioned iSCSI LUN, monitoring file

system space is indispensable to ensure enough space is available.
CAUTION
Insufficient space on a virtually provisioned LUN will result in
data unavailability and might result in data loss. In addition to
monitoring file system usage, set an event notification for file
system full condition and enable automatic file system extension
with a conservative threshold. If enough disk space is not available
for file extension, data loss might still occur.
The automatic file system extension option supports virtually
provisioned iSCSI LUNs, but not regular iSCSI LUNs. You cannot
create a regular iSCSI LUN on a file system enabled for automatic
extension. Nor can you enable automatic file system extension if a
regular iSCSI LUN is present.
iSCSI snapshots for

Linux hosts
When creating a snapshot, the CBMCLI command issues the sync()

system call to flush pending I/O in the Linux buffer cache to the
iSCSI LUN storage. However, because this call does not stop I/O, it is
possible that not all data is written to the LUN, resulting in data
inconsistency. Before using such a snapshot for data restoration or
backup, you need to use a utility (such as fsck) to verify the data. This
behavior is identical for other back ends, such as CLARiiON or
NetApp, because it is a fundamental limitation of Linux.
To ensure data consistency in a snapshot, either quiesce the
application or unmount (umount) the file system for the iSCSI LUN
before creating the snapshot.
Rapid mounting and unmounting
A Linux issue, seen in all supported versions including Red Hat
Enterprise 3 and 4 and SUSE 9, can occur with frequent
mount/umount operations of an iSCSI LUN within a short time (less
than a minute). The symptom is apparent data corruption on the
iSCSI LUN on the Linux file system cache, although the backend data
is consistent. When this issue occurs, the following kernel error
message appears in /var/log/messages:
kernel: VFS: Busy inodes after unmount. Self-destruct in
5 seconds. Have a nice day...
When you see this message, unmount the iSCSI LUNs, reboot the
Linux host as soon as possible, and then mount the LUNs.
53
Microsoft iSCSI
Initiator support
EMC Celerra iSCSI supports the most current version of the

Microsoft iSCSI Software Initiator which is available as a free
download from Microsofts website (search for iSCSI initiator).
This section contains known issues with the Microsoft iSCSI Software
Initiator.
Initiator only uses first network portal of a target
An iSCSI target on the Celerra Network Server may have multiple
network portals configured for it; however, the Microsoft iSCSI
initiator GUI uses only the first network portal of a target.
Applies only to Microsoft iSCSI Software Initiator versions before 2.0.
Workaround: Use the Microsoft initiator CLI to access other portals.

Celerra VSS Provider
for iSCSI issues
This section contains information about patches you may need to

install to use VSS on your Windows Server 2003 machines.
Apply Microsoft hotfixes
To support VSS and the Celerra VSS Provider for iSCSI, you must
apply the following Microsoft hotfixes to your Windows iSCSI hosts:
Hotfix 887827 and 833167 - Volume Shadow Copy Service update.

Microsoft intends to include both of these hotfixes in Windows Server
2003 Service Pack 1.
Hotfix 883646 - Updated Storport Storage driver.
Hotfix 831112 - Cannot import transportable shadow volumes.
In addition, if you use shadow copies with Microsoft Exchange 2003,

you must apply hotfix 838864.
All of these hotfixes are available from Microsoft Product Support
Services. To obtain detailed information about these hotfixes, go to
the Microsoft Knowledge Base at http://support.microsoft.com and
search for the appropriate hotfix number.
Apply patches and hotfixes for VERITAS applications
Celerra VSS Provider for iSCSI has been tested with VERITAS Backup
Exec and VERITAS NetBackup. The following hotfixes and patches
should be applied to these applications:
54
For VERITAS Backup ExecThere is a problem with the

VERITAS Volume Snapshot Provider (VSP) and VSS that causes
volumes to be marked as read-only after creating shadow copies.
In order to avoid this problem, you must apply the VERITAS
hotfix 9.1.4691 VSP 1.04 - Advanced Open File Option.
You can find installation instructions for this fix on VERITAS
website. To find the fix, go to the VERITAS Knowledge Base at
http://support.veritas.com and search for document ID 272771.
For VERITAS NetBackupYou should apply Maintenance Pack 4

to NetBackup. The latest MP4 patch is available from
http://support.veritas.com.
Existing replications
involving file systems
used for iSCSI
Replications created prior to 5.5 in which the source file systems were
also file systems used to house iSCSI LUNs are no longer supported.
These file systems will continue to run in their current state but you
will not be able to actively manage them (e.g., suspend or resume
them). Any attempt to perform such an operation will report that the
"file system is already in use for iSCSI". The only allowed operation is
to abort/delete the replication. We recommend that you convert any
existing replications of this type to the new iSCSI Replication
available in 5.5 at your earliest convenience.
Miscellaneous host
connectivity issues
In addition to the issues contained in these release notes, you should

also check the following guides for other issues related to the
Microsoft initiator, QLOGIC HBAs, and compatibility with the
Windows operating system:
Host Connectivity Guide for Windows

Host Connectivity Guide for Linux
These guides are available from Powerlink

(http://Powerlink.EMC.com).
To locate these guides:
1. Log in to Powerlink.
2. Go to the following URL:
http://Powerlink.EMC.com/support/library/
3. From the Library Documents menu on the left, select Host

Connectivity.
4. Locate the host connectivity guides for Windows and Linux.
55
Celerra Manager
The following notes apply to Celerra Manager. In this release, EMC

requires the use of Internet Explorer 6.0 SP2, Internet Explorer 7.0, or
Netscape 6.2.3 and later.
Changes from other

user interfaces
Any changes made to the network server using another user interface
such as the Celerra CLI are not automatically shown in the Celerra
Manager. Refresh the browser to see these changes.
Checkpoint
scheduling
If you plan to use the automated checkpoint (task) scheduling feature

in Celerra Network Server version 5.5, and you have previously
created a checkpoint schedule in version 5.1 or version 5.2, the old
schedule format is automatically translated into the version 5.5
schedule format during the upgrade.
However, if the version 5.1 schedule (which was created with the
Celerra Native Manager) has irregular monthly refreshes scheduled
(for example, January, March, and August), or if you have created
schedules using version 4.2, these schedules do not automatically
translate upon upgrade, and you must re-create them using the
Checkpoint Schedules tab of the version 5.5 Celerra Manager
interface.
End on date
When creating an automated checkpoint schedule using the Celerra
Manager > Checkpoints > Schedule page, the online help states that
you can change the End On date after the schedule completes, and
the schedule begins to run again as scheduled. The ability to change
to End On date once the schedule completes is not available in V5.5.
Dealing with data

update problems on
pages
Because a Celerra Manager operation can involve several internal

steps and requires internal communication of statuses over a
network, you may not see data on a page updated immediately as
you expect.
If you do not see data updated on a page as you expect, try these
corrective steps:
1. In the same browser session, select and open another Celerra
Manager page, and then select and re-open the page on which
you want to see updated data. The updated data should appear
as you expect. If not, go to step 2.
2. Wait a short time (15 to 30 seconds), and then click the Celerra
Manager Refresh button one time to update the page. This
update may take significant time, because it refreshes all data on
the page.
56
Avoid clicking Refresh several times, because each click sends

another complete update request. If this does not seem to work,
go to step 3.
3. Check the Task Status page to see if the operation status is
recorded there. The progress of long-running operations and
operations on multiple objects (such as deletions of many file
systems) is recorded on the Task Status page. If you do not see
the operation listed on this page, go to step 4.
4. There may be problems preventing the operation from running.
Click the Celerra Network Server name in the navigation pane to
open the Celerra Home: Status page, and check the hardware
statuses and software alerts on the page for issues that could be
preventing the operation from running
Firefox 1.5
EMC requires the use of Firefox 1.5 or greater to run Celerra Manager.
Internet Explorer
browser
Extended use
Internet Explorer may crash or hang after extended use with Celerra
Manager. Extended use includes repeated cycles of navigating to
applet pages, which is the bulk of what you do with normal Celerra
Manager use. The problem appears to be related to a possible
memory leak issue reported for versions of the Java 1.4.2 plug-in.
These is no known fix for this problem. To work around it when you
encounter a crash or hang, log out of Celerra Manager if you can and
then log in again to the Celerra Network Server. If you cannot log out,
close or kill the browser session.
Log in to a second Celerra
Before you can log in to a second Celerra Network Server from a
Celerra Manager session through Internet Explorer, you must enable
acceptance of first-party and third-party cookies. To enable these
cookies in Internet Explorer, select Tools > Internet Options >
Privacy > Advanced, select Override automatic cookie handling,
and select Accept for both First-party Cookies and Third-party
Cookies.
Managing multiple Celerras
When managing multiple Celerra servers through Celerra Manager
and Internet Explorer 6, the browser might incorrectly block cookies,
causing the display of the Login screen after successful
authentication. If this problem occurs, view the privacy report for the
blocked Celerra, and find the entry for the cookie called "Ticket" from
the Login screen. Set the browser to always accept cookies from that
site and then log in again.
57
Java console
When using Celerra Manager, some versions of the Sun Java Runtime
Environment (JRE) may report the exception below in the Java
console. The exception is benign. It can be safely ignored.
java.lang.InternalError: obsolete interface used
at sun.java2d.NullSurfaceData.getRaster(Unknown Source)
at sun.java2d.loops.OpaqueCopyAnyToArgb.Blit(Unknown Source)
at sun.java2d.loops.GraphicsPrimitive.convertFrom(Unknown
Source)
at sun.java2d.loops.MaskBlit$General.MaskBlit(Unknown Source)
at sun.java2d.loops.Blit$GeneralMaskBlit.Blit(Unknown Source)
at sun.java2d.pipe.DrawImage.blitSurfaceData(Unknown Source)
at sun.java2d.pipe.DrawImage.renderSurfaceData(Unknown Source)
at sun.java2d.pipe.DrawImage.copyImage(Unknown Source)
... trace back continues
Length limits for file

system name and
custom mountpoint
path
Celerra Manager now limits the maximum length of a file system

name and custom mountpoint path to 240 characters. The interface
pages will not allow entry of a name or path longer than 240
characters. Online help pages that currently document a larger length
limit will be modified in the future to reflect this change.
Migration file systems
The 5.5 release includes possible corrective actions in a right-click

menu from most alerts displayed on the Celerra Manager home
status page. The list includes an alert if the usage of a migration file
system type exceeds the usage threshold. In this case, the alert
right-click actions that concern file systems (Extend File System,
Predict Full, File System Properties) will display the following error:
=================
Query file systems All:<mgfsName>. File system not found.
The file system name listed cannot be found.
=================
The File System List action will take the user to the File System list
page, but the migration file system will not be displayed there as
migration file systems are managed separately on the File System tab
of the Data Migrations page. The Properties and Extend actions are
available on this tab, but the Predict Full action is not available for
migration file systems.
Netscape with Celerra
Manager
Enabling addition of Celerras to a Celerra Manager session

When using Netscape, you may be unable to add or access Celerra
Network Servers in a Celerra Manager session. Instead of a login
pop-up for the server, you will see a pop-up with this error message:
<Celerra_name_or_IP_address> received a message with
incorrect Message Authentication code. If the error
occurs frequently, contact the website administrator.
58
To correct this problem, make the following changes:

1. On each client on which you see the problem using Netscape:
a. Open Netscape and go to Edit > Preferences > Privacy and
Security > Certificates.
b. Select Manage Certificates.
c. Select the Authorities tab and clear all certificates for Celerra
Network Servers and any certificates that show up as
<localhost.localdomain>.
2. On each Celerra Network Server you want to access using
Netscape:
a. Log in or su to root.
b. Run the command nas_config -ssl and answer y (yes) to this
prompt:
Installing a new SSL certificate requires
restarting the Apache web server.
Do you want to proceed? [y/n]:
These changes clear out and regenerate all SSL certificates and should
resolve the problem.
Log out when using Netscape
If you are using Netscape, simply closing a browser window when
you are finished may not clear authentication information
maintained in Netscape's cache of non persistent cookies. If any other
browser windows are open, the non persistent cache is not cleared.
Directing a subsequent window or a previously existing window
back to the Control Station will not require reauthentication. This is
because you did not terminate your session on the Control Station
and Netscape has retained your authentication token.
To ensure that your Web UI session with the Control Station is
terminated, click the Logout icon to explicitly log out of the Celerra
Manager.
Managing multiple Celerras
When managing multiple Celerra servers through Celerra Manager
and Netscape, if the browser displays an alert about an invalid
certificate, you should clear the certificate information for all Celerra
servers from the browser and from the Java plug-in control panel.
Close all browsers. Regenerate certificates on the Celerra servers by
running (as root) the command /nas/sbin/nas_config -ssl
and restarting Apache.
59
In general with all browsers, you should avoid using

localhost.localdomain certificates, and specify the complete host and
domain name. When connecting to a host, use the same name as the
certificate. Note that performing an upgrade can reset certificates
back to localhost.localdomain.
Navigation
There is a known issue when using Netscape 6.2 with Celerra
Manager. When navigating, you may not be able to click on any text
entry fields to type. The workaround is to click on a non-Netscape
window on your desktop, and then click on the Netscape window.
The recommended solution is to upgrade to Netscape 7.0 or later.
Password remember
If password remember feature is enabled on Netscape, the default
passwords will be filled in password fields on the CIFS server, CHAP
new screens, and so on (which ever page that has a password field).
When using Celerra Manager with Netscape, be sure to configure the
browser so that it does not remember passwords. Otherwise, the
browser might supply default passwords into password fields, which
could be saved unintentionally.
Preserving list of alerts
during a CFS Software
upgrade
The list of alerts displayed on the Celerra Home > Status tab
are deleted when the CFS software is upgraded. To preserve
the list of alerts, use this procedure.
1. Log in to the Celerra Control Station.
2. Select Celerras > [Celerra_name] > Tools in the navigation
pane to display the Celerra Tools page.
3. Select the SSH Shell. Log in to the SSH login window as
nasadmin.
4. In the SSH Shell window, type the following commands to
copy the alert_log files to the $NAS_DB/log directory:
$ cd $NAS_DB/log/webui
$ cp alert_log* ..
Upgrade the Celerra software.
5. Log in to the Celerra Control Station.

6. Select Celerras > [Celerra_name] > Tools in the navigation
pane to display the Celerra Tools page.
7. Select the SSH Shell. Login to the SSH login window as
nasadmin.
60
8. In the SSH Shell window, type the following commands to

move the files back into the $NAS_DB/log/webui directory.
$ cd $NAS_DB/log/webui
$ mv ../alert_log* .
Note: You may not see the updated Celerra status until the Status page and
Navigation pane are refreshed.
Required Java
Runtime Environment
EMC strongly recommends the use of Java Runtime Environment

(JRE) 1.4.2_06 or a higher 1.4.2 revision on client machines that will
run Celerra Manager. You can get an installable version of JRE 1.4.2
from the Sun Microsystems website at
http://java.sun.com/j2se/1.4.2/download.html or from the Celerra
Applications and Tools CD.
The use of Celerra Monitor, provided with Celerra Manager Advanced Edition, requires installation on the client computer of the
Java Web Start package. This package is part of JRE 1.4.2 revisions.
You can use JRE 1.4.0 or 1.4.1 with Celerra Manager, but you may see
performance issues with these earlier revisions.
A bug in version 1.4.2 of the Java Runtime Environment (JRE)

sometimes causes the browser to crash when switching between
objects in the navigation pane in Celerra Manager or when clicking
rapidly without giving the browser time to respond.
Required Web browsers for Celerra Manager use
Celerra Manager has been tested with the Internet Explorer and
Netscape browsers.
secmap
Internet Explorer 6.0 is the minimum version required to run

Celerra Manager. EMC recommends using Internet Explorer 7.0.
Netscape 6.2.3 or a higher 6.x or 7.x revision is required to run

Celerra Manager. For the best Netscape experience with Celerra
Manager, EMC recommends using Netscape 7.1 or a higher 7.x
revision.
The server_cifssupport server_x -secmap -import -file

command doesn't import anything in the secmap database. The
command fails with the error message :
Error 5005 : server_3 : Unexpected MAPPING_RECORD element
in Import request.
Import records are read from the specified import file, but are not
sent to the Celerra. Only empty record elements are sent.
61
This operation does not change the secmap database. After the
import, the displayed secmap list is empty. This will be fixed in a
future release.
Selecting "Shell
Access" after session
timeout
When a Celerra Manager session times out, you are prompted to re

authenticate when you attempt to go to any page. Since Shell Access
provides only the applet window without a Celerra Manager screen,
you do not see the login/password prompts if you first select "Shell
Access" in the navigation pane. In this case, Shell Access does not
proceed because it is waiting for you to re authenticate your Celerra
Manager session. To proceed, either refresh your browser window or
select any other item in the navigation pane to re authenticate and
then select "Shell Access."
Storage capacity for

new and extended file
systems
When you create a new file system or extend an existing file system,
you enter the desired storage capacity in megabytes. However, the
available storage for the selected storage profile is displayed in
megabytes (MB), gigabytes (GB), or terabytes (TB).
To convert gigabytes to megabytes, multiply gigabytes by 1024. To
convert terabytes to gigabytes, multiply terabytes by 1024. For
example, if the selected storage profile has 450 GB available, and you
want to create a file system that uses all of the available space, enter
460800 in the storage capacity text box.
Other browser issues
When adding another Celerra Network Server to the Celerra

Manager display, specify the new Celerra by its fully qualified
domain name, not its IP address; otherwise an error message about
blocked cookies might appear. In addition, the IP address of the
Celerra should be added to the list of trusted sites. If you receive the
error message, double-click the blocked cookie icon and indicate in
the dialog box that cookies should be accepted from that site in the
future.
Outdated (stale) information in the browser cache can sometimes
cause unexpected behavior, such as the appearance of extra
confirmation screens. Clear the browsers cache and close the
browser, then try again.
On systems with a very large configuration (such as a large number
of Data Movers and/or file systems), JServer may not start up or may
start after a very long delay.
62
Celerra Monitor
Java Runtime
Environment
If you are unable to launch the Celerra Monitor from the Celerra
Manager, make sure that the appropriate Java Runtime Environment
(JRE) has been installed on the management workstation (refer to the
Installing Celerra Management Applications technical module for
details) and that the Control Station hostname and network setting
for the system you want to monitor have been set using the Celerra
Manager.
To install and launch Celerra Monitor without Celerra Manager,
perform the following steps:
1. If JRE1.4.0 or higher is not installed, install it (see Installing Celerra
Management Applications technical module for details)
2. In your browser (Netscape or IE), type the following URL:
http://Control_station_address/cmv2/cmv2.jnlp
CLARiiON backend
CLARiiON upgrades
Occasionally, CLARiiON storage connected to a Celerra Network

Server requires a software upgrade. When upgrading CLARiiON
software, do not run heavy loads on the Celerra, as operations may be
extremely slow or interrupted.
Notes
CLARiiON array software cannot be upgraded unless the Celerra

Control Station software is upgraded first. Ignoring this rule may
make the CLARiiON unmanageable by the Celerra.
Incremental CLARiiON arrays cannot be added to systems

running older Celerra software unless they are running a
compatible release of CLARiiON software.
A warning will appear when you upgrade the Celerra software if

the CLARiiON software needs to be updated. If you receive this
warning, finish upgrading the Celerra software and then upgrade
the CLARiiON software as soon as possible to avoid
compatibility issues.
Consult the E-Lab Interoperability Navigator at

http://Powerlink.EMC.com for all software interdependencies.
63
Configuring iSCSI
ports
While running MPFS Configuration Manager (MCM) to configure

iSCSI ports on the backend, the security settings need to be correct.
As of release 5.5.30.4, the default username and password for MCM is
now different than the default security settings that the NAS install
creates when initializing the Global Security on the CLARiiON. MCM
uses emc/emc for username and password. The Global Security
password is nasadmin/nasadmin. This causes MCM to not connect
and therefore iSCSI configuration cannot be done.
Workaround:
Create a user/password for Global Security on the CLARiiON that is
the same as the default username/password for MCM via the
following procedure:
1. Login into the Control Station as nasadmin and su to root.
2. Run /nas/sbin/naviseccli -h spa -User <default username:
nasadmin> -Password <default password: nasadmin> -Scope 0
security -adduser emc -password emc -scope 0 -role
administrator.
Refer to Primus Solution emc167671 for more detailed information on
this MCM issue.
Reserved disks
EMC reserves LUN numbers 0-15 (from the Data Mover perspective)
on the first storage system configured on the Celerra (the Celerra boot
storage). If a CLARiiON was misconfigured prior to 5.3 it was
possible for the administrator to create volumes/file systems on these
disks. Starting with NAS V5.5, you will be prevented from allocating
volumes/file systems for user data on these disks.
Any attempt to allocate this storage will result in Error 3147:
volume <disk-id> resides on a reserved disk. If a
reserved disk is already used in an AVM storage pool, no additional
storage will be allocated from the volume built on the reserved disk.
This restriction affects only new space allocation. Existing file
systems will continue to function normally. If you see this error,
contact your EMC Customer Support Representative so corrective
action can be taken.
Storage system status

properties
64
On Celerra systems with CLARiiON backend storage where one of

the Storage Processors is down, some of the storage system status
properties displayed by nas_storage -info for that storage
system may be stale. In particular, the write cache state may display
"Enabled" when in fact it would be "Disabled" without both storage
processors being active. Other properties such as failed_over and any
status may also be inaccurate.
Control Station
NS Integrated to
Gateway upgrades
fail
When performing an Integrated to Gateway conversion on any NS

platform running NAS version 5.5.x, the upgrade hangs. The
following message is continuously repeated on the screen: Move
CLARiiON to public network and change the /etc/hosts
entries on Control Station to reflect the new IP
addresses.
A fix is currently in development. In the meantime refer to Primus

emc155369 for more detailed information and updates.
Non-ASCII characters
The Control Station cannot handle non-ASCII characters for shares,

NetBIOS names, domain names, and passwords. When creating these
names and when issuing commands to the Control Station, be sure to
use only alphanumeric ASCII characters.
Reboot may be
required after initial
install
NS700 software is typically installed at the factory, but in some cases,

a fresh initial installation may be required at the customer site by an
EMC Customer Support Representative. In these cases, the
representative will follow the factory install procedure. Once the
installation procedure is completed, the Control Station must be
rebooted. Refer to the appropriate setup guide for the configuration
you are installing.
Security and
authentication
To ensure security, change the default Control Station root and

nasadmin user passwords before configuring network connections.
The default password for both root and nasadmin is nasadmin.
To change the password, using Celerra Manager:
1. Log in as root (or nasadmin).
2. Select Control Stations from the navigation pane.
3. Click Change Password next to the Username field.
4. Write down your changes and store them in a secure location.
Repeat this procedure for nasadmin username.
FLARE operating
environment
The backend storage requirements check during upgrade may fail

because the version found on an array is 0.0 and Is_Flare_Commited
is False, as follows:
#./setup
Setting up system, please wait...
Backend Storage Requirements Check Failed:
65
REQUIREMENT: CX700_Software >= 2.06 AND CX700_Software <

2.08
ID: APM00034000062 FOUND: 0.0 STATUS: FAILED
REQUIREMENT: Is_Flare_Committed = True
ID: APM00034000062 FOUND: False STATUS: FAILED
Warning: System does not meet requirements for upgrade.
Do you wish to continue with this installation? [yes or
no]:
This may be because the array software was installed individually

instead of installing the FLARE Operating Environment package.
Use the following command to list all the software packages on the
array: /nas/sbin/navicli -h A_[ID from Failure] ndu -list
for example:
$ /nas/sbin/navicli -h A_APM00034000062 ndu -list
Name of the software package: Navisphere
Revision of the software package: 6.6.0.2.44
Commit Required: NO
Revert Possible: NO
Active State: NO
Is installation completed: YES
Is this System Software: NO
...
If none of the packages list "Base" or "FLARE Operating

Environment" after the field "Name of the software package", inform
the storage administrator that the arrays should be installed with the
"FLARE Operating Environment" package instead of with individual
packages.
This warning will not affect the Celerra upgrade. The upgrade can be
continued, and the array software updated at a later date.
New installations of
5.5.31.6 on CX arrays
running FLARE 19 fail
Celerra Network Server versions 5.5.31.6 and later contain a new

version of naviseccli (Revision 6.26) which is not
backward-compatible with its older version on FLARE 19 CX
backend. This issue causes fresh installations of version 5.5.31.6 or
later to fail.
To use naviseccli, the CLARiiON must be running the following
patch releases (minimum) or any release of R24 or R26:
66
2.22.xxx.5.505 or 3.22.xxx.5.505.
2.19.xxx.5.034
File-Level Retention

when using File-Level Retention on Celerra:
File-Level Retention does not support NFSv4 protocol.
File-level Retention prevents FLR files from being modified or

deleted by NAS clients and users. This version of File-Level
Retention is intended for self-regulated archiving. The Celerra
administrator is the only trusted user of the file system with
File-Level Retention enabled.
This version of File-Level Retention is not intended for high-end
compliance applications such as pharmaceuticals, aerospace, or finance.
FileMover
It is strongly recommended the antivirus (AV) administrator

update the virus definition files on all resident AV engines in the
CAVA pools, and periodically run a full file system scan of the file
system to detect infected FLR files. The Using Celerra AntiVirus
Agent technical module provides additional information about
virus definition updates.
It is not possible to enable the Celerra FileMover functionality on

a file system with File-Level Retention enabled. Hence, it is not
possible to use a file system with File-Level Retention enabled as
primary storage in a FileMover environment. However, it is
possible to create a file system with File-Level Retention enabled
as secondary storage.
Although File-Level Retention supports all backup functionality,

the FLR attribute is not preserved in the NDMP backup.
Therefore, the Celerra administrator must be sure the file system
is restored to a Celerra file system with File-Level Retention
enabled.

when using Celerra FileMover:
FileMover cannot be enabled on a read-only file system not

previously enabled for FileMover functionality.
A File-Level Retention (FLR) capable file system cannot be used
as primary storage in a FileMover environment. However, it is
possible to use an FLR file system as secondary storage. Contact
your software vendor to determine whether your policy and
migration software supports this configuration.
67
A file in the FLR state on secondary storage cannot be altered or

deleted by NAS clients or users. A file in the FLR state can be
deleted only after its retention period expires. The Using File-Level
Retention on Celerra technical module contains information about
FLR storage technology and file system behavior.
Celerra FileMover cannot be enabled on an MGFS file system.
Celerra FileMover can manage a maximum path length of 1024

bytes; however, the maximum pathname length depends on the
policy and migration software.
Celerra FileMover does not support migration of alternate data

streams.
Celerra FileMover does not support NFSv4 protocol.
Celerra FileMover does not support using a DFS-based UNC path

to connect to CIFS based secondary storage. The Celerra must
connect to the real server and share name when using CIFS
connections to secondary storage.
To avoid issues when accessing files migrated from primary

storage multiple times, be sure your chosen policy and migration
software supports maintaining multiple versions of files on
secondary storage.
A multihop configuration allows you to move data from primary
to secondary storage, from secondary to tertiary storage, and
back. A FileMover-enabled file system can be secondary storage.
In a multihop configuration, secondary storage must be the
Celerra Network Server while tertiary storage can be the Celerra
Network Server NFS/CIFS file server.
Although FileMover supports a multihop configuration up to two
hops, your policy engine vendor might not yet support this
functionality. Contact your policy engine software vendor
directly for the most up-to-date information about multihop
configurations.
68
If you want to use a Windows Microsoft Cluster Service (MSWC)

based file server as secondary storage in the FileMover
environment, you need Celerra Network Server version 5.4.20.0,
5.5, or later.
MMC snap-in
Data Mover security
settings snap-in
The Data Mover security settings snap-in is installed as a Microsoft

Management Console (MMC) snap-in into the Celerra Management
Console on a Windows 2000, Window Server 2003, or Windows XP
system.
Problem: The Group Policy Object (GPO) applied to the CIFS server is
not reflected in the Data Mover security settings.
Workaround: If reverse lookup zones are defined in DNS, the DNS
name and DNS domain returned by a reverse lookup on your CIFS
servers IP address must be the same name as your CIFS server.
For example, assuming the name of the CIFS server is gpo, its DNS
domain is celerra.emc.com and its IP address is 192.168.15.81, the
following results should occur:
MPD translation of
read-only file
systems
The forward lookup on the name gpo and

gpo.celerra.emc.com should result in the IP address
192.168.15.81.
The reverse lookup on the IP address 192.168.15.81 should result

in the DNS name gpo.celerra.emc.com.
If no reverse lookup DNS record is configured for the CIFS

servers IP address, the snap-in should operate normally and no
action is required.
For each of the following MPD examples, later read-write mounts of

read-only file systems will not automatically translate them.
Although not required, these file systems can be manually translated
by following the procedure documented in the Using Multi-Protocol
Directories with Celerra version 5.5 technical module (starting on page
13), available on Powerlink.
Automatic translation of directories from non-MPD (compat) to

MPD (DIR3) format cannot occur, or even be scheduled.
Non-MPD directories are left in a readable but obsolete format.
Inability to schedule MPD Translation at upgrade time may result
in an upgrade error message. This error message is informational
only, and does not impact the upgrade.
Root fs, if read-only, will not be translated, but operates properly

in compat mode.
69
Multi-Path File
System (MPFS)
formerly HighRoad
If an un-translated checkpoint is ever restored (disaster recovery

for instance), it does not automatically get translated to MPD
format, but operates properly in compat mode. If the PFS is
already translated, more recent checkpoints taken from it are
already in the MPD format. New file systems are created in MPD
format, so there is no translation required.
MPFS currently does not support more than 4,000 devices in a single
Symmetrix. This implies that:
An MPFS file system cannot reside on any disk volume above

4000
and
An MPFS client cannot have any disk volume mapped that is

above 4000.
For MPFS, EMC recommends a 256 KB Celerra stripe size. This has
proven (on average) to achieve the best performance for Windows
and UNIX MPFS clients.
MPFS-supported NAS
features
Table 7
MPFS-supported NAS features
MPFS
NonMPFS
Celerra AntiVirus Agent (CAVA)
Yes
Yes
Provides virus checking of files written with MPFS Windows clients.
Celerra Data Migration Service

(CDMS)
Yes
Yes
After the system is migrated, it can be used as an MPFS file system.

During the migration, the data is called using standard NAS protocol,
even if it is mounted as MPFS on the client side.
Celerra Replicator
Yes
Yes
Support on source file system only.
CIFS mandatory locking
No
Yes
File-Level Retention
Yes
Yes
FileMover
Yes
Yes
All the files stored on primary storage are retrieved through MPFSi. All
files stored on secondary storage fall through to normal NAS protocols,
either NFS or CIFS.
LAN Backup
Yes
Yes
Fall through to normal NAS protocols also occurs during a LAN backup
if the backup application is not FileMover-aware and data has been
migrated to the secondary storage.
NDMP
Yes
Yes
Product or feature
70
The following table lists Celerra Network Server features supported

with MPFS-enabled Data Movers.
Note
Table 7
MPFS-supported NAS features (continued)
MPFS
NonMPFS
Rainfinity
No
Yes
SnapSure
Yes
Yes
Symmetrix Remote Data

Facility/Asynchronous (SRDF/A)
Yes
Yes
Symmetrix Remote Data

Facility/Synchronous (SRDF/S)
Yes
Yes
TimeFinder/FS
Yes
Yes
User, Group, and Tree Quotas
Yes
Yes
Product or feature
Note
Only production file systems (PFS) can be used with MPFS/i. MPFS/i is
not used when accessing ckpt-type file systems.The system
administrator is not notified when the restore is complete. The restore
completion status can be checked in the server_log file and can also be
configured using the SVFS facility of nas_event.
SRDF is active on the primary site only. After an SRDF failover, the
MPFS client is able to access the secondary sites file systems over
NFS or CIFS.
When a quota limit is reached, all traffic falls back to standard NAS.
NDMP
CAUTION
Not all NDMP vendor versions support this version of Celerra
Network Server. For the latest NDMP Vendors/Versions/Patches
and restrictions supported in this version of Celerra Network
Server, refer to the Backup section of the E-Lab Interoperability
Navigator.
Backups in ASCII
mode
If a Data Mover is in ASCII mode and you back up a file system with
asciifilter off, the file name with international characters cannot be
recovered if the asciifilter it turned on during the restore.
More information about asciifilter can be found in the shadow facility
section of the Parameters Guide.
Backups with
Integrated
Checkpoints feature
By design in V5.5, the Backups with Integrated Checkpoints feature

automatically creates a checkpoint, uses it for a specified backup
operation, and then deletes it. (The checkpoint deletion occurs
whether or not the backup is successful.) However, in the event that
the checkpoint does not auto-delete, you should manually delete the
checkpoint. Note: EMC recommends using the Celerra Manager to
select and delete the specially-named checkpoint, due to its long
name length.
71
If you use the Backups with Integrated Checkpoints feature in the

backup process, and there is insufficient storage space available in the
original pool to create and maintain the checkpoint, the checkpoint
becomes inactive and the backup operation fails. Before using this
feature, ensure sufficient pool space exists to sustain the checkpoint
during the backup operation.
EMC Data Manager
The default value for parameter readWriteBlockSizeInKB must be

changed to 128 when using EMC Data Manager for backup and
restore operations.
Legato backup
software
When configuring NDMP backups using Legato backup software, if

the variable DIRECT=y is specified, the Legato indexed restore does
not recover directory permissions or backup date/time. (This is now
the default behavior in V5.5.)
To restore permissions and backup date/time for directories, proceed
as follows:
On the Legato backup host (UNIX/Linux), in .profile or .login,
depending on shell, add:
NSR_NDMP_RECOVER_DIR=yes
export NSR_NDMP_RECOVER_DIR
On the Legato backup host (Windows):

Click My Computer>Properties, and then click the Advanced tab.
Click Environment Variable, and then enter the environment
variable, as shown below, and click OK. Then click OK again
.
72
nas_fsck -aclchk
command
rawfs to uxfs
conversion
Security scan software
Symbolic links
Tuning parameters
The NDMP backup process must be stopped on the Data Mover

before you can use the nas_fsck -aclchk command.
You cannot convert a file system created as rawfs to a uxfs type unless
you first perform a full destructive restore to the rawfs.
If you are running an IP security scan software package, such as
IP360 or similar from NCircle, you need to exclude the NDMP port
10,000 from the scanning software.
NDMP does not follow symbolic links among file systems. Therefore,
an NDMP backup does not include a symbolic link's target file
system. For more information on symbolic links, refer to the
Managing Celerra for a Multi-Protocol Environment technical module.
The maximum values of the following PAX parameters have been
adjusted to make better use of system resources without negatively
impacting performance:
PAX.nRestore was reduced from 64 to 16;
PAX.nPrefetch was reduced from 32 to 8;
PAX.paxWriteBuff was reduced from 1024 to 256;
PAX.paxReadBuff was reduced from 512 to 256;
PAX.paxStatBuff was reduced from 2048 to 512.
During an upgrade to version 5.5.27.5 or later, when Data Movers are
rebooted, existing parameters with higher values than the new
maximum are rejected and the default values are applied. An attempt
to set these parameters to a higher value than the new maximum will
fail, and an error message will be recorded in the server_log.
Refer to the Celerra Network Parameters Guide, available on Powerlink,
for more detailed information about these parameters.
Using IBM Tivoli and

NDMP Backups to Disk
When using IBM Tivoli and NDMP Backups to Disk (NDMPB2D)

with Celerra Network Server Version 5. 5, note the following:
VTLU (Virtual Tape

Library Unit)
When creating new storage, always specify barcode prefix in

uppercase
Celerra VTLU is supported for NDMP backup of Celerra File Systems

to the Celerra VTLU only. The Celerra VTLU cannot be used as a
target for non-Celerra host backup.
73
NFS
Cached files
If a client mounts exported parent and child paths of the same file
system, it is possible to access the same file from two different mount
points. If client side caching is enabled, the client may end up
maintaining two different versions of a cached file.
Access Control Lists

(ACLs)
If ACLs have been set on a file accessed with Linux over NFSv4, they
are displayed correctly as they were set ("getfacl") from a Linux or
Solaris machine. If the same file is queried from Windows (via the
security tab of the file properties through Explorer), you will receive a
security message indicating the permissions are incorrectly ordered
and offering to sort the permissions correctly. If you choose to allow
Windows Explorer to sort the ACEs, Linux and Solaris may not be
able to display the ACL afterwards.
NFSv4
This is merely an error in reporting and does not represent an exposure in

terms of how access control to file-system objects is performed and the actual
access permissions, based on the original settings of the ACL, are applied
normally.
Celerra Data
Migration System
Access to online migration file systems while using CDMS is not

supported by NFSv4 clients until migration completes.
When NFSv4 is used, it triggers directory migration of sub directories
and requires global lock for the entire Data Mover. Therefore, the
Data Mover will be inaccessible during the migration.
Delegations
Under rare circumstances, a race condition can occur that can block
access to a single file. This issue only occurs in NFSv4 environments
with delegation enabled. If the condition occurs, the workaround is to
disable delegation in NFSv4.
To learn how to disable delegations, refer to the Configuring NFS on
Celerra technical module, P/N 300-002-693, rev A01, which is
available on the Celerra Network Server documentation CD, or on
Powerlink.
NS700
Incorrect Control
Station setup causes
communication error
messages
74
If the Control Station is not correctly configured, the NS700 is unable

to connect to its storage system. The error message: Array software
on <ip_address> not responding, keep trying... is written to
the Control Station log file and displayed on the Celerra Manager
main page.
Verify that the Control Station network configuration is correct. Using

Celerra Manager, select Celerra, Control Station. On the Control
Station Properties page, verify the following values: Hostname, IP
Address, Netmask, Gateway, DNS Domain Name, DNS Domain
Search Order, and DNS Servers.
Be sure the default Hostname emcnas_i0 has been replaced with the
actual hostname of your Control Station. Also, be sure to use
fully-qualified names in the DNS Domain Name and DNS Domain
Search Order text boxes.
Speed settings
When configuring a Fabric-Connected NS700G with CLARiiON, all

Fibre Switch ports must be set to the highest speed setting (1 Gb or 2
Gb) supported by the switch. The Data Mover will auto negotiate its
Fibre port speed to match the switch. Switch ports should not be set
to auto negotiate.
Quotas
Creating and
modifying explicit
user/group quotas
When you create a new, explicit user/group quota, or modify an

existing user/group quota, the quota will not appear in the quota list
if the user/group has no usage on the disk and you set the quotas as
follows:
Storage hard limit: 0
Storage soft limit: 0
File count hard limit: (blank)
File count soft limit: (blank)
Note: In the CLI, the limits are entitled "block" (not storage), and inode (not
file count), but the same behavior applies.
When using the Celerra Manager to modify a user/group quota

under the above conditions, and clicking "Apply", a non manageable
page appears. Click "Back" to return to the previous page in the
Manager.
Restrictions for listing
user/group/tree
quotas
Due to memory restrictions, the following behavior may result when

listing file system quotas on the Celerra Network Server:
The Celerra Manager's File System Quotas > Users and File
System Quotas > Group pages may not display in a timely
manner when a large number of users and groups is associated
with a file system. Lists in excess of 10,000 users and groups may
time out due to an out of memory condition.
75
The Control Station's nas_quotas -tree -list command takes

considerable time, or times out, when attempting to produce a list of
more than 2,000 quota trees. Similarly, the Celerra Manager's File
System Quotas > Tree page may not display in a timely manner
when a large number (typically more than 2,000) quota trees is
associated with a file system.
Soft quota limits
Soft quotas are always set lower than hard quotas. When users try to
store files that exceed the hard quota limit and the deny disk space
flag is set to yes, the Data Mover returns an error and does not store
the file. The behavior of soft quotas depends on the type of client. For
example, Microsoft Windows (CIFS) users receive a pop-up warning
message when they exceed a soft quota (if configured). UNIX users
never receive a message when they exceed a soft quota (only the hard
quota). In both cases, a server_log message is generated. All users can
exceed the soft quota for a configurable grace period before they are
blocked from saving files.
Replication
Replication sessions
Remote replications
Before running remote replication sessions, make sure both systems
are running the same Celerra code family. For example, you cannot
have version 5.4 on the source and version 5.5 on the destination, or
visa-versa.
Replication restart
Under rare circumstances, following a Data Mover reboot, Replicator
sessions may go inactive and need to be restarted. If this occurs,
restart the Replicator sessions manually.
VDM displays status of

"temporarily
unloaded"
When replicating CIFS environments, a VDM being configured for

replication (as the destination) displays a status of "temporarily
unloaded" from the time that it is created/mounted until replication
begins. The VDM displays as temporarily unloaded because the rootfs
is a rawfs file system type and only after it is converted to uxfs will the
VDM state change to "mounted."
SnapSure
Checkpoint name
76
While using SnapSure, do not use the extension .ckpt to name a

checkpoint file system. This extension (.ckpt) is used as a virtual
directory entry to enable NFS or CIFS client access to online
checkpoints in the production file systems namespace, and it can be
used for this purpose only.
SnapSure allows the same checkpoint name to be used when creating

different checkpoint schedules of the same PFS and of different PFSs.
No alert or error message is provided. Avoid using the same
checkpoint names when creating multiple schedules. Using
SnapSure's default checkpoint names avoids this problem.
Checkpoint
scheduling
EMC recommends that you not schedule checkpoint creation/

refreshes of the same PFS to occur within 15 minutes of each other.
This includes checkpoints in the same schedule or between schedules
running on the same PFS. Always specify schedule checkpoint times
that are more than 15 minutes apart when scheduling checkpoints
using the Celerra Manager. (The 15-minute restriction does not apply
to creatingrefreshing checkpoints using the command line interface).
Data migration
Do not perform data migration tasks (using a feature such as the

Celerra Data Migration Service) while SnapSure is active. Since
migration activity produces significant changes in the data
environment, it is best to complete migration tasks before using
SnapSure to create and manage checkpoints. This avoids consuming
SnapSure resources to capture information that is unnecessary to
checkpoint.
File system
environment
Parameter file
All parts of a production file system (PFS), including all parts of its
associated checkpoint (SavVol), must reside on the same storage
array.
Do not insert a new-line character at the end of a nas_param file
(/nas/site/ or /nas/sys/). The empty line causes fs_ckpt to
return an encodeable.extract error.
SRDF
Additional upgrade
guidance based on
new SRDF behavior
As documented in the Using SRDF/S with Celerra for Disaster

Recovery and Using SRDF/A with Celerra technical modules for this
release, SRDF now employs a lock mechanism during the restore
process. With the addition of this lock mechanism, it is now strongly
recommended that you upgrade the source Celerra Network Server
first.
To avoid a prolonged outage and potential exposure to a disaster
scenario, limit the amount of time between your source and
subsequent destination upgrades. In such a disaster scenario, you can
activate a failover, but a restore would require intervention by an
EMC Customer Service Representative.
77
Task Scheduler
The Task Scheduler in version 5.3 runs in a separate thread, but in the
same process domain as the Task Manager. The easiest way to
determine if the scheduler is running is to check the number of
threads in the apl_task_mgr daemon. In NFS version 5.3 and later, the
apl_task_mgr daemon should have 4 threads running. The Linux
command ps axfw | less will show the list of processes in a tree
form with child processes hanging off of their parent processes.
Usermapper
Use of iPLanet for user mapping in a heterogeneous (CIFS and NFS)

environment is not recommended. Other user mapping methods,
such as NIS, provide the same capability, with better performance.
Note that in a CIFS only environment, no external mapping
mechanism is necessary, and the internal Usermapper solves any
CIFS mapping needs.
Other restrictions
CAUTIONS
78
Do not install Linux non-Celerra applications on the Control

Station. Doing so consumes resources and might adversely
affect system operation. Celerra Control Station is a dedicated
controller and not an applications platform.
Do not manually edit the nas_db database without consulting
EMC Customer Service. Any changes to this file could cause
problems with your Celerra installation.
cifssyncwrite
The cifssyncwrite option was implemented to the

server_mount command to enhance support for CIFS on the
Celerra. This special mount option guarantees that any write to the
file server is done synchronously. For Windows NT, it is important to
make sure this option is specified if the Celerra will be used to store
certain database files. This is recommended to avoid chances of data
loss or file corruption across various failure scenarios, for example,
loss of power.
Command line scripts
The NAS database stores specific information required for each Data
Mover. The Celerra Network Server automatically performs a backup
of the entire database every hour and saves it to a file named
nasdb_backup.1.tar.gz. These backups start one minute after
the hour. EMC recommends that you check your scripts to make sure
they are not scheduled to start at the same time as the backups.
During backup, the database becomes locked or inaccessible, and
some commands that rely on the database might fail.
Data Migration using

Robo Copy
Permission inheritance did not work as expected after migrating data

to a Celerra file server using Robo Copy. The migration was done via
Robo Copy in two steps using the following command lines:
robocopy source target /b /e /copy:d // to copy the data
only
robocopy source target /b /e /copy:tsa /IS // this is
copying time stamps, attributes and DACL
To avoid this problem, migrate data and permissions in one step.

Problem with Internet
Explorer back button
When clicking between the links too fast or using the back button of
IE to go back to the previous screen during the processing, the
contents of IFrame are not displayed. The following message is
displayed in place of IFrame contents:
This application does not support refreshing independently or
loading separately the contents of a list. Please try your operation
again by refreshing the entire page or by going through
<<https://192.24.168.141/action/networkServersDisplay>>
or <<https://192.24.168.141>>
Solution: If you click any link on the left navigation pane, you should
be able to view the contents properly.
You can do one of the following things after you encounter this
exception:
Click any of the links as suggested on the IFRAME page.
(<<https://192.24.168.141/action/networkServersDisplay>> or
<<https://192.24.168.141>>) in this case.
Click the link in the left navigation pane to go to the same page.
Schedules during
upgrades
server_cpu
server_export
pathname limitation
As part of the upgrade procedure, schedules that run between

midnight and 1:00 A.M. may lose those run times. This affects only
schedules that run every X hours of the day, (X being from 1 through
12), and not schedules that run at specific hours of the day. This is a
known problem and can be corrected by modifying the affected
schedules after the upgrade using Celerra Manager.
The server_cpu ALL -halt now command does not complete
when using Celerra NSX Network Servers. This is because the NSX
needs at least one Data Mover to be up in order to access its NAS
partitions through NBS.
If the Data Mover has ASCII mode for CIFS, when creating a share
using the server_export command, the share name cannot exceed
12 characters.
79
server_http
server_sysconfig
Symmetrix microcode
5771/5772
UNIX User
Management
The server_http command fails if you issue it on a Data Mover

while running fsck. Other commands (including DHSM API) may
not work until fsck completes.
If you use the option string of server_sysconfig, the entire string
must be enclosed by double quotes. Currently, the command does not
return an error message indicating a failure.
Symmetrix microcode version 5772 and some 5771 versions (.86.95,
.97.95, .91.99, .94.102, .95.103 and later) have been enhanced to
prevent prefetch tasks spawned by sequential reads from flooding
Symmetrix cache. This is done by quickly reusing Symmetrix cache
used by the prefetch tasks. For environments where it is highly
desirable that this prefetched data not be flushed out, (for example,
OLTP or configurations where a file is read sequentially into cache
and the cached data is repeatedly reused), contact your EMC
Customer Support representative and reference Primus article
emc164574.
Celerra UNIX User Management snap-in, Celerra UNIX Attributes
Migration tool, and Celerra UNIX Users and Groups property page
extensions in Active Directory Users and Computers modify the
schema by adding attributes and classes to it. These changes to Active
Directory are permanent, and uninstalling the snap-in will not
remove them.
The Celerra UNIX User tab and the Celerra UNIX Group tab do not
appear on the Properties page if you select Action-Find menu to
locate a user or group in the Active Directory Users and Computers
console. If you want to assign, remove, or modify UNIX attributes for
a user or group in the Active Directory Users and Computers console,
scroll through the list of users and groups and select the one you
want to view, or use the Celerra UNIX User Management snap-in
found in the Celerra Management Console.
Uploading dumps
from Data Mover
When a Data Mover panics, the /nas/var/ partition may not be

large enough to automatically upload the dump to /nas/var/dump.
When this occurs, an entry is logged in /nas/log/sys_log. You
can manually upload the dump from the slot dump area using the
the/nas/sbin/dump_slot utility.
The following is an example of /nas/log/sys_log entry:
Mar 14 20:48:08 2005 OTHERS:7:1 Slot 2 recover_slot
handle panic Uptime: 000 days, 03:36:44
80
Use this command to upload a dump:

/nas/sbin/dump_slot -d /tmp/slot2.dump/name/slot2.dump0
-v -F 2
Where:
2 = the slot where a dump is located
-v = verbose output
-F = full dumps
-d = the full path of the dump file
EMC recommends that dumps be uploaded to an NFS mounted file
system, for example, /mnt/celerraip1.
CAUTION
If you choose the default location (which is not recommended), the
dump is written to /nas/var/dump. It needs to be moved out as
soon as possible to avoid being deleted by the NAS cleanup scripts
that operate regularly.
Once you have uploaded the dump and ensured it is safe, you may
choose to delete it from the slot dump area.
Virtual device naming
When creating virtual devices, such as Fail-Safe Network (FSN)

devices, Ethernet channels, and link aggregations, do not use the
following names as the name of the device:
amgr
i82596
nd
rfa
tcpnfsmesh
blksrv
icmp
nfs
scsicfg
tcprpc
client
icmp6
nfsPrimary
smb
tftp
cmfap
ip
nwh
smpte
toeframechk
dep
ip6
pax
spstream
udp
dfe
iscsi
pcnfs
ssl
udpnfsmesh
echo
lockd
pfkey
sslpipe
elp
lpfn
pipe
statd
escon
mac
pmap
StreamHead
http
meshforwarder
rcp
tcp
81
Problem prevention
Problem prevention
The EMC Customer Service center has reported a number of calls
concerning the following issues. Please take a moment to familiarize
yourself with these potential issues.
Administrative
activities
Login as nasadmin
Shutting down a Data

Mover or deleting an
internal interface
As with any UNIX-based operating system, refrain from performing

normal administrative activities as Root user. Login as nasadmin
instead.
Shutting down a Data Mover or deleting an internal interface will
prevent you from issuing any server_ commands against that Data
Mover. Whenever you issue any command on the Celerra Control
Station CLI server, the command is sent to and executed by a Data
Mover. If that Data Mover is down or unreachable, the command
will fail.
Celerra AntiVirus
Agent (CAVA)
AntiVirus (AV) threads
If a large number of files need to be checked at the same time, the

CAVA file checking process may be slow depending on number of
threads configured, number of CAVA servers, and speed of the CAVA
servers. If the CAVA servers cannot process file checks quickly
enough, then all AntiVirus (AV) threads will become busy and the AV
collector queue will increase, indicating that file check requests are
queueing up. Increasing the number of AV threads (10 by default)
will allow more files to be checked in parallel and may improve
performance. However, the number of CAVA engines may need to be
increased to satisfy a heavy load.
Number of CAVA
engines
It is possible to configure your EMC Celerra installation with only

one Celerra AntiVirus Agent (CAVA) engine. However, this
nonhigh-availability configuration can make your installation
susceptible to data unavailability in the event that the CAVA engine
goes offline for any reason. The best method for preventing this type
of outage is to increase the number of available CAVA engines.
Additional information for configuring the Celerra AntiVirus Agent
can be found in the Using Celerra Anti Virus Agent technical module.
82
Problem prevention
This document describes the parameter setting that will specify

shutdown actions to take when no servers are available.
Removing AV user
from Data Mover
password file
If you remove an AV user from the Data Mover password file, CAVA
will not work. Depending on the setting you have on CIFS, you could
use up all the CIFS threads because they are used for virus checking
or it will shut down CIFS altogether.
Common Internet
File Systems
Deleting machine
account from active
directory
Deleting server out of

the domain
Naming CIFS shares

NetBIOS name
conflict
Deleting a machine account from active directory will result in loss of

access to the CIFS server and its resources. Deleting the computer
account of any CIFS server, including a Celerra Data Mover, from the
domain of which it is a member will prevent that CIFS server from
being able to authenticate both itself and users attempting to connect
to it with the domain. This will mean that users will not be able to
access data hosted by the CIFS server.
Deleting a server out of the domain will result in shares associated
with the server to be unavailable. The same applies to deleting the
machine account.
Do not name your CIFS share "all" or you risk deleting all of your
shares when trying to delete it.
Instead of destroying the conflicting NetBIOS, simply rename it using
the existing server_cifs <server_x> -rename -netbios
<oldname> <newname> command. Changing a NetBIOS name using
this command will update the share and local group and enable you
to reuse it.
LGDB and ShareDB information associated with the server are
persistent. If the target name of a rename conflicts with LGDB or
ShareDB information, recreate the old server and rename it to a name
that will not conflict with your naming structure to free the old slot.
Control Station
Applications
The Control Stations primary function is to manage the Celerra. The

use of applications that do not come installed on the Control Station
is not recommended. The Celerra Control Station is not tested or
certified with any of these applications and there could be issues with
memory usage or CPU time that may affect normal operation.
83
Problem prevention
Dual Control Stations
If you have dual Control Stations on either (CFS/CNS/NSxxx) and

they concurrently try to access the /nas file systems, the nasdb may
become corrupted and the Data Movers will be unable to
communicate with the Control Station.
Control Station zoning
Do not upgrade the CFS-14 or CNS-14 if the Control Station WWN is

present in more than one zone set. The Control Station must only be
zoned to a single Symmetrix FA or CLARiiON fibre port. The Control
Station Fibre Channel WWN must be present in only one zone.
In an environment with more than one Symmetrix or CLARiiON, the
Control Station must be in a zone that contains only one Symmetrix
or CLARiiON.
File System
Deleting files from
Data Mover or file
system
Files that you delete will be permanently lost unless you have a
backup copy. Files can be backed up and restored through the
Celerra SnapSure, TimeFinder/FS, Replicator or by using standard
tape backup methods.
Deleting file systems
File systems that you delete will be permanently lost unless you have
a backup copy. Backups can be created via TimeFinder/FS,
Replicator or other backup methods.
File system full
Insufficient space in a file system will prevent users from writing to

that file system. In addition to monitoring file system usage, set an
event notification for file system full condition and enable automatic
file system extension with a conservative threshold.
84
Mounting file system

to /tmp
Some variants of UNIX or Linux periodically delete some or all of the

files under the /tmp directory. If you mount a file system under /tmp
you risk having the contents of that file system deleted by the
automatic cleanup process.
Splitting BCVs
When a production file system (PFS) is mounted in a read/write

fashion to a Data Mover and has BCVs established, the command to
split (or mirror -off) the BCVs from the PFS volumes must be initiated
using the Celerra fs_timefinder command. The fs_timefinder
command ensures consistency of the BCV snapshot by managing file
system access and Data Mover log operations when the volumes are
split. These additional considerations are not available directly from
Solution Enabler commands (symmir).
Problem prevention
Name Resolution
SunOne/iPlanet
When using SunOne/iPlanet, make sure that the defaultServerList

attribute on the LDAP server is set to use IP addresses and not
hostnames. If hostnames are specified and the Celerra is configured
to use SunOne/iPlanet for host resolution, a deadlock can occur.
Networking
Changing an IP
address
Changing or
removing DNS settings
Changing IP address on a Data Mover can result in a naming

(DNS/NIS) mismatch and render the IP address inaccessible.
When you make changes to the DNS server, such as IP addresses,
make sure you reflect those changes in the Data Mover. Failure to do
so may result in accessibility issues.
Removing the DNS server will prevent the CIFS server from
communicating with the domain. Any CIFS server, including those
configured on a Celerra Data Mover that is a member of a Windows
2000 or later domain will use DNS to locate the Windows domain
controllers that it needs to communicate with to operate. Disabling
DNS in such an environment will prevent the CIFS server from
locating Windows domain controllers. This will prevent the CIFS
server from authenticating users attempting to connect to it. This will
mean that users will be unable to access the data hosted by the CIFS
server.
Connecting an
internal switch to an
outside network
The internal networks in the Celerra are designed to provide private

communication paths between Data Movers and the Control Station.
If that network is connected to an external network ,it is likely that
the resulting routing, addressing issues, and/or volume of network
traffic will disrupt the communications between the Data Movers and
Control Station, which will in turn disrupt the operation of the
Celerra.
Duplicate IP
Duplicating an IP address on two Data Mover interfaces will have the

same result as duplicating IP addresses on any two machines in a
network. The network will either be unable to determine which
location it should deliver packets to or block all traffic to both
machines on the assumption that one or other of them is attempting
to gain unauthorized access to information by impersonating the
other.
85
Problem prevention
FSN devices
Incorrectly adding
network route
Create Fail-Safe Network (FSN) devices on multiple network

interface cards (NICs) to ensure connectivity in the event of a
hardware failure.
Incorrectly adding a network route will cause network access to be
limited to the same subnet/broadcast domain. The default gateway,
or route, address tells any network client, including a Celerra Data
Mover or Control Station, where to send IP packets that are outside
its local subnet. If the default gateway is misconfigured, the network
client will not be able to contact any host that is not within its local
subnet. If the Celerra relies on hosts, such as domain controllers,
outside its local subnet, it may not be able to offer service.
Parameters
Changing parameters
Changing parameters could significantly affect the function of the

Celerra Network Server. Take the following precautions before you
make changes to Celerra parameters:
Do not change system parameters without first consulting with

EMC Customer Service.
Change server parameters one at a time, in small increments,

preferably in a nonproduction environment. Be sure to
thoroughly test each change before making additional changes.
Keep a detailed log of which server parameters you changed and

the reason for the change. This information will be useful if a
problem arises because of the change.
Managing
parameters
EMC recommends using the server_param command or the Celerra

Manager to manage parameters. If you manually edit the param file,
note the following:
Parameters are case-sensitive and must be entered exactly as

described in the Celerra Network Server Parameters Guide. A
typographical error, the omission of a dot, incorrect
capitalization, or a value out of range could result in adverse
effects.
Replication
.etc/passwd file
86
When configuring interconnect security for remote iSCSI replication,

an entry for the interconnect passphrase is added to the .etc/passwd
file for the Data Mover. Do not delete these entries, otherwise
authentication between Data Movers will fail.
Problem prevention
fs_copy and
fs_replicate
Using the fs_copy and fs_replicate commands between different

major releases of Celerra Network Server is not supported and may
lead to system instability or data loss. Be sure that the source and
destination are using the same major version of Celerra Network
Server, for example, 5.4 or 5.5, before using the fs_copy and
fs_replicate commands.
Multiple LUN 0s
Assigning multiple LUN 0s will result in Data Movers being unable

to boot successfully.
Storage
Powering down all

SAN switches
The Celerra does not store any user data itself. All data is stored on
the Symmetrix or CLARiiON disk array(s) that the Celerra is attached
to. With a Celerra gateway model the Celerra is attached to the disk
array(s) it uses via a Fibre Channel SAN. If that SAN is unavailable
then the Celerra will not be able to access the disk array(s) it relies on
and hence will not be able to offer normal service.
Reassigning volumes
with SDDR or making
fibre changes
Like any host of a Fibre Channel SAN or array, the Celerra will
remember the address of the LUNs that it uses in the attached disk
arrays. Any change in the identity or location of those LUNs without
a corresponding reconfiguration of the Celerra may result in the
Celerra being unable to access the LUNs it needs to provide normal
service.
Reassigning Celerra
LUN with Navi
Unbinding/rebinding
LUNs
Write disabling R1 and
R2
Same as Reassigning volumes with SDDR or making Fibre changes.

Same as Reassigning volumes with SDDR or making Fibre changes above.
If the disks that any host, including a Celerra, uses are set to be
read-only, then the host will be unable to write to those disks and
may not be able to offer normal service.
87
Technical notes
Technical notes
Configuration
guidelines
The values listed in the following section represent the highest or

lowest values tested by EMC. In some cases, actual capacities may
exceed those tested by EMC.
Table 8
Guideline/Specification
CIFS TCP connection
Share name length
88
CIFS guidelines
Maximum Tested
Value
Comment
64K (default and

theoretical max.), 20K
(max. tested)
Param tcp.maxStreams sets the maximum number of TCP connections

a Data Mover can have. Maximum value is 64K, or 65535. TCP
connections (streams) are shared by other components and should be
changed in monitored small increments.
80 characters
(Unicode).
12 chars (ASCII)
Unicode: The maximum length for a share name with Unicode enabled is
80 characters.
ASCII: In ASCII mode, the maximum share length is 12 characters.
Number of CIFS shares
10,000 per Data Mover

40,000 per Celerra
Number of NetBIOS
names/compnames per
Data Mover
509 (max.)
Limited by the number of network interfaces available on the Data Mover.

From a local group perspective, the number is limited to 509. NetBIOS
and compnames must be associated with at least one unique network
interface.
NetBIOS name length
15
NetBIOS names are limited to 15 characters (Microsoft limit) and cannot

begin with an@ (at sign) or - (dash) character. The name also cannot
include white space, tab characters, or the following symbols: / \ : ; , =
* + | [ ] ? < > . If using compnames, the NetBIOS form of the name is
assigned automatically and is derived from the first 15 characters of the
<comp_name>.
Comment (ASCII chars)

for NetBIOS name for
server
48
Limited to 48 ASCII characters. International characters are not currently

supported for comments.
Restricted Characters: You cannot use double quotation ("),
semi-colon (;), accent (`), and comma (,) characters within the body of a
comment. Attempting to use these special characters results in an error
message. You can only use an exclamation point (!) if it is preceded by a
single quotation mark ().
Default Comments: If you do not explicitly add a comment, the system
adds a default comment of the form EMC-SNAS:T<x.x.x.x> where
<x.x.x.x> is the version of the NAS software.
Note: You may notice a performance impact when managing a large

number of shares.
Technical notes
Table 8
CIFS guidelines (continued)

Maximum Tested
Value
Comment
Compname length
63 bytes
Number of domains
10 tested
512 (theoretical max.)
Block size negotiated
64 KB
Maximum buffer size that can be negotiated with Microsoft Windows

clients. To increase the default value, change param cifs.W95BufSz,
cifs.NTBufSz or cifs.W2KBufSz. Refer to the System Parameters Guide
for parameter information.
127
This value is fixed and defines the number of requests a client is able to
send to the Data Mover at the same time, (for example, a change
notification request). To increase this value, change the maxMpxCount
parameter. Refer to the System Parameters Guide for parameter
information.
200,000
(NS350/500/700,
DM514, DM514T, NSX,
NS40, NS80)
40,000 (DM510)
Note: A large number of open files could require high memory usage on
the Data Mover and potentially lead to out-of-memory issues.
Number of simultaneous
requests per CIFS session
(maxMpxCount)
Total number
files/directories opened
per Data Mover
In a Windows 2000 or Windows Server 2003-compatible CIFS server, the

CIFS server computer name length can be up to 21 characters when
UTF-8 (3 bytes char) is used.
The maximum number of Windows domains a Data Mover can be a
member of. To increase the max default value from 32, change param
cifs.lsarpc.maxDomain. Refer to the System Parameters Guide for
parameter information.
Number of Home
Directories supported
20,000 tested
Number of Windows/UNIX
users
20,000 tested
4 billion max value of UID. If NIS is used, there is a limitation of 64,000

entries in the password config file on particular UNIX machines. If using
internal Usermapper or Secmap, you are limited by the number of inodes
available on root file system.
Number of users per TCP

connection
16K (default), 64K

(max)
To increase the default value, change param cifs.listBlocks (default 64,

max 255). The value of this parameter times 256 = max number of users.
Note: TID/FID/UID shares this param and cannot be changed individually
for each ID. Use caution when increasing this value as it could lead to an
out-of-memory condition. Refer to the System Parameters Guide for
parameter information.
89
Technical notes
Table 8
Maximum Tested
Value
Comment
Number of CIFS
connections (share
accessed) per TCP
connection
16K (default), 64K

(max)

max 255). The value of this parameter times 256 = max number of
connection.
for each ID. Refer to the System Parameters Guide for parameter
information. Use caution when increasing this value as it could lead to an
out-of-memory condition.
Number of files/directories
opened per CIFS
connection
16K (default), 64K

(max) per CIFS
connection

max 255). The value of this parameter times 256 = max number of
files/directories opened per CIFS connection.
for each ID. Use caution when increasing this value as it could lead to an
out-of-memory condition. Be sure to follow the recommendation for total
number of files/directories opened per Data Mover. Refer to the System
Parameters Guide for parameter information.
96 (1GB memory)
256 (>1 GB memory)
Number of threads for CIFS users (if there is 1 GB of memory on the Data
Mover, the default is 96 threads; however, if there is over 1 GB of memory,
the default number of threads is 256).
29 (tested)
Note: The total number of VDMs, File Systems, and Checkpoints across a
whole cabinet cannot exceed 2048.
Number of CIFS threads
Number of VDMs per Data

Mover
Table 9
Max connections to
secondary storage per
primary (Celerra) file
system
Number of HTTP threads
for servicing FileMover
API requests per Data
Mover
90
CIFS guidelines (continued)
FileMover
Maximum Tested
Value
Comment
1024
64
16 (default), can be increased using server_http command, max

(tested) is 64.
Number of threads available for recalling data from secondary storage is
half the number of whichever is the lower of CIFS or NFS threads.
Technical notes
Table 10
File systems guidelines
Maximum Tested
Value
Mount point name length
240 bytes (ASCII)
File system name length
240 bytes (ASCII)

19 chars display for list
option
For nas_fs list, the name of a file system will be truncated if it is more
than 19 characters. To display the full file system name, use the info
option with a file system ID (nas_fs -i id=<fsid>).
Filename length
255 bytes (NFS),

255 characters (CIFS)
With Unicode enabled in an NFS environment, the number of characters

that can be stored depends on the client encoding type such as latin-1.
For example: With Unicode enabled, a Japanese UTF-8 character may
require three bytes.1
With Unicode enabled in a CIFS environment, the maximum number of
characters is 255. For filenames shared between NFS and CIFS, CIFS
allows 255 characters. NFS truncates these names when they are more
than 255 bytes in UTF-8, and manages the file successfully.
Pathname length
1,024 bytes
Note: Make sure the final path length of restored files is less than 1024
bytes. For example, if a file is backed up which originally had path name
of 900 bytes, and it is restoring to a path with 400 bytes, the final path
length would be 1300 bytes and would not be restored. 1
Directory name length
255 bytes.
This is a hard limit and is rejected on creation if over 255 limit. The limit is
bytes for UNIX names, Unicode characters for CIFS. 1
65,533
This is a hard limit, code will prevent you from creating more than 65,533
directories.
Subdirectories (per parent

directory)
Number of file systems
per Celerra
4096 (NSX, NS80, and

NS40)
2048 (All other models)
Number of file systems

per Data Mover
2048
Maximum Disk Volume

Size
2TB minus 2 MB
Comment
The "/" is used when creating the mount point and is equal to one
character. If exceeded, Error 4105:
Server_x:path_name: invalid path specified is
returned.
This max number includes VDM and checkpoint file systems.
Mount operation will fail when the number of file systems reaches 2048
with an error indicating maximum number of file systems reached. This
max number includes VDM and checkpoint file systems.
91
Technical notes
Table 10
Total storage for a Data
Mover (Fibre Channel
Only)
File systems guidelines (continued)

Maximum Tested
Value
NS
10 TB =NS350, NS500
16 TB =NS700, NS20
20 TB =NS40
24 TB =NS80, NSX
CNS
8 TB =514/514T DM
4 TB = 510 DM
Comment
These total capacity values represent Fibre Channel disk maximum with
no ATA drives. Fibre Channel capacity will change if ATA is used.
Refer to the Celerra Capacity Limits tables for more information,
including mixed disk type configurations.
Note: On a per-Data-Mover basis, the total size of all file systems, and
the size of all SavVols used by SnapSure, and the size of all SavVols
used by the Celerra Replicator feature, must be less than the total
supported capacity. Exceeding these limit can cause out of memory
panic.
All other industry-standard caveats, restrictions, policies, and best practices prevail. This includes, but is not limited to, FSCK
times (now made faster through multi-threading), backup and restore times, number of objects per file system, snapshots, file
system replication, performance, availability, extend times, and layout policies. Proper planning and preparation should occur prior
to implementing these guidelines.
File system size
See right column for TB

information per Data
Mover
16 TB (hard limit, enforced) NS700, NSX, NS20, NS40, NS80

10 TB NS350
x8 TB NS500 and 514/514T Data Mover
4 TB 510 Data Mover
When planning a file system, with respect to maximum desired size, be
sure to factor in the time required for FSCK, ACL checks, file system
build & restore.
This guideline applies to both Fibre Channel and ATA drives.
File size
16 TB (4 TB with
filesize policy of quota)
This hard limit is enforced and cannot be exceeded.
Number of directories
supported per file system
Number of files per
directory
92
Same as number of inodes in file system. Each 8 KB of space = 1 inode.
500,000
Exceeding this number will cause performance problems.
Technical notes
Table 11
iSCSI guidelines
Maximum Tested
Value
Comment
Minimum: 3 MB
Maximum: 2 TB less 1
MB
Because an iSCSI LUN requires overhead, you cannot create a LUN size
equal to the total available space; the maximum size for a LUN is slightly
less (typically 1 MB less) than the amount of available space in the file
system.
256
Each target can have as many as 256 LUNs (numbered 0 through

255).Windows, however, limits each target to 255 LUNs (numbered 0
through 254). To comply with the Windows limitation, Windows-compatible mode, which is enabled by default on the Data Mover, limits
the number of LUNs to 255.
If you use a non-Windows iSCSI initiator, you can disable Windows-compatible mode by modifying the iscsi.WinCompat parameter. Refer to
the System Parameters Guide for parameter information.
Minimum: 8 MB
Maximum: same as
LUN size
Due to a volume formatting limitation on Microsoft platforms, when

extending a LUN, the total LUN size after extension must be greater than
129 MB.
255 bytes
This name is an alias for the formal iSCSI name. The alias is an
alphanumeric string of as many as 255 characters. The alias can include
Unicode characters in UTF-8 encoding. Because the alias is the key field
in the Celerra Network Servers iSCSI databases, each alias must be
unique on a given Data Mover. 1
iSCSI qualified name

(IQN)
223 characters
EMC strongly recommends you leave this field empty and let the Celerra
Network Server generate a legal iSCSI name.
If you type a name, be sure it is globally unique and in either IQN or EUI
format. An IQN can have as many as 223 characters. The Celerra
Network Server converts all uppercase characters in an IQN or EUI to
lowercase characters. Only letters, numbers, "-", ".", and ":" are allowed
in iSCSI IQN names.
CHAP secret
255 characters
CHAP secrets must be a minimum of 12 characters and a maximum of

255 characters. Microsoft iSCSI initiator supports CHAP secrets of 12 to
16 characters.
LUN size
LUNs per target
LUN extension size
Target alias
Snapshots per LUN
2000 (theoretical limit)

1000 tested
Replication session alias
256 characters
Replication application
label
256 characters
Maximum includes temporary writable snapshots.
93
Technical notes
iSCSI guidelines (continued)
Table 11
Maximum Tested
Value
4096 bytes
Replication attached
comment text or binary
data
Configured replication
sessions per Data Mover
1024
Concurrent active
replication sessions per
Data Mover
128 (theoretical max)
Destinations per source

file system
iSCSI timeout
See comments
Disk volumes (LUNs)

supported in MPFS/
MPFSi configurations per
file system
Stripe Size
recommendation
Max # of SCSI/ISCSI
devices (LUNs)
Max # of threads per Data
Mover
94
Number of configured replication sessions that can be actively

transferring data at one time.
Exceeding this number may cause performance implications on the
source file system.
EMC recommends client iSCSI initiator timeout be longer than failover
time. MS iSCSI Initiator has a default timeout of 1 minute.
MPFS guidelines
Table 12
Comment
Maximum Tested
Value
2048
Comment
Note: When new LUNs are added to the CLARiiON, MPFS client does
not automatically recognize them unless they are added to the storage
group in the same zone, and the MPFS client is rebooted. The switch
must be properly zoned and LUN masking must be properly configured in
order for the MPFS clients to see the new LUNs.
CLARiiON: 256
Symmetrix: 32
RHEL3: 255
RHEL4: 300
128
CLARiiON storage systems operating with MPFS(i) require a stripe size

depth of 256 KB (or 262144 bytes).
Symmetrix storage systems operating with MPFS(i) require a stripe size
depth of 32 KB (or 32768 bytes).
This is a client limitation.
The default number of threads is 16. As you increase the number of

threads, the Data Mover allocates more resources, primarily memory, to
accommodate the new threads. (Each thread consumes 16 KB of
memory; 128 threads consume 2048 KB.) You must consider the
resources consumed when modifying the number of threads.
Technical notes
Table 12
Host name length
Maximum Tested
Value
Comment
16 chars
Table 13
Number of DNS domains
Number of NIS Servers

per Data Mover
NIS record capacity
Naming Services guidelines

Maximum Tested
Value
Table 14
Comment
3 - WebUI
unlimited - CLI
Three DNS servers per Data Mover is the limit if using WebUI. There is
no limit when using the CLI (command line interface).
10
You can configure up to 10 NIS servers in a single NIS domain on a Data

Mover.
Note: A Data Mover supports only one NIS domain. Each time you
configure a NIS domain and specify the servers, it overwrites the
previous configuration.
1004 bytes
Number of DNS servers

per DNS domain
Password length
MPFS guidelines (continued)
A Data Mover can read 1004 bytes of data from a NIS record.
Network Data Management Protocol (NDMP) guidelines

Maximum Tested
Value
15
Comment
Passwords must be between 6 and 15 characters, and cannot contain
white spaces, tab characters, or any of the following symbols: / \ : ; , =
* + | [ ] ? < >. The password limitation is not NDMP specific, and
applies to any user account passwords created by server_user.
95
Technical notes
Table 15
Maximum Tested
Value
Number of NFS exports
Network File System (NFS) guidelines
2,048 per Data Mover

tested
Unlimited theoretical
max.
Number of concurrent
NFS clients
64K with TCP

(theoretical)
Unlimited with UDP
(theoretical)
Number of
Windows/UNIX users
Number of UNIX groups

supported
Table 16
Link aggregation/ether
channel
Number of VLANs
supported
Number of interfaces per
Data Mover
96
Note: You may notice a performance impact when managing a large

number of exports using Celerra Manager.
Limited by TCP connections.
4 billion max value of UID. If NIS is used, there is a limitation of 64,000

entries in the password config file on particular UNIX machines. If using
internal Usermapper or Secmap, you are limited by the number of inodes
available on root file system.
Netgroup line size
Comment
16383
The maximum line length that the Data Mover will accept in the local
netgroup file on the Data Mover or the netgroup map in the NIS domain
that the Data Mover is bound to.
64K
2 billion max value of any GID. The maximum number of GIDs is 64K,
but an individual GID can have an ID in the range of 0- 2147483648
Networking guidelines
Maximum Tested
Value
8 ports (ether channel)

12 ports (link
aggregation (LACP)
4094
45 tested
Comment
Ether channel: the number of ports used must be a power of 2 (2, 4, or
8). Link aggregation: any number of ports can be used. All ports must
be the same speed.
Mixing different NIC types (i.e., copper and fibre) is not recommended.
IEEE standard.
Theoretically 509.
Technical notes
Table 16
Maximum Tested
Value
Number of FTP
connections
Networking guidelines (continued)
Comment
Theoretical value 64K
By default the value is (in theory) 0xFFFF, but it is also limited by the
number of TCP streams that can be opened. To increase the default
value, change param tcp.maxStreams (set to 0x00000800 by default).
If you increase it to 64K before you start TCP, you will not be able to
increase the number of FTP connections. Refer to the System
Parameters Guide for parameter information.
255 characters
SNMP community names used when specifying SNMP trap address for
any of the four types of notifications are limited to 255 characters (A-Z,
a-z, 0-9, . and - ). Community names can also contain the following
special characters ~ ! @ $ % ^ * + = { } : ? _ #.
Community name used to

specify SNMP trap
address
Note: These are shown with a spaces for readability, but no spaces are
allowed.
Table 17
Quotas guidelines
Maximum Tested
Value
Number of Tree Quotas
2047
Max size of Tree Quotas
4 TB
Max. number of unique

groups
64K
Table 18
Comment
per file system
per file system
Replication guidelines
Maximum Tested Value
Number of Replication
sessions per Data Mover
64 (NSX-NSX only)
(other DMs) : 32
Comment
VDM replication sessions must be included, assuming adequate
resources are available.
Note: For best performance, do not use timeouts less than 5 minutes or
high water mark less than 128 MBytes on Celerra Replicator. Delta sets
should not be created faster than every five minutes in most
environments.
97
Technical notes
Table 19
SnapSure guidelines
Number of checkpoints
per file system
96 (other DMs)
64 (510 DM)
32 (510 DM & PFS
replicated)
Table 20
Comment
Up to 96 checkpoints per file system are supported.
Virus Checking guidelines
Comment
Virus checker file pending
90% of available VNODES in

system
The percentage of total cached open files available in the system

which can be pending on Virus Checking. An event is sent to the
Control Station when the maximum is reached. To change the
default value, modify param viruschk.vnodeMax(default=2000)
and/or param viruschk.vnodeHWM(default=0x5a(90%). Refer
to the System Parameters Guide for parameter information.
1 On systems with Unicode enabled, a character may require between 1 and 3
bytes, depending on encoding type or character used. For example, a

Japanese character typically uses 3 bytes in UTF-8. ASCII characters
require 1 byte.
98
Technical notes
Capacity guidelines
File system size
Table 21
The maximum file system size supported per Data Mover is 16 TB, as
listed in the following table. These guidelines apply to both Fibre
Channel and ATA drives.
Maximum file system size
NS350
NS500
NS600 NS700
5.5
5.5
5.5
10
10
NSX
NS20
NS40
NS80
5.5
5.5
XB60/65
5.5
5.5
5.5
XB60/65
16
24
16
20
24
16
24
16
(FCmax16
)
20
24
16
128TB4, 7
N/A
DART Version
FC Limit (TB) per
DM/Blade
FC+ATA Limit (TB) per
DM/Blade
10
16
6 (FC max
6)
(FC max (FC max
10)
10)
N/A
B2D w/ATA Limit (TB)

per DM/Blade
10
32
<4
32
32
32
32
32
Max FS Size (TB)
10
16
16
16
16
16
2048
2048
2048
2048
2048
2048
2048
2048
2048
2048
2048
2048
4096
2048
4096
4096
Max replication sessions

per DM/Blade
32
32
32
32
64
32
32
64
Max # of checkpoints
per PFS2
96
96
96
96
96
96
96
96
Max # FS per DM/Blade1

Max # FS per cabinet
16
6 (FC max
6)
(FC max
10)
128TB4, 5 128TB4, 5
MPFS with FC or ATA

Limit per DM/Blade
2048 is maximum for all FSIDs (File System IDentifiers)

PFS is Production File System
3 Limited to 32 if PFS is replicated.
4
No snaps or replication. If planning to use snaps or replication, the Data Mover
capacity limit will be the same as non-MPFS.
5
Requires DART V5.4.28.0 or later or DART V5.5.22.0 or later with MPFS for
Windows V5.0.12.0 or later or MPFS for Linux V4.3.20.0 or later.
2
When planning a file system, with respect to maximum desired size, be sure
to factor in the time required for FSCK, ACL checks, file system build and
restore.
99
Technical notes
Table 22
Celerra capacity Limits - CNS series

CNS (510 Data
Mover)
CNS (514 Data

Mover)
5.5
5.5
FC Limit (TB) per DM/Blade
FC + ATA Limit (TB) per DM/Blade
16
16
B2D w/ ATA Limit (TB) per DM/Blade)
32
32
Max FS Size (TB)
Max # FS per DM/Blade1
2048
2048
Max # FS per Cabinet1
2048
2048
Max Replication Sessions per DM/Blade
32
32
Max # Checkpoints per PFS3
643
96
DART Version
2048 is maximum for all FSIDs (File System IDentifiers)

is Production File System
3
Limited to 32 if PFS is replicated.
2 PFS
100
Configuring CX
storage system
arrays
For information about configuring CX storage system arrays using

predefined configuration templates, refer to the EMC Celerra Network
Server Version 5.5 Configuring CX Storage System Arrays technical note.
Control Station
Zoning
Requirements
Do not upgrade the CFS-14 or CNS-14 if the Control Station WWN is

present in more than one Zone Set. The Control Station must only be
zoned to a Single Symmetrix FA or CLARiiON Fibre Port. The
Control Station Fibre Channel WWN must be present in only one
Zone. In an environment with more than one Symmetrix or
CLARiiON, the Control Station must be in a zone that contains only
one Symmetrix or CLARiiON.
iSCSI Replication
with Replication
Manager (RM)
For detailed information about iSCSI disaster recovery procedures

when using Replication Manager, refer to the iSCSI Disaster Recovery
Solution Implementation Guide, available on Powerlink (under
Support > Technical Documentation and Advisories > Software P-R
Documentation > Replication Manager).
Technical notes
Mapping between
network ports and
software device
names
On some NS700 Data Movers, the network ports are labeled 11

through 16 and correspond to software device names cge0 through
cge5. These ports support 10 Mb, 100 Mb, or 1 Gb operation and are
connected to your LAN.
The network port that is labeled 01 corresponds to the software
device fxp0. The fxp0 device is reserved for the internal network and
is connected to the switch inside the cabinet. Some NS700 Data
Movers are marked with actual port names (cge0, and so on).
Refer to Configuring and Managing Celerra Networking technical
module for more information on networking ports.
Celerra scripting
guidelines
Table 23
The following table lists guidelines to keep in mind when creating

Celerra command scripts.
Scripting guidelines
Issue
Action
Scheduling
NAS Database Backups:

The Celerra Network Server backs up the NAS database that stores specific configuration
information required for each Data Mover every hour, at one minute after the hour. During part of
the backup, the database is locked, and some commands that rely on the database might not have
access. EMC recommends that command scripts avoid starting at one minute after the hour. Note
that scripts with complex commands that run for an extended period may overlap the backup
period.
The duration of the backup may vary. Use the following Linux command to check the state of the
backup process prior to executing scripts: ps -ef | grep nasdb_backup. If a lock condition occurs,
wait a few minutes and retry.
Replication:
If you are using the Celerra Replicator restartable checkpoint feature to protect from out-of-sync
issues, please be aware that the system uses a CRON job to determine if a restartable checkpoint
needs to be refreshed at 25 minutes after the hour. This process may generate a checkpoint
refresh, depending on the state of the replication session. During a checkpoint refresh, the
database is locked. If another operation requires a conflicting lock, it receives an "unable to
acquire locks" error. If this occurs, wait a few minutes and retry.
In version 5.4, the checkpoint refresh process is sequential and may impact performance. Version
5.5 supports concurrent checkpoint refreshes, which are less resource-intensive.
Command sequencing
Some commands must lock the database in order to execute. If multiple user-entered commands
or scripts are active at the same time, some of these commands may lock the database and
prevent other commands from executing. To avoid this, you should sequence commands
whenever possible.
101
Technical notes
Table 23
Scripting guidelines (continued)
Issue
Action
Sleep statements
Some processes within a script can take time to complete. Use proper timing and adequate sleep
statements to prevent timing-related issues.
Pipe and grep
Piping script outputs through grep is a helpful tool to check the status of the script. Use periodic
checks to grep for file or database locked messages, timeouts, resource unavailable warnings,
and other failure or success messages, and use this information to check status, pause the script,
or halt it. Detailed information about error messages can be found in the Error Messages Guide.
Return code check
All commands return a UNIX-style return code (for example: 0 for success or 1 for failure) or a
text-based status code (for example, done) which can be used to help determine if the command
completed or if there was an error or a conflict with the NAS database backup, or other commands
being run. If a lock condition occurs, wait a few minutes and retry. If you create and run scripts, be
sure to incorporate return code checks and verify for proper return codes from individual
operations.
Scripting examples
The RECOVERABLE variable contains the following errors to retry

on:
Unable to acquire lock
Resource temporarily unavailable
Unable to connect to host
Socket: All ports in use
Database resource vanished
Connection timed out
NAS_DB locked object is stale

Following is an example of what the RECOVERABLE variable looks
like:
RECOVERABLE="unable to acquire lock|Resource
temporarily unavailable|unable to connect to
host|socket: All ports in use|database resource
vanished|Connection timed out|NAS_DB locked object is
stale".
The res variable contains the command output.

#!/bin/sh
########################################################
# File: nas_cmdrcvr.sh
# Created by: NAS Engineering
# Date: Thursday, May 25, 2006
# Version: 1.0
# Notes:
102
Technical notes
# 1) script will retry commands for specified period of

time
#2) script will log messages to file only if there's
available disk space
########################################################
NAS_DB=/nas
export NAS_DB
PATH=$PATH:$NAS_DB/bin:$NAS_DB/sbin:/usr/sbin:.
export PATH
RETRIES=60
SLEEPTIME=30
LOGDIR="/home/nasadmin"
LOGFILE="$0.log"
LOGLAST="$0.last"
DISKFULL=98
RECOVERABLE="Resource temporarily unavailable|\
unable to acquire lock|\
unable to connect to host|\
socket: All ports in use|\
database resource vanished|\
Connection timed out|\
NAS_DB locked object is stale"
#
# function to log messages to a file
#
nas_log()
{
DISKCHK=`df -k $LOGDIR|awk 'NR>1{print $5}'|sed 's/\%//'`
# if there's enough free disk space, append to log
if [ $DISKCHK -lt $DISKFULL ]; then
TDSTAMP=`date +%Y-%m-%d' '%T`
echo $TDSTAMP: $LOGMSG >> $LOGDIR/$LOGFILE
fi
# regardless of available space, always write last error
echo $TDSTAMP: $LOGMSG > $LOGDIR/$LOGLAST
}
#
# function to execute (and potentially retry) commands
#
nas_cmd()
{
# initialize variable(s)
103
Technical notes
retry_count=0
# loop until either successful of retry count exceeded
while [ $retry_count -le $RETRIES ]; do
# execute command and gather response
RES=`$CMD 2>&1`
# check if response means command is recoverable
if [ `echo "$RES"|egrep -c "$RECOVERABLE"` -ne 0 ]; then
# check retry count
if [ $retry_count -ne $RETRIES ]; then
# retry count has not been exceeded
LOGMSG="Command ($CMD) failed with ($RES)...retrying in
$SLEEPTIME s"
nas_log
sleep $SLEEPTIME
else
# retry count has been exceeded
LOGMSG="Command ($CMD) failed with ($RES)...exiting
(retry count of $RETRIES exceeded)"
nas_log
exit 1
fi
else
# command was either successful or failed for an unknown
reason
LOGMSG="Command ($CMD) successful with ($RES)"
nas_log
retry_count=$RETRIES
exit 0
fi
#increment counter for retries
retry_count=`expr $retry_count + 1`
done
}
#
# main
#
CMD="nas_volume -d mtv1"
nas_cmd
Using the NAS

database and query
facility
104
EMC has partially changed the layout or format of Celerra internal

databases. This change can impact the use of awk or grep utilities
when used in scripts that assume specific positions of fields in
databases.
Technical notes
To enable searching of the NAS database, Celerra has developed a

new query subsystem that appears as a hidden option on some of the
nas_commands. This query subsystem enables you to specify the
information you are interested in, allows you to format the output,
and is independent of the database format.
CAUTION
Do not use grep and awk to scan the database files. Database
positions may change and substrings may return false matches for
database objects.
Following is an example of a query to view unused disks:
nas_disk -query:inuse==n -format:'%s\n' -Fields:Id
To filter root disks, refer to the List all non-root disks that are not in use
in Table 10.
Examples
Use the following commands to view the tags (fields) that you can
query:
nas_disk query:tags
nas_fs query:tags
nas_volume query:tags
nas_slice -query:tags
Table 24 contains a list of examples to help you get started. Note that
these commands can be run on the Control Station CLI, so the
hardcoded values can be replaced with shell script variables.
Table 24
Task and query examples
Task
Example
Query the ID of a named file system
nas_fs -query:Name==RLL_fs10 -format:'%s\n'

-Fields:Id
Query the ID of a named file system

without the new line
nas_fs -query:Name==RLL_fs10 -format:'%s'

-Fields:Id
Query the name of a file system ID that

corresponds to a particular ID
nas_fs -query:id==20 -format:'%s\n' -Fields:Name
List of all server IDs
nas_server -query:* -format:'%s\n' -Fields:Id
List of all server names
nas_server -query:* -format:'%s\n' -Fields:Name
List of all the checkpoint file systems
nas_fs -query:type==ckpt -fields:name

-format:"%s\n"
105
Technical notes
Table 24
106
Task and query examples (continued)
Task
Example
List type of file system with ID 20
nas_fs -query:id==20 -format:'%s\n' -Fields:Type
List the file systems that are in use
nas_fs -query:inuse==y -format:'%s\n' -Fields:Name

or
nas_fs -query:inuse==y -format:'%s\n' -Fields:Id
Identify file system of which file system ID

28 is a backup
nas_fs -query:id==28 -format:'%s' -Fields:BackupOf
List the name of the server with ID 2
nas_server -query:id==2 -format:'%s\n'

-fields:name`"
View which volume file system is built on
nas_fs -query:Name==my_fs -format:%d

-fields:VolumeID
View the block count of meta volume
nas_volume -query:Name==my_meta3 -format:%d

-fields:Blocks
View the block size of meta volume
nas_volume -query:Name==JAH_meta3 -format:%d

-fields:BlockSize
Find which server IDs use fs123
nas_fs -query:name==fs123 -format:'%s\n'

-fields:ServersNumeric
List all non-root disks that are not in use
nas_disk -query:inuse==n:IsRoot==False
-format:"%s\n" -fields:name
List unused volumes that contain dc in

the volume name
nas_volume -query:inuse==n:IsRoot==False:name=dc
-format:"%s\n" -fields:name
List all available disks on a particular

storage device (symm_id is a script/env
variable)
nas_disk -query:inuse==n:SymmID==$symm_id:IsRoot==False -format:"%s\n"

-fields:name
Environment and system requirements
Query operators
Table 25
2007 Daylight
Saving Time change
for United States
and Canada
Use these operators when building your queries:

Operator definitions
Operator
Definition
Having the pattern (sub-string match)
==
Exact string match
=-
Integer Minimum (not less than)
=+
Integer Maximum (not more than)
=*
Any
=^
NOT having the pattern
=^=
NOT an exact match
=^-
NOT Integer Minimum (is less than)
=^+
NOT Integer Maximum (is more than)
=^*
NOT Any (None)
for the United States and Canada. It is now observed from the second
Sunday in March to the first Sunday in November, adding about a
month to the total DST period.
Information about the 2007 daylight saving time change is detailed in
the EMC Celerra Network Server U.S. and Canada 2007 Daylight Saving
Time Change technical note, available on Powerlink.

Celerra internal
private subnet has
changed
The Celerra internal private subnet configuration has changed on all

Celerra integrated systems (NS350/NS20/NS40/NS80) that are
pre-installed by manufacturing. This change was necessary because
the software default values based on the 192.168 subnets were
increasingly conflicting with customers networks. The new internal
private subnet configuration is in the 128.221 subnet. This is an
EMC-specific subnet and should never be encountered in a customer
network. Refer to ETA emc173555 for more detailed information. This
change supersedes all documentation that identifies the internal
private IP addresses as being on the 192.168 subnet.
107
Microcode
The minimum required microcode levels supported by Celerra

Version 5.5 are listed below. Refer to the E-Lab Interoperability
Navigator for the latest microcode information.
Symmetrix
5567.52.29
5568.56.22
Symmetrix DMX1 and 2
5669.47.25
5670.27.29
5671.31.35
Symmetrix DMX3
5771.68.75
Symmetrix DMX4
5772.83.75
Symmetrix Enginuity
5772.55.51
CX300/CX400/CX500/CX600/CX700
NS700G fabric attach systems support: CX300, CX400, CX500, CX600,
CX700, Symmetrix 5 series, and DMX.
The supported releases of CLARiiON are FLARE 19 (CX400/CX600
only), FLARE 24 and FLARE 26. If you are running an older release,
you must upgrade the CLARiiON. You will receive a warning during
the upgrade if the CLARiiON software is not up to proper revision.
System
requirements for
SnapSure and
Replicator
108
On a per-Data-Mover basis, the total size of all file systems, and the
size of all SavVols used by SnapSure, and the size of all SavVols used
by the Celerra Replicator feature, must be less than 16 TB.
To view the size of a file system or a SnapSure SavVol, use the
nas_fs -size command. To view the size of a Replicator SavVol,
use the fs_replicate -info command. For more details, refer to
the E-Lab Interoperability Navigator.
Documentation
Documentation
Documentation
clarification and
corrections
Celerra Manager
For Celerra Manager Notifications help, the descriptions should be as

follows:
Celerras > [Celerra_name] > Notifications
Destination of notification. Format depends on Action: use absolute
path on Control Station for log file, one or more SMTP addresses
separated by commas for mail, or IP address; community name or
hostname; community name for trap.
Celerras > [Celerra_name] > Notifications > (right click)
Notification Properties
Type a destination for the notification. The format depends on the
action: Use an absolute path on the Control Station for log file, an IP
address; community name or hostname; community name for trap,
or one or more valid, comma-separated SMTP addresses in the
format user@domain (as in storage_admin@emc.com). Any address
which does not include the domain (as in nasadmin) is assumed to be
local to the host from which the mail originates.
Celerras > [Celerra_name] > Notifications and click New
Type a destination for the notification. Format depends on Action:
Use an absolute path on the Control Station for log file, an IP address;
community name or hostname; community name for trap, or one or
more valid, comma-separated SMTP addresses in the format
user@domain (as in storage_admin@emc.com). Any address which
does not include the domain (as in nasadmin) is assumed to be local
to the host from which the mail originates.
Extending a file
system after
replication has been
suspended
If the source file system is extended automatically or manually after

replication has been suspended, perform the following steps to
ensure that the source and destination file systems are the same size
before restarting replication.
Note: If you do not perform these steps before restarting replication, the
restart will fail with an error because the source file system and destination
file system are no longer the same size (that is, the block counts differ). In this
case, you must perform these steps to recover from the error.
109
Documentation
Step action
1. Verify that the destination file system type is set to rawfs. If it is
set to uxfs, convert the destination file system from uxfs to rawfs
using this command syntax:
nas_fs -Type rawfs <dstfs> -Force
Where:
<dstfs> = name of the destination file system
Example:
$ nas_fs -Type rawfs dst_ufs1 -Force
Note: A read only file system must be set to rawfs prior to a file system
extension and/or replication restart.
2. Extend the destination file system manually using the same size
as the source file system:
nas_fs -xtend <dst_fs> size=<integer>[G|M] -option
<options>
Where:
<integer> = the size of the secondary file system in gigabytes or
megabytes
<option> = Specifies any comma-separated option(s), such as
slice={y|n}, which specifies whether the disk volumes used by
the file system may be shared with other file systems using a slice.
Example:
To extend the destination file system dst_ufs1 by 2 megabytes,
using the slice option, to match the source file system extension to
the same size, type:
$ nas_fs -xtend dst_ufs1 size=2M slice=y
3. Restart the replication relationship using this command syntax

from the source site:
$ fs_replicate -restart <srcfs> <dstfs>:cel=<cel_name>
Where:
<scrfs> = name of the source file system
<cel_name> = name of the destination Celerra Network
Server
Example:
To restart a replication relationship, type:
$ fs_replicate -restart src_ufs1 dst_ufs1:cel=cs110
110
Documentation
MirrorView
Related documentation for this release may include references to the

MirrorView/Synchronous with Celerra for Disaster Recovery feature.
For example, the man page for the nas_storage command includes
the -modify option, which is currently intended for use by
MirrorView/Synchronous only. This feature is currently available
only through the RPQ (request for price quotation) process.
For more information about the MirrorView/Synchronous with
Celerra feature, contact your local EMC sales representative or your
local EMC Service Provider.
NDMP backups
SnapSure
In the Configuring NDMP Backups on Celerra technical module, note

the following:
Minor parameter correction:
The range for the bufsz parameter is 64-1168, rather than 0-1168.
The following two corrections apply to page 18 of the Using SnapSure
on Celerra technical module, version 5.5, P/N: 300-002-720:
1. Section Restoring checkpoints from unmounted PFS has to be
updated to include a procedure for manually extending SavVol
size.
Revised text:
The section should be read as follows:
Restoring checkpoints from unmounted PFS
The SavVol automatic extension feature does not work on
unmounted file systems. If you are restoring checkpoints from an
unmounted PFS, first verify if there is enough space available on
the SavVol. If there is insufficient space, you must manually
extend it. To manually extend SavVol, use the CLI command
nas_fs -xtend <fs_name> size=<integer>[T|G|M|%].
2. The title Custom SavVol creation using SnapSure's options is

misleading. The text therein documents the procedure to shrink
the size of SavVol.
Revised text:
The title should be read as follows:
Shrinking SavVol size
System Configuration
Guide
Incorrect text in the Celerra Network Server System Configuration

Guide (300-000-832 rev a06). This document incorrectly states to zone
Control Station 0 to both SP A and SP B. This is incorrect. The Control
Station should have a single path to SP A. Fibre Channel zone 0
should contain the Control Station and SP A. SP B should not be in
this FC zone. The pre-upgrade health check will fail if you have SP B
in FC zone 0.
111
Documentation
TimeFinder/FS
The TimeFinder/FS FarCopy configuration shown in Using

TimeFinder/FS, NearCopy, and FarCopy on Celerra, "Task 5: Create a
FarCopy snapshot of the imported file system," the fs_timefinder
<fs_name> -Snapshot command line should now include -option
disktype=STD to ensure proper selection of the disktype.
Documentation CD
New browser window
If you have Autorun enabled and want the Celerra Network Server
Documentation CD to open in a new browser window instead of
replacing an open browser window, modify your browser's options
for opening new windows or tabs. Refer to your browser's online
help for details.
External links
To select hyperlinks to external sources on the Documentation CD,

right-click and select Open in New Window from the context menu
in Internet Explorer or Open Link in New Window from the context
menu in Netscape Navigator.
Using Celerra features
The Celerra Network Server Version 5.5.31 Documentation CD provides

online HTML guides, a set of Adobe Portable Document Format
(PDF) technical documents, a copy of the online help systems for the
Celerra Network Servers management applications, and EMC
Customer Support contact information.
EMC recommends that you read the document associated with a
Celerra feature to familiarize yourself with the concepts, options, and
best practices designed to optimize functionality before beginning the
installation or setup procedures.
Celerra Network
Server training
Celerra Network
Server Version 5.5
documentation
112
EMC Customer Education Courses are designed to help you learn

how EMC storage products work together and integrate within your
environment in order to maximize your entire infrastructure
investment. EMC Customer Education features online and hands-on
training in state-of-the-art labs conveniently located throughout the
world. EMC customer training courses are developed and delivered
by EMC experts. For course information and registration, refer to the
EMC Powerlink website: http://Powerlink.EMC.com.
This section lists the complete set of user documentation for version
5.5 software. All components of this set, unless otherwise indicated,
reside on the Celerra Network Server Version 5.5.31 Documentation CD.
You can also download the documents from the EMC Powerlink
website at http://powerlink.emc.com.
Documentation
Documentation kit
contents
The EMC Celerra Network Server Version 5.5 user documentation kit
contains the following.
Kit Contents
Getting started
Table 26
Item
Part Number
Celerra Network Server Version 5.5.31 Documentation CD
053-001-470 Rev A07
Table 26 lists getting started information described in technical

modules on the Celerra Network Server Version 5.5.31 Documentation
CD, online help, or in hard-copy format.
Getting started
Part number
Name
Description
300-002-692
Celerra Network Server Release

Notes
Contains new announcements, updates, requirements, and precautions

related to software installation or upgrades, as well as a list of known issues
and bug fixes. You can obtain the latest release notes on EMC Powerlink.
Online
Whats New in Celerra V5.5
A feature of the Celerra Network Server Version 5.5.31 Documentation CD

user interface, Whats New in Celerra V5.5 identifies the new features and
functional enhancements offered by Celerra Network Server Version 5.5
software.
Online
Learning about Celerra
A feature of the Celerra Network Server Version 5.5 Documentation user

interface, the Learning about Celerra topic introduces you to the EMC Celerra
Network Server models and their components. Includes a high-level
conceptual overview of Celerra Network Server networking, volumes and file
systems, file access protocols, iSCSI support, and business continuance
features, such as high availability and data protection.
Online
Getting started with Celerra topic

on the Celerra Network Server
Version 5.5.31 Documentation
CD and in Celerra Manager
online help
A feature of the CD user interface and Celerra Manager online help. Getting
started with Celerra provides a high-level outline of essential Celerra
management tasks that includes links and references to Celerra Manager
documentation, Celerra command line interface documentation, and Celerra
Wizards that help you perform the tasks.
300-002-677
Celerra Configuration
Quick Start
Describes the steps required to configure a Celerra Network Server using the
Set Up Celerra wizard in Celerra Manager.
300-002-702
Celerra Network Server System

Operations
Explains how to power up and power down the Celerra cabinet, perform an
emergency shutdown, halt and reboot the Control Station, Data Movers and
blades, manage the Control Station using CLI commands, and use preventive
troubleshooting to maintain system operation.
113
Documentation
Celerra management
Table 27
114
Table 27 lists the Celerra management information described in

technical modules and online help on the Celerra Network Server
Version 5.5.31 Documentation CD or in the applications online help.
Celerra management
Part number
Name
Description
300-002-681
Installing Celerra Management

Applications
Provides a brief description of, and explains how to access or install, the
Celerra Network Servers management and monitoring applications and
interfaces, including Celerra Monitor and Celerra management snap-ins and
tools.
Online Help
Celerra Manager
Provides guidance as you use the Celerra Web Manager to perform Celerra
administrative tasks from a remote computer with a supported browser and
network connection to the Celerra Control Station. Helps you manage and
configure Celerra file systems, Data Movers, networks, hosts, volumes, and
storage systems.
Online Help
Celerra Monitor
Provides guidance as you use the Celerra Monitor, a Java-based client/server

application software package, to monitor Symmetrix performance and any
Data Movers mounted on the Celerra Network Server. Celerra Monitor allows
you to observe system operation, performance, configuration, and statistics;
check Symmetrix error logs; and control access and polling of monitored data.
Online Guide
Monitoring Celerra
Provides an overview of the various monitoring functions in Celerra Manager

and Celerra Monitor.
Online Help
Celerra Data Mover Security

Settings
Explains how to manage audit policy by selecting which Data Mover security
events are logged into the Windows Security log, and how to manage user
rights by selecting which users and groups can log in to or execute tasks on a
Data Mover.
Online Help
Celerra Home Directory

Management
Explains how to enable home directories on Celerra.
Online Help
Celerra UNIX Property Page

Extensions in Active Directory
Users and Computers
Explains how to modify the Celerra UNIX user and group attributes using
property page extensions in Active Directory Users and computers.
Online Help
Celerra UNIX Attributes Migration Explains how to migrate UNIX attributes from the NIS server or UNIX
Tool
password and group files, or remove user and group attributes from the Active
Directory.
Online Help
Celerra UNIX User Management
Explains how to view and modify Celerra UNIX user and group attributes.
Online Help
Celerra AntiVirus Management
Explains how to modify CAVA virus-checking configuration parameters on a

Celerra Data Mover from a Windows 2000 or NT server.
300-002-708
Configuring Celerra Events and

Notifications
Describes how to configure the Celerra Network Server to record and display
various system events, such as errors, alerts, and commands to facilitate
system administration and troubleshooting.
Documentation
Table 27
Celerra management (continued)
Part number
Name
Description
300-002-727
Configuring Standbys on Celerra
Describes Control Station and Data Mover failover, and how to implement a
standby Control Station or standby Data Mover(s) on the Celerra Network
Server.
300-002-717
Managing Administrative
Accounts on Celerra
Describes how to create, modify, and delete user accounts on the Celerra
Network Server Control Station. Control Station user accounts provide
administrative access to the Celerra and its resources.
300-002-686
Using Quotas on Celerra
Explains how to configure and manage disk usage limits (quotas) for users,
groups, and trees within a file system, and users and groups within a tree, on
Networking
Table 28
Table 28 lists networking information described in technical modules

on the Celerra Network Server Version 5.5.31 Documentation CD.
Networking
Part number
Name
Description
300-002-707
Configuring and Managing

Celerra Networking
Describes how to configure and manage a Celerra network, including network

devices and interfaces, address resolution, routing, and virtual local area
networks (VLANs). It also provides reference information on Packet Reflect
mode and configuring Linux service on Celerra.
300-002-706

Celerra Network High Availability
Explains how to configure a Celerra Network Server to enable high-availability

network communications links. High-availability network configurations allow
a Data Mover to respond to a failed link in a virtual device (an Ethernet
channel or link aggregation), or the failure of all links in virtual device, by
automatically rerouting traffic to the available links (virtual device failover) or to
the available component devices (Fail-Safe Networking). Topics also include
how to manage and troubleshoot a high-availability network environment.
300-002-709
Configuring Celerra Time

Services
Explains how to use Time Services, including the Network Time Protocol
(NTP) and Simple Network Time Protocol (SNTP) functions, to synchronize
the Data Movers with a network time source.
300-002-710
Using FTP on Celerra Network

Server
Describes the File Transfer Protocol (FTP) function, and how to configure FTP
on Celerra Network Server to enable the transfer of files from one machine to
another.
300-002-711
Using TFTP on Celerra Network

Server
Describes the Trivial File Transfer Protocol (TFTP) function, and how to
configure TFTP on Celerra Network Server to enable the transfer of files from
one machine to another.
115
Documentation
Volumes and file

systems
Table 29
116
Table 29 lists the file system information described in technical

CD.
Volumes and file systems
Part number
Name
Description
300-002-682
Using International
Character Sets with
Celerra
Describes when and how to enable a variety of international character sets

(also known as I18N) on the Celerra Network Server to serve local language
requirements. Topics include:
How to configure the Celerra Network Server to store one or more
international character sets
How to enable international character support for a newly installed Celerra
Network Server
How to upgrade an existing system to support international characters,
including file and directory name conversion
300-002-705
Managing Celerra Volumes

and File Systems Manually
Provides an overview of Celerra volumes and file systems. Contains

planning consideration information and explains the procedures to manually
create, configure, and manage Celerra volumes and file systems (including
nested mount file systems). Provides instructions for performing basic
Celerra file system administration tasks.
300-002-689
Managing Celerra Volumes

and File Systems with
Automatic Volume
Management
Explains how to configure disk volumes and create and extend file systems
with Automatic Volume Management (AVM). Provides a comprehensive
explanation of how AVM works with Automatic File System Extension. Also
provides planning consideration information and procedures for managing
AVM storage pools and other related file system administration tasks.
300-002-686
Using Quotas on Celerra
Explains how to configure and manage disk usage limits (quotas) for users,
groups, and trees within a file system, and users and groups within a tree, on
300-002-716
Controlling Access to
Celerra System Objects
Explains the Celerra Network Server's access control lists (ACLs) and
describes how to implement them to control the level of user access to
system objects.
300-002-675
Celerra File Extension

Filtering
Describes how Windows users can control the types of files that are saved to
a CIFS share or a Data Mover.
Documentation
Table 29
Volumes and file systems (continued)
Part number
Name
Description
300-002-703
Using Celerra FileMover
Explains how to enable FileMover on the Celerra Network Server to

automatically migrate infrequently used files to slower, less-expensive
storage devices. Discusses planning considerations, file migration and
retrieval processes, and procedures for performing administration tasks on
FileMover-enabled file systems.
300-002-704
Using File-Level Retention

on Celerra (formerly Using
Celerra WORM)
Provides an overview of File-Level Retention (FLR) storage technology. It

explains how to manage files in an FLR file system to create a permanent,
unalterable set of files and directories.
Copying file systems
Table 30
Table 30 lists the Celerra backup protection information described in

technical modules on the Celerra Network Server Version 5.5.31
Documentation CD.
Copying file systems
Part number
Name
Description
300-002-701
Using TimeFinder/FS, NearCopy,

and FarCopy with Celerra*
(formerly Using TImeFinder/FS
with Celerra and Using
TimeFinder/FS NearCopy and
FarCopy with Celerra)
Explains how to use TimeFinder/FS to create a snapshot (point-in-time copy)

of a production file system based on local volumes, or how to use
TimeFinder/FS NearCopy or FarCopy to create a snapshot of a file system
that resides on a remote Celerra Network Server. You can export this
read-only copy from a Data Mover on the remote Celerra Network Server and
use it for backup and application testing. TimeFinder/FS NearCopy can work
with SRDF/Synchronous (SRDF/S); TimeFinder/FS FarCopy can work with
SRDF adaptive copy write pending mode.
300-002-725
Using Celerra Replicator
Describes how to create and manage a point-in-time, read-only, replica of a

production (source) file system on a Celerra Network Server that can be used
by a Data Mover in the same cabinet or a Data Mover at a remote site. This
replica can be used for backup, data recovery, and application testing without
interrupting the processing of client transactions to and from the source file
system.
300-002-720
Using SnapSure on Celerra
Explains how to create and manage read-only, point-in-time, logical images

(checkpoints) of a production file system, and restore a PFS from a
checkpoint.
*Celerra TimeFinder/FS and Celerra TimeFinder/FS NearCopy and FarCopy are supported only on a Celerra Network Server
attached to a Symmetrix system.
117
Documentation
User management
Table 31
User management
Part number
Name
Online Help
Celerra UNIX Attributes Migration Explains how to migrate UNIX attributes from the NIS server or UNIX
Tool
password and group files, or remove user and group attributes from the Active
Directory.
Online Help
Celerra UNIX User Management
Explains how to view and modify Celerra UNIX user and group attributes.
300-002-713
Configuring Celerra Naming

Services
Explains how to use Naming Services, including local files, the Network
Information Service (NIS), Domain Name System (DNS), and the Sun Java
System Directory Server (iPlanet).
300-002-715
Configuring Celerra User

Mapping
Describes how to configure and manage Windows user ID mapping using

methods and tools such as Internal Usermapper, MMC snap-ins, local files,
and NIS. Recommends the technique to use based on whether you have a
Windows-only or multiprotocol environment.
300-002-714
Configuring External Usermapper Describes how to implement and manage External Usermapper, a Celerra
for Celerra
Network Service feature that automatically assigns UIDs and GIDs to
Windows users and groups.
300-002-719
Using NTMigrate with Celerra
CIFS configuration
Table 32
118
Table 31 lists user management information described in technical

modules and online help on the Celerra Network Server Version 5.5.31
Documentation CD or in the applications online help.
Description
Describes how to configure and manage NTMigrate, a tool that adds

Windows users to an existing UNIX UID/GID database.
Table 32 lists the Microsoft Windows-based product configuration

and support information described in technical modules on the
Celerra Network Server Version 5.5.31 Documentation CD.
CIFS configuration
Part number
Name
Description
300-002-678
Configuring CIFS on Celerra
Describes how to configure CIFS support on the Celerra Network Server for
Windows environments using the Celerra Manager and the command line
interface (CLI).
300-002-676
Managing Celerra for a

Multiprotocol Environment
Describes how to manage CIFS support on the Celerra Network Server for
mixed Windows and UNIX environments.
Note: To configure NFS services for file systems, refer to the Configuring NFS
on Celerra technical module.
300-002-679
Managing Celerra for the

Windows Environment
Describes how to manage CIFS support on the Celerra Network Server for
Windows environments.
Documentation
Table 32
CIFS configuration (continued)
Part number
Name
Description
300-002-715
Configuring Celerra User

Mapping
Describes how to configure and manage Windows user ID mapping using

methods and tools such as Internal Usermapper, MMC snap-ins, local files,
and NIS. Recommends the technique to use based on whether you have a
Windows-only or multiprotocol environment.
300-002-714
Configuring External Usermapper Describes how to implement and manage External Usermapper, a Celerra
for Celerra
Network Service feature that automatically assigns UIDs and GIDs to
Windows users and groups.
300-002-719
Using NTMigrate with Celerra
Describes how to configure and manage NTMigrate, a tool that adds

Windows users to an existing UNIX UID/GID database.
300-002-684
Using EMC Utilities for the CIFS

Environment
This technical module describes how to use EMC Windows utilities, which you
install on a Windows server from the Celerra Network Server Applications and
Tools CD, to facilitate the migration of files and directories from a Windows
server to the Celerra Network Server.
300-002-675
Celerra File Extension Filtering
Describes how Windows users can control the types of files that are saved to
a CIFS share or a Data Mover.
300-002-726
Configuring Virtual Data Movers

for Celerra
Describes how to configure and manage a Virtual Data Mover (VDM), which
enables you to administratively separate CIFS servers and their associated
resources, like file systems, into virtual containers, or VDMs. VDMs allow
administrative separation between groups of CIFS servers, enable replication
of CIFS environments, and allow the movement of CIFS servers from Data
Mover to Data Mover.
300-002-724
Replicating Celerra CIFS

Environments
Describes the planning and configuration tasks required to use the Celerra
functionality to deliver asynchronous data recovery for your CIFS
environments by leveraging Celerra Replicator.
300-002-680
Using Windows Administrative

Tools with Celerra
Contains instructions for using Microsoft Windows tools to perform the

following administrative tasks on a Celerra Network Server configured with the
CIFS protocol:
Create shares and groups on the Data Mover
Configure access policy for shares and groups on the Data Mover
Monitor Data Mover attributes at the session, share, and file level
Audit the CIFS users and objects
119
Documentation
NFS configuration
Table 33
Table 33 lists NFS information described in technical modules on the

NFS configuration
Part number
Name
Description
300-002-693
Configuring NFS on Celerra
Provides an overview of the Celerra NFS environment and explains how to

mount, export, and manage access to NFS file systems on the Celerra
Network Server.
Note: The Managing Celerra for a Multiprotocol Environment technical
module provides information about configuring the Celerra Network Server to
support both NFS and CIFS; the Configuring CIFS on Celerra technical
module provides information about CIFS only.
High availability
Table 34
Table 34 lists the high availability information described in technical

CD.
High availability
Part number
Name
Description
300-002-706

Celerra Network High
Availability
Explains how to configure a Celerra Network Server to enable high-availability

network communications links. High-availability network configurations allow
a Data Mover to respond to a failed link in a virtual device (an Ethernet
channel or link aggregation), or the failure of all links in virtual device, by
automatically rerouting traffic to the available links (virtual device failover) or to
the available component devices (Fail-Safe Networking). Topics also include
how to manage and troubleshoot a high-availability network environment.
300-002-699
Using SRDF/S with Celerra

for Disaster Recovery*
Explains how to configure source and destination Celerra Network Servers for
Symmetrix Remote Data Facility/Synchronous (SRDF/S), a limited-distance
data replication facility that provides complete disaster recovery.
300-002-698
Using SRDF/A with

Celerra*
Explains how to configure source and destination Celerra Network Servers for
Symmetrix Remote Data Facility/Asynchronous (SRDF/A), an
extended-distance asynchronous data replication facility that provides
continuous availability of file systems.
300-002-034
Configuring Standbys on
Celerra
Describes Control Station and Data Mover failover, and how to implement a
standby Control Station or standby Data Mover(s) on the Celerra Network
Server.
* Celerra SRDF is supported only on a Celerra Network Server attached to a Symmetrix system.
120
Documentation
Backup
Table 35
Table 35 lists the Celerra backup protection information described in

Documentation CD.
Backup
Part number
Name
Description
300-002-721
Configuring NDMP Backups to

Disk on Celerra
Describes how to create, configure, and manage a virtual tape library unit
(VTLU) on the Celerra Network Server. VTLUs allow you back up data to
disk-based storage by means of the Network Data Management Protocol
(NDMP).
300-002-722
Configuring NDMP Backups on

Celerra
Describes how to configure and manage the Celerra Network Server for
Network Data Management Protocol (NDMP) backups, and how to configure
your backup software for the Celerra Network Server.
300-002-723
Using the Celerra server_archive

Utility
Describes how to use the server_archive utility to perform manual,

local backup and restore operations, and to copy directories. As a backup
option, server_archive is for local, low-capacity backups.
Celerra iSCSI
Table 36
Table 36 lists the Celerra iSCSI information described in technical

modules and online help on the Celerra Network Server Version 5.5.31
Documentation CD or in the applications online help.
Celerra iSCSI
Part number
Name
Description
Online Help
iSCSI SnapSure Manager
Explains how to use Celerra iSCSI SnapSure Manager. This application is

available on a separate CD for ease of installation into a Windows
environment.
Online Help
iSCSI SnapSure Manager for

Exchange 2000
Explains how to use Celerra iSCSI SnapSure Manager for Exchange 2000.
This application is available on a separate CD for ease of installation into a
Windows environment.
300-002-695
Installing Celerra iSCSI Host

Components
Describes how to install Celerra VSS Provider for iSCSI, SnapSure Manager
for iSCSI, and iSCSI SnapSure Manager for Exchange 2000 on iSCSI hosts.
The technical module also covers installing the Microsoft iSCSI Software
Initiator and configuring Celerra iSCSI LUNs as disk devices in Windows.
300-002-694
Configuring iSCSI Targets on

Celerra
Describes how to configure iSCSI on the Celerra Network Server and covers
the creation of targets, LUNs, and LUN masks.
300-002-696
Using Celerra Replicator for

iSCSI
This technical module describes how to set up a Celerra Network Server for
asynchronous replication of iSCSI LUNs, and covers iSCSI replication
management functions performed on the Celerra Network Server.
121
Documentation
Antivirus
Table 37
Antivirus
Part number
Name
Description
Online Help
CAVA Calculator
Describes using the CAVA Calculator sizing tool to estimate the number of
CAVAs required to provide a user-defined level of performance in a CAVA
pool, based upon user information.
Online Help
Celerra AntiVirus Management
Explains how to modify CAVA virus-checking configuration parameters on a

Celerra Data Mover from a Windows 2000 or NT server.
300-002-685
Using Celerra AntiVirus Agent
Describes how to install and configure Celerra AntiVirus Agent (CAVA), which
provides an antivirus solution to clients using an EMC Celerra Network
Server in a Microsoft Windows Server 2003, Windows 2000, or Windows NT
domain. CAVA uses third-party antivirus software to identify and eliminate
known viruses before they infect files on the storage system.
Celerra MPFS
(formerly HighRoad)
Table 38
Table 38 lists the Celerra MPFS (MultiPath File System) information

described in technical modules on the Celerra Network Server Version
5.5.31 Documentation CD.
Celerra MPFS
Part number
Name
Description
300-002-728
Using MPFS on Celerra
This technical module explains how to provide EMC HighRoad MPFS clients
with access to Celerra file systems using the Celerra HighRoad feature.
Originating with metadata requests through the Celerra Network Server to the
storage device, MPFS delivers requested data over direct Fibre Channel or
iSCSI connections to the client.
Data migration
Table 39
122
Table 37 lists information for antivirus protection described in

technical modules and online help on the Celerra Network Server
Version 5.5.31 Documentation CD or in the applications online help.
Table 39 lists the Celerra data migration information described in

Documentation CD.
Data migration
Part number
Name
Description
300-002-712
Celerra CDMS Version 2.0 for

NFS and CIFS
Describes the planning considerations and procedures required to migrate

data to the Celerra Network Server in a CIFS or NFS environment.
Documentation
Table 39
Data migration (continued)
Part number
Name
Description
300-001-990
Using EMC Utilities for the CIFS

Environment
Describes how to use EMC Windows utilities, which you install on the a
Windows server from the Celerra Network Server Version 5.5 Applications
and Tools CD-ROM, to facilitate the migration of files and directories from a
Windows server to the Celerra Network Server. The utilities are:
LGDUP.exe duplicates the local groups database
EMCOPY.exe duplicates a directory tree from one server to another
with Windows security intact
EMCACL.exe modifies all the security properties of existing files and
directories on a Celerra Network Server
SHAREDUP.exe duplicates the shares from one Windows server to
another with the Windows security intact
Troubleshooting
Table 40
Table 40 lists troubleshooting information described in technical

CD.
Troubleshooting
Part number
Name
Description
300-002-690
Celerra Problem Resolution

Roadmap
Provides a generic description of how to troubleshoot Celerra system

problems and guidelines to use when contacting EMC Customer Support.
300-002-708
Configuring Celerra Events and

Notifications
Describes how to configure the Celerra Network Server to record and display
various system events, such as errors, alerts, and commands to facilitate
system administration and troubleshooting.
Reference
Table 41
Table 41 lists reference information described in technical modules on

the Celerra Network Server Version 5.5.31 Documentation CD.
Reference
Part number
Name
Description
300-002-697
Celerra Network Server

Command Reference Manual
Details the command set, including the command function, storage

application, syntax, usage prerequisites, and precautions. This manual is also
available in hard copy in the product kit, and its command reference
information is available online as man pages that you can access from the
CLI.
300-002-688
Celerra Network Server Error

Messages Guide
Lists and describes Celerra Network Server error messages. It covers how
error codes are presented, methods you can use to look them up, and where
possible, directs you to scripts, online resources, or other tools where you can
learn more about each error message.
123
Installation
Table 41
Reference (continued)
Part number
Name
Description
300-002-691
Celerra Network Server

Parameters Guide
Lists and describes specific system attributes that are set up by default in the
Celerra Network Server software. Describes how to establish, modify, or
override some of these attributes for the entire Celerra Network Server or for a
specific Data Mover.
300-002-687
Celerra Network Server User

Information Glossary
This is a glossary of network-attached storage (NAS) terminology used in

reference to the Celerra Network Server. Terms that are specific to a
particular product are designated as such in the definition.
300-002-690
Celerra Problem Resolution

Roadmap
Provides a generic description of how to troubleshoot Celerra system

problems and guidelines to use when contacting EMC Customer Service.
Installation
EMC or its partners performs the installation of Celerra Network
Server hardware and software. Please contact your EMC Support
Representative for detailed information about the installation
procedure.
Celerra installation
process
How to prepare for
Celerra installation
This section lists some facts you should know about the Celerra
Network Server and the installation and configuration process before
proceeding further.
To prepare for Celerra installation and configuration, complete the
following tasks:
1. Perform the pre-site installation tasks as required by EMC to
complete the installation and connectivity of the Celerra Network
Server. Tasks include preparing electrical facilities, wiring the
network, adding a telephone line, checking environmental
conditions and related requirements, as stated in the Professional
Services Statement of Work (PSSOW).
For a copy of the PS-SOW document, contact your Professional Services
Account Team. For examples of Celerra network topologies, refer to the
Configuring and Managing Celerra Networking technical module on the
2. Configure your environment for UNIX clients, Microsoft

Windows clients, or a combination of Windows and UNIX
clients, as applicable.
124
Installation
3. Complete network assessment to ensure proper performance and

protocol configurations. (Optional EMC services can be ordered
to assist with this process.)
Celerra Network
Server installation and
configuration
EMC Customer Service installs the Celerra Network Server

hardwarea process that includes the physical setup, cabling, power
application, and diagnostic testing of the Celerra Network Server.
Next, EMC Professional Services or an EMC partner might configure
and integrate the Celerra Network Server. For example, for a Celerra
Network Server with a Symmetrix system, this process typically
includes the following tasks:
Coordinating the activities to enable installation, cabling, and

connectivity to system cabinets
Loading and configuring the Celerra Network Server software
Creating, mounting, and exporting file systems
Configuring network protocols
Integrating with a supported, existing user authentication

environment
Configuring standby Data Mover(s), if applicable
Configuring and demonstrating system management and

administration tools
EMC Professional Services can perform additional Celerra Network

Server configuration tasks, such as host provisioning (with Celerra
MPFS), backup protection, and network connectivity, depending on
the individual EMC customer agreement defined in the PSSOW.
These services are part of the EMC Network-Attached Storage Design
and Implementation service offering, which is designed to provide
continuous technical expertise as you manage your organizations
growing network storage needs. For more information about this
service offering, contact your EMC Customer Support representative.
After EMC installs, configures, and tests the Celerra Network Server,
it is fully operational and ready to support networked storage
operations.
125
Installation
Reboot required at
upgrade
CNS software
The Celerra Network Server (CNS) software is comprised of two

installation components, a Linux package and a CNS software
package. The software currently running on the Celerra and the CNS
upgrade software versions determine the number of Control Station
and Data Mover reboots required and whether Data Mover reboots
can be scheduled to a later time after the upgrade is complete. The
DART software runs on the Data Movers and is part of the CNS
software package. The new version of DART is loaded and run on the
Data Movers only after a Data Mover boot.
Control Station
rebooting
When upgrading CNS software, if either of the first two digits of the
new version number change (for example, 5.3.x to 5.5.x), the Control
Station requires rebooting. The setup upgrade script asks if you wish
to continue with the following question: Continuing with this
install will require a reboot. Do you wish to
continue? [yes or no].
If you enter no, the upgrade terminates immediately and you are
returned to the root prompt.
Data Mover rebooting
The Data Movers require a reboot in order to run the new version of
DART. When upgrading CNS software, if either of the first two digits
of the new version number change (i.e., 5.3.x to 5.5.x), you may not
schedule or delay the rebooting of the Data Movers. The setup
upgrade script asks if you wish to continue with the following
question.
!!! PLEASE READ CAREFULLY !!!
All servers will need to be rebooted to run the new
release. The upgrade will automatically reboot all
servers. Do you wish to continue with the upgrade? [yes
or no]:
If you enter no , the upgrade terminates immediately and you are

When upgrading DART, if both of the first two digits of the new
version of software remain the same (for example, 5.3.x to 5.3.y), you
may delay the rebooting of the Data Movers.
The setup upgrade script asks if you wish to continue with the
following question:
126
Installation
!!! PLEASE READ CAREFULLY !!! All servers will need to be

rebooted to run the new release. You will have the
opportunity to reboot all servers during the upgrade or,
you may choose to reboot each server individually after
the upgrade has completed. Do you wish to continue with
the upgrade? [yes or no]:
If you enter no , the upgrade terminates immediately and you are

After the CNS software package is upgraded, the setup script asks if
you wish to reboot the Data Movers now. If you enter yes, the Data
Movers are rebooted as part of the CNS software upgrade. If you
enter no, the Data Movers continue running the current
(pre-upgrade) version of CNS software until they are rebooted.
Licensing terms
Apache License,
Version 2.0
Copyright 1999-2006, Apache Software Foundation.

Licensed under the Apache License, Version 2.0.
TERMS AND CONDITIONS FOR USE, REPRODUCTION, AND
DISTRIBUTION
Definitions
"License" shall mean the terms and conditions for use, reproduction,
and distribution as defined by Sections 1 through 9 of this document.
"Licensor" shall mean the copyright owner or entity authorized by
the copyright owner that is granting the License.
"Legal Entity" shall mean the union of the acting entity and all other
entities that control, are controlled by, or are under common control
with that entity. For the purposes of this definition, "control" means
(i) the power, direct or indirect, to cause the direction or management
of such entity, whether by contract or otherwise, or (ii) ownership of
fifty percent (50%) or more of the outstanding shares, or (iii)
beneficial ownership of such entity.
"You" (or "Your") shall mean an individual or Legal Entity exercising
permissions granted by this License.
"Source" form shall mean the preferred form for making
modifications, including but not limited to software source code,
documentation source, and configuration files.
127
Installation
"Object" form shall mean any form resulting from mechanical

transformation or translation of a Source form, including but not
limited to compiled object code, generated documentation, and
conversions to other media types.
"Work" shall mean the work of authorship, whether in Source or
Object form, made available under the License, as indicated by a
copyright notice that is included in or attached to the work (an
example is provided in the Appendix below).
"Derivative Works" shall mean any work, whether in Source or Object
form, that is based on (or derived from) the Work and for which the
editorial revisions, annotations, elaborations, or other modifications
represent, as a whole, an original work of authorship. For the
purposes of this License, Derivative Works shall not include works
that remain separable from, or merely link (or bind by name) to the
interfaces of, the Work and Derivative Works thereof.
"Contribution" shall mean any work of authorship, including the
original version of the Work and any modifications or additions to
that Work or Derivative Works thereof, that is intentionally submitted
to Licensor for inclusion in the Work by the copyright owner or by an
individual or Legal Entity authorized to submit on behalf of the
copyright owner. For the purposes of this definition, "submitted"
means any form of electronic, verbal, or written communication sent
to the Licensor or its representatives, including but not limited to
communication on electronic mailing lists, source code control
systems, and issue tracking systems that are managed by, or on behalf
of, the Licensor for the purpose of discussing and improving the
Work, but excluding communication that is conspicuously marked or
otherwise designated in writing by the copyright owner as "Not a
Contribution."
"Contributor" shall mean Licensor and any individual or Legal Entity
on behalf of whom a Contribution has been received by Licensor and
subsequently incorporated within the Work.
Grant of Copyright License
Subject to the terms and conditions of this License, each Contributor
hereby grants to You a perpetual, worldwide, non-exclusive,
no-charge, royalty-free, irrevocable copyright license to reproduce,
prepare Derivative Works of, publicly display, publicly perform,
sublicense, and distribute the Work and such Derivative Works in
Source or Object form.
128
Installation
Grant of Patent License.

Subject to the terms and conditions of this License, each Contributor
hereby grants to You a perpetual, worldwide, non-exclusive,
no-charge, royalty-free, irrevocable (except as stated in this section)
patent license to make, have made, use, offer to sell, sell, import, and
otherwise transfer the Work, where such license applies only to those
patent claims licensable by such Contributor that are necessarily
infringed by their Contribution(s) alone or by combination of their
Contribution(s) with the Work to which such Contribution(s) was
submitted. If You institute patent litigation against any entity
(including a cross-claim or counterclaim in a lawsuit) alleging that
the Work or a Contribution incorporated within the Work constitutes
direct or contributory patent infringement, then any patent licenses
granted to You under this License for that Work shall terminate as of
the date such litigation is filed.
Redistribution
You may reproduce and distribute copies of the Work or Derivative
Works thereof in any medium, with or without modifications, and in
Source or Object form, provided that You meet the following
conditions:
You must give any other recipients of the Work or Derivative

Works a copy of this License; and
You must cause any modified files to carry prominent notices

stating that You changed the files; and
You must retain, in the Source form of any Derivative Works that
You distribute, all copyright, patent, trademark, and attribution
notices from the Source form of the Work, excluding those notices
that do not pertain to any part of the Derivative Works; and
If the Work includes a "NOTICE" text file as part of its

distribution, then any Derivative Works that You distribute must
include a readable copy of the attribution notices contained
within such NOTICE file, excluding those notices that do not
pertain to any part of the Derivative Works, in at least one of the
following places: within a NOTICE text file distributed as part of
the Derivative Works; within the Source form or documentation,
if provided along with the Derivative Works; or, within a display
generated by the Derivative Works, if and wherever such
third-party notices normally appear. The contents of the NOTICE
file are for informational purposes only and do not modify the
License.
129
Installation
You may add Your own attribution notices within Derivative

Works that You distribute, alongside or as an addendum to the
NOTICE text from the Work, provided that such additional
attribution notices cannot be construed as modifying the License.
You may add Your own copyright statement to Your

modifications and may provide additional or different license
terms and conditions for use, reproduction, or distribution of
Your modifications, or for any such Derivative Works as a whole,
provided Your use, reproduction, and distribution of the Work
otherwise complies with the conditions stated in this License.
Submission of Contributions
Unless You explicitly state otherwise, any Contribution intentionally
submitted for inclusion in the Work by You to the Licensor shall be
under the terms and conditions of this License, without any
additional terms or conditions. Notwithstanding the above, nothing
herein shall supersede or modify the terms of any separate license
agreement you may have executed with Licensor regarding such
Contributions.
Trademarks
This License does not grant permission to use the trade names,
trademarks, service marks, or product names of the Licensor, except
as required for reasonable and customary use in describing the origin
of the Work and reproducing the content of the NOTICE file.
Disclaimer of Warranty
Unless required by applicable law or agreed to in writing, Licensor
provides the Work (and each Contributor provides its Contributions)
on an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS
OF ANY KIND, either express or implied, including, without
limitation, any warranties or conditions of TITLE,
NON-INFRINGEMENT, MERCHANTABILITY, or FITNESS FOR A
PARTICULAR PURPOSE. You are solely responsible for determining
the appropriateness of using or redistributing the Work and assume
any risks associated with Your exercise of permissions under this
License.
130
Installation
Limitation of Liability
In no event and under no legal theory, whether in tort (including
negligence), contract, or otherwise, unless required by applicable law
(such as deliberate and grossly negligent acts) or agreed to in writing,
shall any Contributor be liable to You for damages, including any
direct, indirect, special, incidental, or consequential damages of any
character arising as a result of this License or out of the use or
inability to use the Work (including but not limited to damages for
loss of goodwill, work stoppage, computer failure or malfunction, or
any and all other commercial damages or losses), even if such
Contributor has been advised of the possibility of such damages.
Accepting Warranty or Additional Liability
While redistributing the Work or Derivative Works thereof, You may
choose to offer, and charge a fee for, acceptance of support, warranty,
indemnity, or other liability obligations and/or rights consistent with
this License. However, in accepting such obligations, You may act
only on Your own behalf and on Your sole responsibility, not on
behalf of any other Contributor, and only if You agree to indemnify,
defend, and hold each Contributor harmless for any liability incurred
by, or claims asserted against, such Contributor by reason of your
accepting any such warranty or additional liability.
Ganymed
Copyright (c) 2005 - 2006 Swiss Federal Institute of Technology (ETH

Zurich), Department of Computer Science (http://www.inf.ethz.ch),
Christian Plattner. All rights reserved.
Redistribution and use in source and binary forms, with or without
modification, are permitted provided that the following conditions re
met:
a.) Redistributions of source code must retain the above copyright
notice, this list of conditions and the following disclaimer.
b.) Redistributions in binary form must reproduce the above
copyright notice, this list of conditions and the following
disclaimer in the documentation and/or other materials provided
with the distribution.
c.) Neither the name of ETH Zurich nor the names of its contributors
may be used to endorse or promote products derived from this
software without specific prior written permission.
131
Installation
THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS

AND CONTRIBUTORS "AS IS" AND ANY EXPRESS OR IMPLIED
WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS
FOR A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT
SHALL THE COPYRIGHT OWNER OR CONTRIBUTORS BE
LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL,
EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING,
BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS
INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY
OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR
TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN
ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF
ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
The Java implementations of the AES, Blowfish and 3DES ciphers
have been taken (and slightly modified) from the cryptography
package released by "The Legion Of The Bouncy Castle".
Their license states the following:
Copyright (c) 2000 - 2004 The Legion Of The Bouncy Castle
(http://www.bouncycastle.org)
Permission is hereby granted, free of charge, to any person obtaining
a copy of this software and associated documentation files (the
"Software"), to deal in the Software without restriction, including
without limitation the rights to use, copy, modify, merge, publish,
distribute, sublicense, and/or sell copies of the Software, and to
permit persons to whom the Software is furnished to do so, subject to
the following conditions:
The above copyright notice and this permission notice shall be
included in all copies or substantial portions of the Software.
THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY
OF ANY KIND, EXPRESS ORIMPLIED, INCLUDING BUT NOT
LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
FITNESS FOR A PARTICULAR PURPOSE AND
NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR
COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES
OR OTHER LIABILITY, WHETHER IN AN ACTION OF
CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR
IN CONNECTION WITH THE SOFTWARE OR THE USE OR
OTHER DEALINGS IN THE SOFTWARE.
132
Installation
GNU GENERAL PUBLIC

LICENSE
Version 2, June 1991

Copyright (C) 1989, 1991 Free Software Foundation, Inc.
51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA
Everyone is permitted to copy and distribute verbatim copies of this
license document, but changing it is not allowed.
Preamble
The licenses for most software are designed to take away your
freedom to share and change it. By contrast, the GNU General Public
License is intended to guarantee your freedom to share and change
free software--to make sure the software is free for all its users. This
General Public License applies to most of the Free Software
Foundation's software and to any other program whose authors
commit to using it. (Some other Free Software Foundation software is
covered by the GNU Lesser General Public License instead.) You can
apply it to your programs, too.
When we speak of free software, we are referring to freedom, not
price. Our General Public Licenses are designed to make sure that
you have the freedom to distribute copies of free software (and
charge for this service if you wish), that you receive source code or
can get it if you want it, that you can change the software or use
pieces of it in new free programs; and that you know you can do
these things.
To protect your rights, we need to make restrictions that forbid
anyone to deny you these rights or to ask you to surrender the rights.
These restrictions translate to certain responsibilities for you if you
distribute copies of the software, or if you modify it.
For example, if you distribute copies of such a program, whether
gratis or for a fee, you must give the recipients all the rights that you
have. You must make sure that they, too, receive or can get the source
code. And you must show them these terms so they know their
rights.
We protect your rights with two steps:
(1) copyright the software, and
(2) offer you this license which gives you legal permission to copy,
distribute and/or modify the software.
Also, for each author's protection and ours, we want to make certain
that everyone understands that there is no warranty for this free
software. If the software is modified by someone else and passed on,
we want its recipients to know that what they have is not the original,
so that any problems introduced by others will not reflect on the
original authors' reputations.
133
Installation
Finally, any free program is threatened constantly by software

patents. We wish to avoid the danger that redistributors of a free
program will individually obtain patent licenses, in effect making the
program proprietary. To prevent this, we have made it clear that any
patent must be licensed for everyone's free use or not licensed at all.
The precise terms and conditions for copying, distribution and
modification follow.
TERMS AND CONDITIONS FOR COPYING, DISTRIBUTION AND
MODIFICATION
0. This License applies to any program or other work which
contains a notice placed by the copyright holder saying it may be
distributed under the terms of this General Public License. The
"Program", below, refers to any such program or work, and a
"work based on the Program" means either the Program or any
derivative work under copyright law: that is to say, a work
containing the Program or a portion of it, either verbatim or with
modifications and/or translated into another language.
(Hereinafter, translation is included without limitation in the
term "modification".) Each licensee is addressed as "you".
Activities other than copying, distribution and modification are
not covered by this License; they are outside its scope. The act of
running the Program is not restricted, and the output from the
Program is covered only if its contents constitute a work based on
the Program (independent of having been made by running the
Program). Whether that is true depends on what the Program
does.
1. You may copy and distribute verbatim copies of the Program's
source code as you receive it, in any medium, provided that you
conspicuously and appropriately publish on each copy an
appropriate copyright notice and disclaimer of warranty; keep
intact all the notices that refer to this License and to the absence
of any warranty; and give any other recipients of the Program a
copy of this License along with the Program.
You may charge a fee for the physical act of transferring a copy,
and you may at your option offer warranty protection in
exchange for a fee.
134
Installation
2. You may modify your copy or copies of the Program or any

portion of it, thus forming a work based on the Program, and
copy and distribute such modifications or work under the terms
of Section 1 above, provided that you also meet all of these
conditions:
a. You must cause the modified files to carry prominent notices
stating that you changed the files and the date of any change.
b. You must cause any work that you distribute or publish, that in
whole or in part contains or is derived from the Program or any
part thereof, to be licensed as a whole at no charge to all third
parties under the terms of this License.
c. If the modified program normally reads commands
interactively when run, you must cause it, when started
running for such interactive use in the most ordinary way, to
print or display an announcement including an appropriate
copyright notice and a notice that there is no warranty (or else,
saying that you provide a warranty) and that users may
redistribute the program under these conditions, and telling the
user how to view a copy of this License. (Exception: if the
Program itself is interactive but does not normally print such an
announcement, your work based on the Program is not
required to print an announcement.)
These requirements apply to the modified work as a whole. If
identifiable sections of that work are not derived from the
Program, and can be reasonably considered independent and
separate works in themselves, then this License, and its terms, do
not apply to those sections when you distribute them as separate
works. But when you distribute the same sections as part of a
whole which is a work based on the Program, the distribution of
the whole must be on the terms of this License, whose
permissions for other licensees extend to the entire whole, and
thus to each and every part regardless of who wrote it.
Thus, it is not the intent of this section to claim rights or contest
your rights to work written entirely by you; rather, the intent is to
exercise the right to control the distribution of derivative or
collective works based on the Program.
In addition, mere aggregation of another work not based on the
Program with the Program (or with a work based on the
Program) on a volume of a storage or distribution medium does
not bring the other work under the scope of this license.
135
Installation
3. You may copy and distribute the Program (or a work based on it,
under Section 2) in object code or executable form under the
terms of Sections 1 and 2 above provided that you also do one of
the following:
a) Accompany it with the complete corresponding
machine-readable source code, which must be distributed under
the terms of Sections 1 and 2 above on a medium customarily
used for software interchange; or,
b) Accompany it with a written offer, valid for at least three years,
to give any third party, for a charge no more than your cost of
physically performing source distribution, a complete
machine-readable copy of the corresponding source code, to be
distributed under the terms of Sections 1 and 2 above on a
medium customarily used for software interchange; or,
c) Accompany it with the information you received as to the offer
to distribute corresponding source code. (This alternative is
allowed only for noncommercial distribution and only if you
received the program in object code or executable form with such
an offer, in accord with Subsection b above.)
The source code for a work means the preferred form of the work
for making modifications to it. For an executable work, complete
source code means all the source code for all modules it contains,
plus any associated interface definition files, plus the scripts used
to control compilation and installation of the executable.
However, as a special exception, the source code distributed need
not include anything that is normally distributed (in either source
or binary form) with the major components (compiler, kernel, and
so on) of the operating system on which the executable runs,
unless that component itself accompanies the executable.
If distribution of executable or object code is made by offering
access to copy from a designated place, then offering equivalent
access to copy the source code from the same place counts as
distribution of the source code, even though third parties are not
compelled to copy the source along with the object code.
4. You may not copy, modify, sublicense, or distribute the Program
except as expressly provided under this License. Any attempt
otherwise to copy, modify, sublicense or distribute the Program is
void, and will automatically terminate your rights under this
License. However, parties who have received copies, or rights,
from you under this License will not have their licenses
terminated so long as such parties remain in full compliance.
136
Installation
5. You are not required to accept this License, since you have not
signed it. However, nothing else grants you permission to modify
or distribute the Program or its derivative works. These actions
are prohibited by law if you do not accept this License. Therefore,
by modifying or distributing the Program (or any work based on
the Program), you indicate your acceptance of this License to do
so, and all its terms and conditions for copying, distributing or
modifying the Program or works based on it.
6. Each time you redistribute the Program (or any work based on
the Program), the recipient automatically receives a license from
the original licensor to copy, distribute or modify the Program
subject to these terms and conditions. You may not impose any
further restrictions on the recipients' exercise of the rights
granted herein. You are not responsible for enforcing compliance
by third parties to this License.
7. If, as a consequence of a court judgment or allegation of patent
infringement or for any other reason (not limited to patent
issues), conditions are imposed on you (whether by court order,
agreement or otherwise) that contradict the conditions of this
License, they do not excuse you from the conditions of this
License. If you cannot distribute so as to satisfy simultaneously
your obligations under this License and any other pertinent
obligations, then as a consequence you may not distribute the
Program at all. For example, if a patent license would not permit
royalty-free redistribution of the Program by all those who
receive copies directly or indirectly through you, then the only
way you could satisfy both it and this License would be to refrain
entirely from distribution of the Program.
If any portion of this section is held invalid or unenforceable
under any particular circumstance, the balance of the section is
intended to apply and the section as a whole is intended to apply
in other circumstances.
It is not the purpose of this section to induce you to infringe any
patents or other property right claims or to contest validity of any
such claims; this section has the sole purpose of protecting the
integrity of the free software distribution system, which is
implemented by public license practices. Many people have made
generous contributions to the wide range of software distributed
through that system in reliance on consistent application of that
system; it is up to the author/donor to decide if he or she is
willing to distribute software through any other system and a
licensee cannot impose that choice.
137
Installation
This section is intended to make thoroughly clear what is

believed to be a consequence of the rest of this License.
8. If the distribution and/or use of the Program is restricted in
certain countries either by patents or by copyrighted interfaces,
the original copyright holder who places the Program under this
License may add an explicit geographical distribution limitation
excluding those countries, so that distribution is permitted only
in or among countries not thus excluded. In such case, this
License incorporates the limitation as if written in the body of
this License.
9. The Free Software Foundation may publish revised and/or new
versions of the General Public License from time to time. Such
new versions will be similar in spirit to the present version, but
may differ in detail to address new problems or concerns.
Each version is given a distinguishing version number. If the
Program specifies a version number of this License which applies
to it and "any later version", you have the option of following the
terms and conditions either of that version or of any later version
published by the Free Software Foundation. If the Program does
not specify a version number of this License, you may choose any
version ever published by the Free Software Foundation.
10. If you wish to incorporate parts of the Program into other free
programs whose distribution conditions are different, write to the
author to ask for permission. For software which is copyrighted
by the Free Software Foundation, write to the Free Software
Foundation; we sometimes make exceptions for this. Our
decision will be guided by the two goals of preserving the free
status of all derivatives of our free software and of promoting the
sharing and reuse of software generally.
NO WARRANTY
11. BECAUSE THE PROGRAM IS LICENSED FREE OF CHARGE,
THERE IS NO WARRANTY FOR THE PROGRAM, TO THE
EXTENT PERMITTED BY APPLICABLE LAW. EXCEPT WHEN
OTHERWISE STATED IN WRITING THE COPYRIGHT
HOLDERS AND/OR OTHER PARTIES PROVIDE THE
PROGRAM "AS IS" WITHOUT WARRANTY OF ANY KIND,
EITHER EXPRESSED OR IMPLIED, INCLUDING, BUT NOT
LIMITED TO, THE IMPLIED WARRANTIES OF
MERCHANTABILITY AND FITNESS FOR A PARTICULAR
PURPOSE. THE ENTIRE RISK AS TO THE QUALITY AND
PERFORMANCE OF THE PROGRAM IS WITH YOU.
138
Installation
SHOULD THE PROGRAM PROVE DEFECTIVE, YOU ASSUME

THE COST OF ALL NECESSARY SERVICING, REPAIR OR
CORRECTION.
12. IN NO EVENT UNLESS REQUIRED BY APPLICABLE LAW OR
AGREED TO IN WRITING WILL ANY COPYRIGHT HOLDER,
OR ANY OTHER PARTY WHO MAY MODIFY AND/OR
REDISTRIBUTE THE PROGRAM AS PERMITTED ABOVE, BE
LIABLE TO YOU FOR DAMAGES, INCLUDING ANY
GENERAL, SPECIAL, INCIDENTAL OR CONSEQUENTIAL
DAMAGES ARISING OUT OF THE USE OR INABILITY TO USE
THE PROGRAM (INCLUDING BUT NOT LIMITED TO LOSS
OF DATA OR DATA BEING RENDERED INACCURATE OR
LOSSES SUSTAINED BY YOU OR THIRD PARTIES OR A
FAILURE OF THE PROGRAM TO OPERATE WITH ANY
OTHER PROGRAMS), EVEN IF SUCH HOLDER OR OTHER
PARTY HAS BEEN ADVISED OF THE POSSIBILITY OF SUCH
DAMAGES.
END OF TERMS AND CONDITIONS
How to Apply These Terms to Your New Programs
If you develop a new program, and you want it to be of the greatest
possible use to the public, the best way to achieve this is to make it
free software which everyone can redistribute and change under
these terms.
To do so, attach the following notices to the program. It is safest to
attach them to the start of each source file to most effectively convey
the exclusion of warranty; and each file should have at least the
"copyright" line and a pointer to where the full notice is found.
one line to give the program's name and an idea of what it does.
Copyright (C) yyyy name of author
This program is free software; you can redistribute it and/or modify
it under the terms of the GNU General Public License as published by
the Free Software Foundation; either version 2 f the License, or (at
your option) any later version.
This program is distributed in the hope that it will be useful, but
WITHOUT ANY WARRANTY; without even the implied warranty of
MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.
See the GNU General Public License for more details.
139
Installation
You should have received a copy of the GNU General Public License
along with this program; if not, write to the Free Software
Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA
02110-1301, USA.
Also add information on how to contact you by electronic and paper
mail. If the program is interactive, make it output a short notice like
this when it starts in an interactive mode:
Gnomovision version 69, Copyright (C) year name of author
Gnomovision comes with ABSOLUTELY NO WARRANTY; for
details type `show w'. This is free software, and you are welcome to
redistribute it under certain conditions; type `show c' for details.
The hypothetical commands `show w' and `show c' should show the
appropriate parts of the General Public License. Of course, the
commands you use may be called something other than `show w' and
`show c'; they could even be mouse-clicks or menu items--whatever
suits your program.
You should also get your employer (if you work as a programmer) or
your school, if any, to sign a "copyright disclaimer" for the program, if
necessary. Here is a sample; alter the names:
Yoyodyne, Inc., hereby disclaims all copyright interest in the program
`Gnomovision' (which makes passes at compilers) written by James
Hacker.
signature of Ty Coon, 1 April 1989
Ty Coon, President of Vice
This General Public License does not permit incorporating your
program into proprietary programs. If your program is a subroutine
library, you may consider it more useful to permit linking proprietary
applications with the library. If this is what you want to do, use the
GNU Lesser General Public License instead of this License.
GNU LESSER GENERAL
PUBLIC LICENSE
Version 2.1, February 1999

Copyright (C) 1991, 1999 Free Software Foundation, Inc.
51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
Everyone is permitted to copy and distribute verbatim copies of this
license document, but changing it is not allowed.
[This is the first released version of the Lesser GPL. It also counts as
the successor of the GNU Library Public License, version 2, hence the
version number 2.1.]
140
Installation
Preamble
The licenses for most software are designed to take away your
freedom to share and change it. By contrast, the GNU General Public
Licenses are intended to guarantee your freedom to share and change
free software--to make sure the software is free for all its users.
This license, the Lesser General Public License, applies to some
specially designated software packages--typically libraries--of the
Free Software Foundation and other authors who decide to use it.
You can use it too, but we suggest you first think carefully about
whether this license or the ordinary General Public License is the
better strategy to use in any particular case, based on the
explanations below.
When we speak of free software, we are referring to freedom of use,
not price. Our General Public Licenses are designed to make sure that
you have the freedom to distribute copies of free software (and
charge for this service if you wish); that you receive source code or
can get it if you want it; that you can change the software and use
pieces of it in new free programs; and that you are informed that you
can do these things.
To protect your rights, we need to make restrictions that forbid
distributors to deny you these rights or to ask you to surrender these
rights. These restrictions translate to certain responsibilities for you if
you distribute copies of the library or if you modify it.
For example, if you distribute copies of the library, whether gratis or
for a fee, you must give the recipients all the rights that we gave you.
You must make sure that they, too, receive or can get the source code.
If you link other code with the library, you must provide complete
object files to the recipients, so that they can relink them with the
library after making changes to the library and recompiling it. And
you must show them these terms so they know their rights.
We protect your rights with a two-step method: (1) we copyright the
library, and (2) we offer you this license, which gives you legal
permission to copy, distribute and/or modify the library.
To protect each distributor, we want to make it very clear that there is
no warranty for the free library. Also, if the library is modified by
someone else and passed on, the recipients should know that what
they have is not the original version, so that the original author's
reputation will not be affected by problems that might be introduced
by others.
141
Installation
Finally, software patents pose a constant threat to the existence of any

free program. We wish to make sure that a company cannot
effectively restrict the users of a free program by obtaining a
restrictive license from a patent holder. Therefore, we insist that any
patent license obtained for a version of the library must be consistent
with the full freedom of use specified in this license.
Most GNU software, including some libraries, is covered by the
ordinary GNU General Public License. This license, the GNU Lesser
General Public License, applies to certain designated libraries, and is
quite different from the ordinary General Public License. We use this
license for certain libraries in order to permit linking those libraries
into non-free programs.
When a program is linked with a library, whether statically or using a
shared library, the combination of the two is legally speaking a
combined work, a derivative of the original library. The ordinary
General Public License therefore permits such linking only if the
entire combination fits its criteria of freedom. The Lesser General
Public License permits more lax criteria for linking other code with
the library.
We call this license the "Lesser" General Public License because it
does Less to protect the user's freedom than the ordinary General
Public License. It also provides other free software developers Less of
an advantage over competing non-free programs. These
disadvantages are the reason we use the ordinary General Public
License for many libraries. However, the Lesser license provides
advantages in certain special circumstances.
For example, on rare occasions, there may be a special need to
encourage the widest possible use of a certain library, so that it
becomes a de-facto standard. To achieve this, non-free programs must
be allowed to use the library. A more frequent case is that a free
library does the same job as widely used non-free libraries.
In this case, there is little to gain by limiting the free library to free
software only, so we use the Lesser General Public License.
In other cases, permission to use a particular library in non-free
programs enables a greater number of people to use a large body of
free software. For example, permission to use the GNU C Library in
non-free programs enables many more people to use the whole GNU
operating system, as well as its variant, the GNU/Linux operating
system.
142
Installation
Although the Lesser General Public License is Less protective of the

users' freedom, it does ensure that the user of a program that is linked
with the Library has the freedom and the wherewithal to run that
program using a modified version of the Library.
The precise terms and conditions for copying, distribution and
modification follow. Pay close attention to the difference between a
"work based on the library" and a "work that uses the library". The
former contains code derived from the library, whereas the latter
must be combined with the library in order to run.
TERMS AND CONDITIONS FOR COPYING, DISTRIBUTION AND
MODIFICATION
0. This License Agreement applies to any software library or other
program which contains a notice placed by the copyright holder
or other authorized party saying it may be distributed under the
terms of this Lesser General Public License (also called "this
License"). Each licensee is addressed as "you".
A "library" means a collection of software functions and/or data
prepared so as to be conveniently linked with application
programs (which use some of those functions and data) to form
executables.
The "Library", below, refers to any such software library or work
which has been distributed under these terms. A "work based on
the Library" means either the Library or any derivative work
under copyright law: that is to say, a work containing the Library
or a portion of it, either verbatim or with modifications and/or
translated straightforwardly into another language. (Hereinafter,
translation is included without limitation in the term
"modification".)
"Source code" for a work means the preferred form of the work
for making modifications to it. For a library, complete source code
means all the source code for all modules it contains, plus any
associated interface definition files, plus the scripts used to
control compilation and installation of the library.
Activities other than copying, distribution and modification are
not covered by this License; they are outside its scope. The act of
running a program using the Library is not restricted, and output
from such a program is covered only if its contents constitute a
work based on the Library (independent of the use of the Library
in a tool for writing it). Whether that is true depends on what the
Library does and what the program that uses the Library does.
143
Installation
1. You may copy and distribute verbatim copies of the Library's

complete source code as you receive it, in any medium, provided
that you conspicuously and appropriately publish on each copy
an appropriate copyright notice and disclaimer of warranty; keep
intact all the notices that refer to this License and to the absence
of any warranty; and distribute a copy of this License along with
the Library.
You may charge a fee for the physical act of transferring a copy,
and you may at your option offer warranty protection in
exchange for a fee.
2. You may modify your copy or copies of the Library or any
portion of it, thus forming a work based on the Library, and copy
and distribute such modifications or work under the terms of
Section 1 above, provided that you also meet all of these
conditions:
"a) The modified work must itself be a software library.
"b) You must cause the files modified to carry prominent notices
stating that you changed the files and the date of any change.
"c) You must cause the whole of the work to be licensed at no
charge to all third parties under the terms of this License.
"d) If a facility in the modified Library refers to a function or a
table of data to be supplied by an application program that
uses the facility, other than as an argument passed when the
facility is invoked, then you must make a good faith effort to
ensure that, in the event an application does not supply such
function or table, the facility still operates, and performs
whatever part of its purpose remains meaningful.
For example, a function in a library to compute square roots
has a purpose that is entirely well-defined independent of the
application. Therefore, Subsection 2d requires that any
application-supplied function or table used by this function
must be optional: if the application does not supply it, the
square root function must still compute square roots.)
These requirements apply to the modified work as a whole. If
identifiable sections of that work are not derived from the
Library, and can be reasonably considered independent and
separate works in themselves, then this License, and its terms,
do not apply to those sections when you distribute them as
separate works. But when you distribute the same sections as
part of a whole which is a work based on the Library, the
distribution of the whole must be on the terms of this License,
144
Installation
whose permissions for other licensees extend to the entire

whole, and thus to each and every part regardless of who
wrote it.
Thus, it is not the intent of this section to claim rights or
contest your rights to work written entirely by you; rather, the
intent is to exercise the right to control the distribution of
derivative or collective works based on the Library. In
addition, mere aggregation of another work not based on the
Library with the Library (or with a work based on the Library)
on a volume of a storage or distribution medium does not
bring the other work under the scope of this License.
3. You may opt to apply the terms of the ordinary GNU General
Public License instead of this License to a given copy of the
Library. To do this, you must alter all the notices that refer to this
License, so that they refer to the ordinary GNU General Public
License, version 2, instead of to this License. (If a newer version
than version 2 of the ordinary GNU General Public License has
appeared, then you can specify that version instead if you wish.)
Do not make any other change in these notices.
Once this change is made in a given copy, it is irreversible for that
copy, so the ordinary GNU General Public License applies to all
subsequent copies and derivative works made from that copy.
This option is useful when you wish to copy part of the code of
the Library into a program that is not a library.
4. You may copy and distribute the Library (or a portion or
derivative of it, under Section 2) in object code or executable form
under the terms of Sections 1 and 2 above provided that you
accompany it with the complete corresponding machine-readable
source code, which must be distributed under the terms of
Sections 1 and 2 above on a medium customarily used for
software interchange.
If distribution of object code is made by offering access to copy
from a designated place, then offering equivalent access to copy
the source code from the same place satisfies the requirement to
distribute the source code, even though third parties are not
compelled to copy the source along with the object code.
145
Installation
5. A program that contains no derivative of any portion of the

Library, but is designed to work with the Library by being
compiled or linked with it, is called a "work that uses the
Library". Such a work, in isolation, is not a derivative work of the
Library, and therefore falls outside the scope of this License.
However, linking a "work that uses the Library" with the Library
creates an executable that is a derivative of the Library (because it
contains portions of the Library), rather than a "work that uses the
library". The executable is therefore covered by this License.
Section 6 states terms for distribution of such executables.
When a "work that uses the Library" uses material from a header
file that is part of the Library, the object code for the work may be
a derivative work of the Library even though the source code is
not. Whether this is true is especially significant if the work can
be linked without the Library, or if the work is itself a library. The
threshold for this to be true is not precisely defined by law.
If such an object file uses only numerical parameters, data
structure layouts and accessors, and small macros and small
inline functions (ten lines or less in length), then the use of the
object file is unrestricted, regardless of whether it is legally a
derivative work. (Executables containing this object code plus
portions of the Library will still fall under Section 6.)
Otherwise, if the work is a derivative of the Library, you may
distribute the object code for the work under the terms of Section
6. Any executables containing that work also fall under Section 6,
whether or not they are linked directly with the Library itself.
6. As an exception to the Sections above, you may also combine or
link a "work that uses the Library" with the Library to produce a
work containing portions of the Library, and distribute that work
under terms of your choice, provided that the terms permit
modification of the work for the customer's own use and reverse
engineering for debugging such modifications.
You must give prominent notice with each copy of the work that
the Library is used in it and that the Library and its use are
covered by this License. You must supply a copy of this License. If
the work during execution displays copyright notices, you must
include the copyright notice for the Library among them, as well
as a reference directing the user to the copy of this License. Also,
you must do one of these things:
146
Installation
a) Accompany the work with the complete corresponding

machine-readable source code for the Library including
whatever changes were used in the work (which must be
distributed under Sections 1 and 2 above); and, if the work is
an executable linked with the Library, with the complete
machine-readable "work that uses the Library", as object
code and/or source code, so that the user can modify the
Library and then relink to produce a modified executable
containing the modified Library. (It is understood that the
user who changes the contents of definitions files in the
Library will not necessarily be able to recompile the
application to use the modified definitions.)
b) Use a suitable shared library mechanism for linking with
the Library. A suitable mechanism is one that (1) uses at run
time a copy of the library already present on the user's
computer system, rather than copying library functions into
the executable, and (2) will operate properly with a modified
version of the library, if the user installs one, as long as the
modified version is interface-compatible with the version
that the work was made with.
"c) Accompany the work with a written offer, valid for at least
three years, to give the same user the materials specified in
Subsection 6a, above, for a charge no more than the cost of
performing this distribution.
"d) If distribution of the work is made by offering access to
copy from a designated place, offer equivalent access to
copy the above specified materials from the same place.
"e) Verify that the user has already received a copy of these
materials or that you have already sent this user a copy.
For an executable, the required form of the "work that uses
the Library" must include any data and utility programs
needed for reproducing the executable from it. However, as a
special exception, the materials to be distributed need not
include anything that is normally distributed (in either
source or binary form) with the major components (compiler,
kernel, and so on) of the operating system on which the
executable runs, unless that component itself accompanies
the executable.
It may happen that this requirement contradicts the license
restrictions of other proprietary libraries that do not
normally accompany the operating system.
147
Installation
Such a contradiction means you cannot use both them and

the Library together in an executable that you distribute.
7. You may place library facilities that are a work based on the
Library side-by-side in a single library together with other library
facilities not covered by this License, and distribute such a
combined library, provided that the separate distribution of the
work based on the Library and of the other library facilities is
otherwise permitted, and provided that you do these two things:
"a) Accompany the combined library with a copy of the same
work based on the Library, uncombined with any other library
facilities. This must be distributed under the terms of the
Sections above.
"b) Give prominent notice with the combined library of the fact
that part of it is a work based on the Library, and explaining
where to find the accompanying uncombined form of the same
work.
8. You may not copy, modify, sublicense, link with, or distribute the
Library except as expressly provided under this License. Any
attempt otherwise to copy, modify, sublicense, link with, or
distribute the Library is void, and will automatically terminate
your rights under this License. However, parties who have
received copies, or rights, from you under this License will not
have their licenses terminated so long as such parties remain in
full compliance.
9. You are not required to accept this License, since you have not
signed it. However, nothing else grants you permission to modify
or distribute the Library or its derivative works. These actions are
prohibited by law if you do not accept this License. Therefore, by
modifying or distributing the Library (or any work based on the
Library), you indicate your acceptance of this License to do so,
and all its terms and conditions for copying, distributing or
modifying the Library or works based on it.
10. Each time you redistribute the Library (or any work based on the
Library), the recipient automatically receives a license from the
original licensor to copy, distribute, link with or modify the
Library subject to these terms and conditions.
You may not impose any further restrictions on the recipients'
exercise of the rights granted herein. You are not responsible for
enforcing compliance by third parties with this License.
148
Installation
11. If, as a consequence of a court judgment or allegation of patent

infringement or for any other reason (not limited to patent
issues), conditions are imposed on you (whether by court order,
agreement or otherwise) that contradict the conditions of this
License, they do not excuse you from the conditions of this
License. If you cannot distribute so as to satisfy simultaneously
your obligations under this License and any other pertinent
obligations, then as a consequence you may not distribute the
Library at all. For example, if a patent license would not permit
royalty-free redistribution of the Library by all those who receive
copies directly or indirectly through you, then the only way you
could satisfy both it and this License would be to refrain entirely
from distribution of the Library.
If any portion of this section is held invalid or unenforceable
under any particular circumstance, the balance of the section is
intended to apply, and the section as a whole is intended to apply
in other circumstances. It is not the purpose of this section to
induce you to infringe any patents or other property right claims
or to contest validity of any such claims; this section has the sole
purpose of protecting the integrity of the free software
distribution system which is implemented by public license
practices. Many people have made generous contributions to the
wide range of software distributed through that system in
reliance on consistent application of that system; it is up to the
author/donor to decide if he or she is willing to distribute
software through any other system and a licensee cannot impose
that choice. This section is intended to make thoroughly clear
what is believed to be a consequence of the rest of this License.
12. If the distribution and/or use of the Library is restricted in certain
countries either by patents or by copyrighted interfaces, the
original copyright holder who places the Library under this
License may add an explicit geographical distribution limitation
excluding those countries, so that distribution is permitted only
in or among countries not thus excluded. In such case, this
License incorporates the limitation as if written in the body of
this License.
13. The Free Software Foundation may publish revised and/or new
versions of the Lesser General Public License from time to time.
Such new versions will be similar in spirit to the present version,
but may differ in detail to address new problems or concerns.
Each version is given a distinguishing version number.
149
Installation
If the Library specifies a version number of this License which

applies to it and "any later version", you have the option of
following the terms and conditions either of that version or of any
later version published by the Free Software Foundation. If the
Library does not specify a license version number, you may
choose any version ever published by the Free Software
Foundation.
14. If you wish to incorporate parts of the Library into other free
programs whose distribution conditions are incompatible with
these, write to the author to ask for permission. For software
which is copyrighted by the Free Software Foundation, write to
the Free Software Foundation; we sometimes make exceptions
for this. Our decision will be guided by the two goals of
preserving the free status of all derivatives of our free software
and of promoting the sharing and reuse of software generally.
NO WARRANTY
15. BECAUSE THE LIBRARY IS LICENSED FREE OF CHARGE,
THERE IS NO WARRANTY FOR THE LIBRARY, TO THE
EXTENT PERMITTED BY APPLICABLE LAW. EXCEPT WHEN
OTHERWISE STATED IN WRITING THE COPYRIGHT
HOLDERS AND/OR OTHER PARTIES PROVIDE THE
LIBRARY "AS IS" WITHOUT WARRANTY OF ANY KIND,
EITHER EXPRESSED OR IMPLIED, INCLUDING, BUT NOT
LIMITED TO, THE IMPLIED WARRANTIES OF
MERCHANTABILITY AND FITNESS FOR A PARTICULAR
PURPOSE. THE ENTIRE RISK AS TO THE QUALITY AND
PERFORMANCE OF THE LIBRARY IS WITH YOU. SHOULD
THE LIBRARY PROVE DEFECTIVE, YOU ASSUME THE COST
OF ALL NECESSARY SERVICING, REPAIR OR CORRECTION.
16. IN NO EVENT UNLESS REQUIRED BY APPLICABLE LAW OR
AGREED TO IN WRITING WILL ANY COPYRIGHT HOLDER,
OR ANY OTHER PARTY WHO MAY MODIFY AND/OR
REDISTRIBUTE THE LIBRARY AS PERMITTED ABOVE, BE
LIABLE TO YOU FOR DAMAGES, INCLUDING ANY
GENERAL, SPECIAL, INCIDENTAL OR CONSEQUENTIAL
DAMAGES ARISING OUT OF THE USE OR INABILITY TO USE
THE LIBRARY (INCLUDING BUT NOT LIMITED TO LOSS OF
DATA OR DATA BEING RENDERED INACCURATE OR
LOSSES SUSTAINED BY YOU OR THIRD PARTIES OR A
FAILURE OF THE LIBRARY TO OPERATE WITH ANY OTHER
150
Installation
SOFTWARE), EVEN IF SUCH HOLDER OR OTHER PARTY

HAS BEEN ADVISED OF THE POSSIBILITY OF SUCH
DAMAGES.
END OF TERMS AND CONDITIONS
How to Apply These Terms to Your New Libraries
If you develop a new library, and you want it to be of the greatest
possible use to the public, we recommend making it free software
that everyone can redistribute and change. You can do so by
permitting redistribution under these terms (or, alternatively, under
the terms of the ordinary General Public License).
To apply these terms, attach the following notices to the library. It is
safest to attach them to the start of each source file to most effectively
convey the exclusion of warranty; and each file should have at least
the "copyright" line and a pointer to where the full notice is found.
one line to give the library's name and an idea of what it does.
Copyright (C) year name of author
This library is free software; you can redistribute it and/or modify it
under the terms of the GNU Lesser General Public License as
published by the Free Software Foundation; either version 2.1 of the
License, or (at your option) any later version.
This library is distributed in the hope that it will be useful, but
WITHOUT ANY WARRANTY; without even the implied warranty of
MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.
See the GNU Lesser General Public License for more details.
You should have received a copy of the GNU Lesser General Public
License along with this library; if not, write to the Free Software
Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA
02110-1301 USA
Also add information on how to contact you by electronic and paper
mail.
You should also get your employer (if you work as a programmer) or
your school, if any, to sign a "copyright disclaimer" for the library, if
necessary. Here is a sample; alter the names:
Yoyodyne, Inc., hereby disclaims all copyright interest in the library
`Frob' (a library for tweaking knobs) written by James Random
Hacker.
signature of Ty Coon, 1 April 1990
Ty Coon, President of Vice
151
Installation
Microsoft
In addition to the license terms for the software (as defined below) in
your agreement with EMC or its authorized reseller, the following
additional terms and conditions shall be applicable to any Microsoft
Corporation software or intellectual property that may be included
with the software. You hereby agree to these following terms and
conditions by installing and/or using the software.
You may install and use this EMC software (hereinafter the
"software") only as a file server as specified in the documentation
accompanying this software, and only to interoperate or
communicate with native Microsoft Windows 2000 Professional or
Windows XP Home or Professional client operating systems (or any
successors thereto). You may not install or use this software on a
personal computer under any circumstances. For purposes of the
foregoing, a personal computer means any computer configured
so that its primary purpose is for use by one person at a time and that
uses a video display and keyboard.
You hereby agree that Microsoft Corporation, as a licensor of
intellectual property that may be included in the software, is
intended to be a third party beneficiary (only to the extent permitted
by applicable law) of this agreement with rights to enforce any terms
herein that affect any included Microsoft intellectual property that
may be included in the software or other Microsoft interest related to
this agreement.
You may install the software on a server that (a) contains not more
than four (4) processors, which processor has a maximum of
thirty-two (32) bits and less than or equal to four (4) gigabytes of
RAM, and (b) is not capable of having memory added, changed or
removed without the requirement that the server on which it is
running be rebooted (hot swapping capabilities).
You may not use the software in conjunction with any software that
supports hot swapping capabilities or clustering capabilities, where
clustering capabilities means the ability to allow a group of
servers (e.g. host servers) to function as a single high-availability
platform for running applications using application failover between
server nodes in the group. You may not use the software in
conjunction with any software that supports hot swapping
capabilities, where hot swapping capabilities means the ability to
have memory added, changed or removed without the requirement
that the server on which it is running be rebooted.
152
Installation
Hardware or software that reduces the number of electronic devices

directly accessing or using services provided by the software does not
reduce the number of devices deemed to be concurrently accessing or
using services provided by the software. The number of electronic
devices accessing or using the software is equal to the number of
devices that access or use, either directly or through a multiplexing
device, services provided by the software.
As used here, a multiplexing device means any hardware or
software that provides or obtains access, directly or indirectly, to
services provided by the software to or on behalf of multiple other
devices concurrently through a reduced number of connections to
such software.
You must acquire and dedicate a separate Windows CAL for each
electronic device that is accessing or using, either directly or through
or from a multiplexing device, services provided by the software. A
Windows CAL means a Windows device Client Access License
(CAL), or a Windows user CAL, in either case for a Microsoft
Windows Server 2003 (Standard Edition, Enterprise Edition, or Datacenter Edition) server operating system product (Windows Server
2003); or (b) a Microsoft core CAL that provides an individual
person or electronic device with rights to access and use Windows
Server 2003, in either of (a) or (b) above that you have acquired for
use with one or more such Microsoft Windows Server 2003 operating
system products or electronic device and that is used on a per user or
per device basis.
In the event the software is being used as part of an evaluation, you
may install and use the software solely for the purposes of testing
and evaluation for time period specified below. Your right to use the
software shall be effective from the date you first install any portion
of the software on any server for a period of one hundred twenty
(120) days.
You may install the Software on a server that (a) contains only one (1)
processor, which processor has a maximum of thirty-two (32) bits and
five hundred twelve (512) megabytes of RAM, (b) contains a network
interface card with a maximum throughput of one hundred (100)
megabits per second, and (c) is not capable of having memory added,
changed or removed without the requirement that the server on
which it is running be rebooted (hot swapping capabilities). You may
not use the software in conjunction with any software that supports
hot swapping capabilities or clustering capabilities, where clustering capabilities means the ability to allow a group of servers to
function as a single high-availability platform for running
applications using application failover between Server nodes in the
group.
153
Installation
Open Source Initiative

OSI - Common
Development and
Distribution License
(CDDL)
Copyright 1994-2007 by Sun Microsystems, Inc., 901 San Antonio

Road, Palo Alto, California, 94303, U.S.A. All Rights Reserved.
1. Definitions
1.1. Contributor means each individual or entity that creates or
contributes to the creation of Modifications.
1.2. Contributor Version means the combination of the Original
Software, prior Modifications used by a Contributor (if any), and
the Modifications made by that particular Contributor.
1.3. Covered Software means
(a) the Original Software, or
(b) Modifications, or
(c) the combination of files containing Original Software with files
containing Modifications, in each case including portions
thereof.
1.4. Executable means the Covered Software in any form other than
Source Code.
1.5. Initial Developer means the individual or entity that first makes
Original Software available under this License.
1.6. Larger Work means a work which combines Covered Software or
portions thereof with code not governed by the terms of this
License.
1.7. License means this document.
1.8. Licensable means having the right to grant, to the maximum
extent possible, whether at the time of the initial grant or
subsequently acquired, any and all of the rights conveyed herein.
1.9. Modifications means the Source Code and Executable form of
any of the following:
a. Any file that results from an addition to, deletion from or
modification of the contents of a file containing Original
Software or previous Modifications;
b. Any new file that contains any part of the Original Software or
previous Modification; or
c. Any new file that is contributed or otherwise made available
under the terms of this License.
1.10. Original Software means the Source Code and Executable form
of computer software code that is originally released under this
License.
154
Installation
1.11. Patent Claims means any patent claim(s), now owned or

hereafter acquired, including without limitation, method,
process, and apparatus claims, in any patent Licensable by
grantor.
1.12. Source Code means
(a) the common form of computer software code in which
modifications are made and
(b) associated documentation included in or with such code.
1.13. You (or Your) means an individual or a legal entity exercising
rights under, and complying with all of the terms of, this
License. For legal entities, You includes any entity which
controls, is controlled by, or is under common control with You.
For purposes of this definition, control means
(a) the power, direct or indirect, to cause the direction or
management of such entity, whether by contract or otherwise,
or
(b) ownership of more than fifty percent (50%) of the outstanding
shares or beneficial ownership of such entity.
2. License Grants
2.1. The Initial Developer Grant.

Conditioned upon Your compliance with Section 3.1 below and
subject to third party intellectual property claims, the Initial
Developer hereby grants You a world-wide, royalty-free,
non-exclusive license:
(a) under intellectual property rights (other than patent or
trademark) Licensable by Initial Developer, to use, reproduce,
modify, display, perform, sublicense and distribute the
Original Software (or portions thereof), with or without
Modifications, and/or as part of a Larger Work; and
(b) under Patent Claims infringed by the making, using or selling
of Original Software, to make, have made, use, practice, sell,
and offer for sale, and/or otherwise dispose of the Original
Software (or portions thereof).
(c) The licenses granted in Sections 2.1(a) and (b) are effective on
the date Initial Developer first distributes or otherwise makes
the Original Software available to a third party under the terms
of this License.
(d) Notwithstanding Section 2.1(b) above, no patent license is
granted: (1) for code that You delete from the Original
Software, or (2) for infringements caused by: (i) the
modification of the Original Software, or (ii) the combination
of the Original Software with other software or devices.
155
Installation
2.2. Contributor Grant

Conditioned upon Your compliance with Section 3.1 below and
subject to third party intellectual property claims, each
Contributor hereby grants You a world-wide, royalty-free,
non-exclusive license:
(a) under intellectual property rights (other than patent or
trademark) Licensable by Contributor to use, reproduce,
modify, display, perform, sublicense and distribute the
Modifications created by such Contributor (or portions
thereof), either on an unmodified basis, with other
Modifications, as Covered Software and/or as part of a Larger
Work; and
(b) under Patent Claims infringed by the making, using, or selling
of Modifications made by that Contributor either alone and/or
in combination with its Contributor Version (or portions of
such combination), to make, use, sell, offer for sale, have made,
and/or otherwise dispose of: (1) Modifications made by that
Contributor (or portions thereof); and (2) the combination of
Modifications made by that Contributor with its Contributor
Version (or portions of such combination).
(c) The licenses granted in Sections 2.2(a) and 2.2(b) are effective
on the date Contributor first distributes or otherwise makes
the Modifications available to a third party.
(d) Notwithstanding Section 2.2(b) above, no patent license is
granted: (1) for any code that Contributor has deleted from the
Contributor Version; (2) for infringements caused by: (i) third
party modifications of Contributor Version, or (ii) the
combination of Modifications made by that Contributor with
other software (except as part of the Contributor Version) or
other devices; or (3) under Patent Claims infringed by Covered
Software in the absence of Modifications made by that
Contributor.
3. Distribution Obligations
3.1. Availability of Source Code.
Any Covered Software that You distribute or otherwise make
available in Executable form must also be made available in
Source Code form and that Source Code form must be distributed
only under the terms of this License.
You must include a copy of this License with every copy of the
Source Code form of the Covered Software You distribute or
otherwise make available.
156
Installation
You must inform recipients of any such Covered Software in

Executable form as to how they can obtain such Covered
Software in Source Code form in a reasonable manner on or
through a medium customarily used for software exchange.
3.2. Modifications
The Modifications that You create or to which You contribute are
governed by the terms of this License. You represent that You
believe Your Modifications are Your original creation(s) and/or
You have sufficient rights to grant the rights conveyed by this
License.
3.3. Required Notices.
You must include a notice in each of Your Modifications that
identifies You as the Contributor of the Modification. You may
not remove or alter any copyright, patent or trademark notices
contained within the Covered Software, or any notices of
licensing or any descriptive text giving attribution to any
Contributor or the Initial Developer.
3.4. Application of Additional Terms.
You may not offer or impose any terms on any Covered Software
in Source Code form that alters or restricts the applicable version
of this License or the recipients rights hereunder. You may choose
to offer, and to charge a fee for, warranty, support, indemnity or
liability obligations to one or more recipients of Covered
Software. However, you may do so only on Your own behalf, and
not on behalf of the Initial Developer or any Contributor. You
must make it absolutely clear that any such warranty, support,
indemnity or liability obligation is offered by You alone, and You
hereby agree to indemnify the Initial Developer an every
Contributor for any liability incurred by the Initial Developer or
such Contributor as a result of warranty, support, indemnity or
liability terms You offer.
3.5. Distribution of Executable Versions.
You may distribute the Executable form of the Covered Software
under the terms of this License or under the terms of a license of
Your choice, which may contain terms different from this License,
provided that You are in compliance with the terms of this
License and that the license for the Executable form does not
attempt to limit or alter the recipients rights in the Source Code
form from the rights set forth in this License.
157
Installation
If You distribute the Covered Software in Executable form under

a different license, You must make it absolutely clear that any
terms which differ from this License are offered by You alone, not
by the Initial Developer or Contributor.
You hereby agree to indemnify the Initial Developer and every
Contributor for any liability incurred by the Initial Developer or
such Contributor as a result of any such terms You offer.
3.6. Larger Works
You may create a Larger Work by combining Covered Software
with other code not governed by the terms of this License and
distribute the Larger Work as a single product. In such a case, You
must make sure the requirements of this License are fulfilled for
the Covered Software.
4. Versions of the License
4.1. New Versions.
Sun Microsystems, Inc. is the initial license steward and may
publish revised and/or new versions of this License from time to
time. Each version will be given a distinguishing version number.
Except as provided in Section 4.3, no one other than the license
steward has the right to modify this License.
4.2. Effect of New Versions.
You may always continue to use, distribute or otherwise make the
Covered Software available under the terms of the version of the
License under which You originally received the Covered
Software. If the Initial Developer includes a notice in the Original
Software prohibiting it from being distributed or otherwise made
available under any subsequent version of the License, You must
distribute and make the Covered Software available under the
terms of the version of the License under which You originally
received the Covered Software. Otherwise, You may also choose
to use, distribute or otherwise make the Covered Software
available under the terms of any subsequent version of the
License published by the license steward.
4.3. Modified Versions.
When You are an Initial Developer and You want to create a new
license for Your Original Software, You may create and use a
modified version of this License if You:
(a) rename the license and remove any references to the name of
the license steward (except to note that the license differs from
this License); and
(b) otherwise make it clear that the license contains terms which
differ from this License.
158
Installation
5. DISCLAIMER OF WARRANTY
COVERED SOFTWARE IS PROVIDED UNDER THIS LICENSE
ON AN AS IS BASIS, WITHOUT WARRANTY OF ANY KIND,
EITHER EXPRESSED OR IMPLIED, INCLUDING, WITHOUT
LIMITATION, ARRANTIES THAT THE COVERED SOFTWARE
IS FREE OF DEFECTS, MERCHANTABLE, FIT FOR A
PARTICULAR PURPOSE OR NON-INFRINGING.
THE ENTIRE RISK AS TO THE QUALITY AND
PERFORMANCE OF THE COVERED SOFTWARE IS WITH YOU.
SHOULD ANY COVERED SOFTWARE PROVE DEFECTIVE IN
ANY RESPECT, YOU (NOT THE INITIAL DEVELOPER OR ANY
OTHER CONTRIBUTOR) ASSUME THE COST OF ANY
NECESSARY SERVICING, REPAIR OR CORRECTION. THIS
DISCLAIMER OF WARRANTY CONSTITUTES AN ESSENTIAL
PART OF THIS LICENSE. NO USE OF ANY COVERED
SOFTWARE IS AUTHORIZED HEREUNDER EXCEPT UNDER
THIS DISCLAIMER.
6. TERMINATION
6.1. This License and the rights granted hereunder will terminate
automatically if You fail to comply with terms herein and fail to
cure such breach within 30 days of becoming aware of the
breach. Provisions which, by their nature, must remain in effect
beyond the termination of this License shall survive.
6.2. If You assert a patent infringement claim (excluding declaratory
judgment actions) against Initial Developer or a Contributor (the
Initial Developer or Contributor against whom You assert such
claim is referred to as Participant) alleging that the Participant
Software (meaning the Contributor Version where the
Participant is a Contributor or the Original Software where the
Participant is the Initial Developer) directly or indirectly infringes any patent, then any and all rights granted directly or
indirectly to You by such Participant, the Initial Developer (if the
Initial Developer is not the Participant) and all Contributors
under Sections 2.1 and/or 2.2 of this License shall, upon 60 days
notice from Participant terminate prospectively and
automatically at the expiration of such 60 day notice period,
unless if within such 60 day period You withdraw Your claim
with respect to the Participant Software against such Participant
either unilaterally or pursuant to a written agreement with
Participant.
159
Installation
6.3. In the event of termination under Sections 6.1 or 6.2 above, all
end user licenses that have been validly granted by You or any
distributor hereunder prior to termination (excluding licenses
granted to You by any distributor) shall survive termination.
7. LIMITATION OF LIABILITY
UNDER NO CIRCUMSTANCES AND UNDER NO LEGAL
THEORY, WHETHER TORT (INCLUDING NEGLIGENCE),
CONTRACT, OR OTHERWISE, SHALL YOU, THE INITIAL
DEVELOPER, ANY OTHER CONTRIBUTOR, OR ANY
DISTRIBUTOR OF COVERED SOFTWARE, OR ANY SUPPLIER
OF ANY OF SUCH PARTIES, BE LIABLE TO ANY PERSON FOR
ANY INDIRECT, SPECIAL, INCIDENTAL, OR
CONSEQUENTIAL DAMAGES OF ANY CHARACTER
INCLUDING, WITHOUT LIMITATION, DAMAGES FOR LOST
PROFITS, LOSS OF GOODWILL, WORK STOPPAGE,
COMPUTER FAILURE OR MALFUNCTION, OR ANY AND
ALL OTHER COMMERCIAL DAMAGES OR LOSSES, EVEN IF
SUCH PARTY SHALL HAVE BEEN INFORMED OF THE
POSSIBILITY OF SUCH DAMAGES. THIS LIMITATION OF
LIABILITY SHALL NOT APPLY TO LIABILITY FOR DEATH OR
PERSONAL INJURY RESULTING FROM SUCH PARTYS
NEGLIGENCE TO THE EXTENT APPLICABLE LAW
PROHIBITS SUCH LIMITATION. SOME JURISDICTIONS DO
NOT ALLOW THE EXCLUSION OR LIMITATION OF
INCIDENTAL OR CONSEQUENTIAL DAMAGES, SO THIS
EXCLUSION AND LIMITATION MAY NOT APPLY TO YOU.
8. U.S. GOVERNMENT END USERS
The Covered Software is a commercial item, as that term is
defined in 48 C.F.R. 2.101 (Oct. 1995), consisting of commercial
computer software (as that term is defined at 48 C.F.R.
252.227-7014(a)(1)) and commercial computer software
documentation as such terms are used in 48 C.F.R. 12.212 (Sept.
1995). Consistent with 48 C.F.R. 12.212 and 48 C.F.R. 227.7202-1
through 227.7202-4 (June 1995), all U.S. Government End Users
acquire Covered Software with only those rights set forth herein.
This U.S. Government Rights clause is in lieu of, and supersedes,
any other FAR, DFAR, or other clause or provision that addresses
Government rights in computer software under this License.
160
Installation
9. MISCELLANEOUS
This License represents the complete agreement concerning
subject matter hereof. If any provision of this License is held to be
unenforceable, such provision shall be reformed only to the
extent necessary to make it enforceable. This License shall be
governed by the law of the jurisdiction specified in a notice
contained within the Original Software (except to the extent
applicable law, if any, provides otherwise), excluding such
jurisdictions conflict-of-law provisions. Any litigation relating to
this License shall be subject to the jurisdiction of the courts
located in the jurisdiction and venue specified in a notice
contained within the Original Software, with the losing party
responsible for costs, including, without limitation, court costs
and reasonable attorneys fees and expenses. The application of
the United Nations Convention on Contracts for the International
Sale of Goods is expressly excluded. Any law or regulation which
provides that the language of a contract shall be construed
against the drafter shall not apply to this License. You agree that
You alone are responsible for compliance with the United States
export administration regulations (and the export control laws
and regulation of any other countries) when You use, distribute or
otherwise make available any Covered Software.
10. RESPONSIBILITY FOR CLAIMS
As between Initial Developer and the Contributors, each party is
responsible for claims and damages arising, directly or indirectly,
out of its utilization of rights under this License and You agree to
work with Initial Developer and Contributors to distribute such
responsibility on an equitable basis. Nothing herein is intended or
shall be deemed to constitute any admission of liability.
Sun Microsystems
Copyright (c) 2005 Sun Microsystems, Inc., 4150 Network Circle,

Santa Clara, California 95054, U.S.A.
Sun Microsystems, Inc. Binary Code License Agreement for the JAVA
2 PLATFORM STANDARD EDITION RUNTIME ENVIRONMENT
5.0
SUN MICROSYSTEMS, INC. ("SUN") IS WILLING TO LICENSE
THE SOFTWARE IDENTIFIED BELOW TO YOU ONLY UPON THE
CONDITION THAT YOU ACCEPT ALL OF THE TERMS
CONTAINED IN THIS BINARY CODE LICENSE AGREEMENT
AND SUPPLEMENTAL LICENSE TERMS (COLLECTIVELY
"AGREEMENT"). PLEASE READ THE AGREEMENT CAREFULLY.
BY DOWNLOADING OR INSTALLING THIS SOFTWARE, YOU
161
Installation
ACCEPT THE TERMS OF THE AGREEMENT. INDICATE

ACCEPTANCE BY SELECTING THE "ACCEPT" BUTTON AT THE
BOTTOM OF THE AGREEMENT. IF YOU ARE NOT WILLING TO
BE BOUND BY ALL THE TERMS, SELECT THE "DECLINE"
BUTTON AT THE BOTTOM OF THE AGREEMENT AND THE
DOWNLOAD OR INSTALL PROCESS WILL NOT CONTINUE.
1. DEFINITIONS. "Software" means the identified above in binary
form, any other machine readable materials (including, but not
limited to, libraries, source files, header files, and data files), any
updates or error corrections provided by Sun, and any user
manuals, programming guides and other documentation
provided to you by Sun under this Agreement. "Programs" mean
Java applets and applications intended to run on the Java 2
Platform Standard Edition (J2SE platform) platform on
Java-enabled general purpose desktop computers and servers.
2. LICENSE TO USE. Subject to the terms and conditions of this
Agreement, including, but not limited to the Java Technology
Restrictions of the Supplemental License Terms, Sun grants you a
non-exclusive, non-transferable, limited license without license
fees to reproduce and use internally Software complete and
unmodified for the sole purpose of running Programs.
Additional licenses for developers and/or publishers are granted
in the Supplemental License Terms.
3. RESTRICTIONS. Software is confidential and copyrighted. Title
to Software and all associated intellectual property rights is
retained by Sun and/or its licensors. Unless enforcement is
prohibited by applicable law, you may not modify, decompile, or
reverse engineer Software.
You acknowledge that Licensed Software is not designed or
intended for use in the design, construction, operation or
maintenance of any nuclear facility. Sun Microsystems, Inc.
disclaims any express or implied warranty of fitness for such
uses. No right, title or interest in or to any trademark, service
mark, logo or trade name of Sun or its licensors is granted under
this Agreement. Additional restrictions for developers and/or
publishers licenses are set forth in the Supplemental License
Terms.
162
Installation
4. LIMITED WARRANTY. Sun warrants to you that for a period of

ninety (90) days from the date of purchase, as evidenced by a
copy of the receipt, the media on which Software is furnished (if
any) will be free of defects in materials and workmanship under
normal use. Except for the foregoing, Software is provided "AS
IS". Your exclusive remedy and Sun's entire liability under this
limited warranty will be at Sun's option to replace Software
media or refund the fee paid for Software. Any implied
warranties on the Software are limited to 90 days. Some states do
not allow limitations on duration of an implied warranty, so the
above may not apply to you. This limited warranty gives you
specific legal rights. You may have others, which vary from state
to state.
5. DISCLAIMER OF WARRANTY. UNLESS SPECIFIED IN THIS
AGREEMENT, ALL EXPRESS OR IMPLIED CONDITIONS,
REPRESENTATIONS AND WARRANTIES, INCLUDING ANY
IMPLIED WARRANTY OF MERCHANTABILITY, FITNESS FOR
A PARTICULAR PURPOSE OR NON-INFRINGEMENT ARE
DISCLAIMED, EXCEPT TO THE EXTENT THAT THESE
DISCLAIMERS ARE HELD TO BE LEGALLY INVALID.
6. LIMITATION OF LIABILITY. TO THE EXTENT NOT
PROHIBITED BY LAW, IN NO EVENT WILL SUN OR ITS
LICENSORS BE LIABLE FOR ANY LOST REVENUE, PROFIT
OR DATA, OR FOR SPECIAL, INDIRECT, CONSEQUENTIAL,
INCIDENTAL OR PUNITIVE DAMAGES, HOWEVER CAUSED
REGARDLESS OF THE THEORY OF LIABILITY, ARISING OUT
OF OR RELATED TO THE USE OF OR INABILITY TO USE
SOFTWARE, EVEN IF SUN HAS BEEN ADVISED OF THE
POSSIBILITY OF SUCH DAMAGES. In no event will Sun's
liability to you, whether in contract, tort (including negligence),
or otherwise, exceed the amount paid by you for Software under
this Agreement. The foregoing limitations will apply even if the
above stated warranty fails of its essential purpose.
Some states do not allow the exclusion of incidental or
consequential damages, so some of the terms above may not be
applicable to you.
7. TERMINATION. This Agreement is effective until terminated.
You may terminate this Agreement at any time by destroying all
copies of Software. This Agreement will terminate immediately
without notice from Sun if you fail to comply with any provision
of this Agreement.
163
Installation
Either party may terminate this Agreement immediately should

any Software become, or in either party's opinion be likely to
become, the subject of a claim of infringement of any intellectual
property right. Upon Termination, you must destroy all copies of
Software.
8. EXPORT REGULATIONS. All Software and technical data
delivered under this Agreement are subject to US export control
laws and may be subject to export or import regulations in other
countries. You agree to comply strictly with all such laws and
regulations and acknowledge that you have the responsibility to
obtain such licenses to export, re-export, or import as may be
required after delivery to you.
9. TRADEMARKS AND LOGOS. You acknowledge and agree as
between you and Sun that Sun owns the SUN, SOLARIS, JAVA,
JINI, FORTE, and iPLANET trademarks and all SUN, SOLARIS,
JAVA, JINI, FORTE, and iPLANET-related trademarks, service
marks, logos and other brand designations ("Sun Marks"), and
you agree to comply with the Sun Trademark and Logo Usage
Requirements currently located at
http://www.sun.com/policies/trademarks. Any use you make
of the Sun Marks inures to Sun's benefit.
10. U.S. GOVERNMENT RESTRICTED RIGHTS. If Software is being
acquired by or on behalf of the U.S. Government or by a U.S.
Government prime contractor or subcontractor (at any tier), then
the Government's rights in Software and accompanying
documentation will be only as set forth in this Agreement; this is
in accordance with 48 CFR 227.7201 through 227.7202-4 (for
Department of Defense (DOD) acquisitions) and with 48 CFR
2.101 and 12.212 (for non-DOD acquisitions).
11. GOVERNING LAW. Any action related to this Agreement will be
governed by California law and controlling U.S. federal law. No
choice of law rules of any jurisdiction will apply.
12. SEVERABILITY. If any provision of this Agreement is held to be
unenforceable, this Agreement will remain in effect with the
provision omitted, unless omission would frustrate the intent of
the parties, in which case this Agreement will immediately
terminate.
164
Installation
13. INTEGRATION. This Agreement is the entire agreement

between you and Sun relating to its subject matter. It supersedes
all prior or contemporaneous oral or written communications,
proposals, representations and warranties and prevails over any
conflicting or additional terms of any quote, order,
acknowledgment, or other communication between the parties
relating to its subject matter during the term of this Agreement.
No modification of this Agreement will be binding, unless in
writing and signed by an authorized representative of each party.
SUPPLEMENTAL LICENSE TERMS
These Supplemental License Terms add to or modify the terms of the
Binary Code License Agreement. Capitalized terms not defined in
these Supplemental Terms shall have the same meanings ascribed to
them in the Binary Code License Agreement . These Supplemental
Terms shall supersede any inconsistent or conflicting terms in the
Binary Code License Agreement, or in any license contained within
the Software.
A. Software Internal Use and Development License Grant. Subject to
the terms and conditions of this Agreement and restrictions and
exceptions set forth in the Software "README" file incorporated
herein by reference, including, but not limited to the Java
Technology Restrictions of these Supplemental Terms, Sun grants
you a non-exclusive, non-transferable, limited license without fees
to reproduce internally and use internally the Software complete
and unmodified for the purpose of designing, developing, and
testing your Programs.
B. License to Distribute Software. Subject to the terms and conditions
of this Agreement and restrictions and exceptions set forth in the
Software README file, including, but not limited to the Java
Technology Restrictions of these Supplemental Terms, Sun grants
you a non-exclusive, non-transferable, limited license without fees
to reproduce and distribute the Software, provided that (i) you
distribute the Software complete and unmodified and only
bundled as part of, and for the sole purpose of running, your
Programs, (ii) the Programs add significant and primary
functionality to the Software, (iii) you do not distribute additional
software intended to replace any component(s) of the Software,
(iv) you do not remove or alter any proprietary legends or notices
contained in the Software, (v) you only distribute the Software
subject to a license agreement that protects Sun's interests
consistent with the terms contained in this Agreement, and (vi)
you agree to defend and indemnify Sun and its licensors from and
165
Installation
against any damages, costs, liabilities, settlement amounts and/or

expenses (including attorneys' fees) incurred in connection with
any claim, lawsuit or action by any third party that arises or
results from the use or distribution of any and all Programs
and/or Software.
C. Java Technology Restrictions. You may not create, modify, or
change the behavior of, or authorize your licensees to create,
modify, or change the behavior of, classes, interfaces, or sub
packages that are in any way identified as "java", "javax", "sun" or
similar convention as specified by Sun in any naming convention
designation.
D. Source Code. Software may contain source code that, unless
expressly licensed for other purposes, is provided solely for
reference purposes pursuant to the terms of this Agreement.
Source code may not be redistributed unless expressly provided
for in this Agreement.
E. Third Party Code. Additional copyright notices and license terms
applicable to portions of the Software are set forth in the
THIRDPARTYLICENSEREADME.txt file. In addition to any terms
and conditions of any third party opensource/freeware license
identified in the THIRDPARTYLICENSEREADME.txt file, the
disclaimer of warranty and limitation of liability provisions in
paragraphs 5 and 6 of the Binary Code License Agreement shall
apply to all Software in this distribution.
F. Termination for Infringement. Either party may terminate this
Agreement immediately should any Software become, or in either
party's opinion be likely to become, the subject of a claim of
infringement of any intellectual property right.
G. Installation and Auto-Update. The Software's installation and
auto-update processes transmit a limited amount of data to Sun
(or its service provider) about those specific processes to help Sun
understand and optimize them. Sun does not associate the data
with personally identifiable information.
You can find more information about the data Sun collects at
http://java.com/data/.
For inquiries please contact: Sun Microsystems, Inc., 4150 Network
Circle, Santa Clara, California 95054, U.S.A.
166
Bug fixes
Bug fixes
Issue Tracker
available on
Powerlink
Fixed bug categories
The Issue Tracker contains a list of bugs, similar to Release Notes, for
select EMC products. This tool allows you to search on keyword
fields to find particular bugs and identify the software release in
which the bugs were identified/fixed. This tool can be used to check
if a bug was fixed in a revision that was suggested to you. It can also
be used if you're upgrading and you want to know about the revision
you're going to. This is not, however, a problem or status tracking
system.
The following table explains the category of the bug fixes.
Category
Meaning
CAVA
Celerra AntiVirus Agent
CDMS
Celerra Data Migration Service
Celerra Management
Applications
Celerra Monitor, Celerra Manager, Celerra Startup

Assistant, MMC snap-ins, Web UI
CFS
Celerra File System
CIFS
Common Internet File System
Control Station
Celerra Control Station
CORE
Celerra Data Movers
CSA
Celerra Startup Assistant
FileMover
Celerra FileMover
Hardware
Hardware, Call Homes
HighRoad (MPFS)
HighRoad (Multi-Protocol File System)
Installation
Celerra installs
Internationalization
Internationalization (I18N)
NDMP/Backup
Network Data Management Protocol and/or Backup
Networking
Networking issues
NFS
Network File System
Replication
Celerra Replicator, SnapSure, Checkpoint
SRDF
Symmetrix Remote Data Facility
Tools
EMCOPY, EMCACL, LGDUP, SHAREDUP
UFS
UNIX File System
Usermapper
Usermapper
167
Version 5.5.32.4
Version 5.5.32.4
CDMS
The following entries describe fixed problems pertaining to Celerra

Data Migration Service:
Impact Level
Severity 1
Symptom
Description
During a CDMS migration, a Data Mover panicked with the following DART panic/fault message:
*** Divide exception. Virt ADDRESS: de77be ***. The stack trace showed:
0x170bb7c: 0x13d890 waitForReboot+0x90
0x170bb9c: 0x13da8d fault_dump+0x67
0x170bbbc: 0x137b3c handle_fatal_fault+0xa
0x170bc10: 0xde76f9 __udivdi3+0x29
0x260e39f0: 0x621ead _ZN9UFS_Vnode4bmapElRlPiS1_+0x233
0x260e3a50: 0x6255bd
xxx_ZN9UFS_Vnode5breadEjR11File_Statusi+0xc1
0x260e3aa0: 0x6256d7
xxx_ZN9UFS_Vnode9blockReadEjR11File_Status+0x4d
0x260e3ae0: 0xb98cdc
xxx_ZN10MGFS_Vnode9blockReadEjR11File_Status+0x48
Fix
A race condition caused a thread to get suspended in between the time it marked an operation as
complete and when it wrote the status. This resulted in the status overwriting another variable when
the thread resumed. Now the operation is not completed until the status is written.
Fixed in version
5.5.32.0
Celerra
Management
Applications
168
The following entries describe fixed problems relating to the Celerra

Management Applications, i.e., Celerra Monitor, Celerra Manager,
WebUI, or Celerra MMC Snap-ins:
Impact Level
Severity 2
Symptom
Description
The osmlog displayed intermittently in Celerra Monitor. Most often, it displayed the first time it was
attempted, but did not display in subsequent attempts.
Fix
There was a race condition where code was entering a wait state after having received notification that
the data had arrived, and waited indefinitely. Code now verifies that it has not received the response
before entering the wait state.
Service Request
19001252
Fixed in version
5.5.32.0
Version 5.5.32.4
Impact Level
Severity 2
Symptom
Description
Once the used space in tree is greater than 2 TB, the GUI shows "Hard storage limit
exceeded" even if there is space left. In this case, GUI shows "Hard storage limit
exceeded", but mouse over popup on "storage used" field shows "96% of hard limit".
Fix
Celerra Manager now returns the proper status value when 'used MB', 'soft limit' and 'hard limit' values
are greater than 2 TB.
Service Request
20153750
Fixed in version
5.5.32.1
Impact Level
Severity 1
Symptom
Description
XML API was programmed to generate tree quota cache everyday at 2:30 AM by sending a request to
APL and storing the response on disk so that it can service quota queries efficiently. When a quota
query was started on a system with more than 30,000 tree quotas and 512 MB of total memory, the
quota cache request consumed a large amount of memory due to the sheer number of attributes
requested for each of these tree quotas. After several hours (in this particular case, 8 hours) of running
this query, the Control Station ran out of memory and panicked.
Note: This issue only occurred on systems with 512 MB of memory and more than 30,000 tree quotas.
Systems with more than 512 MB of memory (e.g., NSX, NS20, NS40, NS80) were not affected. Other
scheduled processes such as nasdb backup were not affected as they run in parallel with quota query.
Fix
As of version 5.5.32.4, during new installs and upgrades, tree quota cache is disabled by default to
prevent this Control Station panic. Tree quota cache can be enabled via the xml.api.enable.quota.poll
parameter in the /nas/sys/xml_api.conf file. For XML API users, enabling this parameter will increase
quota query efficiency. However, this parameter should not be enabled on systems with 512 MB of
memory, and over 30,000 tree quotas. You can check your tree quota usage by selecting the Tree
Quota tab in Celerra Manager.
Service Request
19617602
Fixed in version
5.5.32.3
169
Version 5.5.32.4
Impact Level
Severity 2
Symptom
Description
New file systems were not appearing in the user's ECC console until jserver was restarted on the
Control Station.
Fix
Two counters in the application indication manager and in jserver were getting out of synch, resulting
in new file systems not be included in the response sent to ECC. This problem no longer occurs.
Service Request
18704664
Fixed in version
5.5.32.0
CFS
170
The following entries describe fixed issues pertaining to the Celerra

File System:
Impact Level
Severity 1
Symptom
Description
When a file system was created with a size that was not a multiple of 64MB and was later extended, if
the file system became nearly full, the Data Mover panicked with the following DART panic/fault
message: alloc failed: file system full. The stack trace showed:
0xb8981560: 0x13d85c waitForReboot+0x90
0xb8981580: 0x13da59 fault_dump+0x67
0xb89815a0: 0x13d959 PANIC+0x29
0xb89815b0: 0x618caa
xxx_ZN14UFS_FileSystem28markCorruptedFsForUnmount
0xb8981630: 0x624d01
xxx_ZN9UFS_Vnode5allocEP20UFS_UpdateDescriptorllj
0xb8981670: 0x6231d8
xxx_ZN9UFS_Vnode10ballocListEP20UFS_UpdateDescrip
0xb8981710: 0x626262
xxx_ZN9UFS_Vnode15blockChainWriteEiP10BlockChaini
0xb8981bd0: 0x625deb
xxx_ZN9UFS_Vnode15blockChainWriteEiP10BlockChain+
Fix
File system extension no longer causes writes against these file systems to trigger a Data Mover panic
when nearly full. When a file system is extended and the last cylinder group is partially allocated, code
now ensures that there will be no mismatch between the cylinder group cache and the ondisk
structures.
Fixed in version
5.5.32.0
Version 5.5.32.4
CIFS
The following entries describe fixed Common Internet File System

problems:
Impact Level
Severity 1
Symptom
Description
CIFS users were unable to rename files and directories unless they were given WRITE_DATA
permission on the containing folder. Only the APPEND_DATA right should be needed to allow
renames to occur.
Fix
WRITE_DATA permission is no longer required in the parent folder to allow objects to be renamed.
Service Request
17889920
Fixed in version
5.5.32.0
Impact Level
Severity 1
Symptom
Description
If GPO updates for a joined CIFS server failed (i.e., unable to connect to the domain controller to
retrieve the GPO settings), and a GPO update was in progress when the user tried to stop CIFS, the
CIFS stop waited for the GPO update to complete. However, if the CIFS stop had progressed to a point
where the GPO update could not complete, the CIFS stop wouldn't complete, and the only correction
was to reboot the Data Mover.
Fix
If the GPO update fails to connect to the domain controller, it now checks to see if CIFS stop is in
progress before proceeding. If CIFS stop is in progress, the GPO update is now aborted so that the
CIFS stop can complete.
Service Request
20176954
Fixed in version
5.5.32.0
Impact Level
Severity 1
Symptom
Description
CIFS clients were unable to rename named streams (alternate data streams) within a file. The Data
Mover reported that a rename had completed, though it was not performed. This caused unexpected
behaviors depending on the users application, such as thumbnail images intermittently not being
displayed.
Fix
The Data Mover was renaming streams from the original name back to the original name, causing the
rename operations to have no effect. Alternate Data Stream renames now complete successfully.
Service Request
20090030
Fixed in version
5.5.32.0
171
Version 5.5.32.4
172
Impact Level
Severity 2
Symptom
Description
Users lost access to a file system when the Windows security log was stored on a user file system that
was replicated or had checkpoints being created or refreshed. The Data Mover needed to be failed
over or rebooted to correct the issue
Fix
The security log was being written and the file system pause process (used with Replicator or
Snapsure) was temporarily blocking access to the file system. This deadlock condition has been
eliminated.
Service Request
19966230
Fixed in version
5.5.32.0
Impact Level
Severity 1
Symptom
Description
The Data Mover panicked when it failed to retrieve the first GPO update for a new compname on a
VDM which had not previously been updated. The panic only occurred on VDMs where another
compname had not previously successfully retrieved a GPO update. DART panic/fault message:
*** Page Fault Interrupt. Virt ADDRESS: 18699f Err code: 0
Target addr: ffd830cd **. The stack trace showed:
0x1706214: 0x13d810 waitForReboot+0x90
0x1706234: 0x13da0d fault_dump+0x67
0x1706254: 0x137abc handle_fatal_fault+0xa
0x17062d0: 0x184f20
xxx_ZN18Stdio_AbstractFile6vprintEPKcPPc+0x1a
0x24a6ddb0: 0x1bf596
xxx_ZN16logObject_Memory7logvmsgE11logFacility11l
0x24a6ddd0: 0x19369f _ZN13logObject_XML5replyEPKcz+0x63
0x24a6de00: 0xaba605
xxx_Z20get_gpo_server_entryP13logObject_XML+0x735
0x24a6de40: 0xabb267
xxx_Z14get_gpo_serverP13logObject_XMLRiPcRK6Strin
Fix
The fix will prevent the Data Mover panicking when it fails to retrieve the first GPO update for the first
compname on a VDM.
Service Request
19867642
Fixed in version
5.5.32.0
Version 5.5.32.4
Impact Level
Severity 2
Symptom
Description
User mounted a file system (default,native,rw) and exported as read-only. User was then able to delete
the files on the exported CIFS share from the CIFS clients. The only way to prevent the files/directories
from being deleted from the CIFS client was to mount the file system
uxfs,perm,rw,accesspolicy=SECURE.
Fix
Added a check to prevent an open for deletion when the share is in ReadOnly. The following message
will appear when attempting to delete a file: Error message Cannot delete
<filename>: Access Denied. The source file may be in use.
Service Request
18808826
Fixed in version
5.5.32.0
Impact Level
Severity 1
Symptom
Description
Memory leaked when using CIFS clients that didn't support Unicode. When memory was completely
exhausted, the Data Mover panicked with the following DART panic/fault message: >>PANIC in
file: ../addrspac.cxx at line: 275 : Out of memory. The stack trace
showed:
0xec9ddb10: 0x13d1db waitForReboot+0x8b
0xec9ddb30: 0x13d3e3 fault_dump+0x67
0xec9ddb50: 0x13d2e3 PANIC+0x2b 0xec9ddb60: 0x13918c
xxx_ZN19Memory_AddressSpace10allocPagesEi12Memory
0xec9ddb80: 0x16bae1
xxx_ZN20Buffer_Uncached_Pool10addBuffersEmim+0x97
0xec9ddbc0: 0x16b9fa
xxx_ZN20Buffer_Uncached_Pool14getUncachedBufE15St
0xec9ddbe0: 0x16a75c allocb+0x72
0xec9ddc10: 0xd73c3c
xxx_ZN14VirtualLU_Disk4readEP6VLUCCBmm+0x26c
Fix
This problem occurred because memory allocated for ASCII to Unicode conversion didnt get freed.
Code was modified to ensure memory allocated for ASCII to Unicode conversion is properly released.
Service Request
20118982
Fixed in version
5.5.32.0
173
Version 5.5.32.4
174
Impact Level
Severity 1
Symptom
Description
CIFS users lost access when there were many CIFS users active, virus checking was enabled, and
one or more of the virus checker servers was being used for non-virus check access to the Celerra.
Fix
Celerra now reserves CIFS threads to ensure it can respond to requests from the virus checker
servers even when there are many CIFS users active. Previously, the reserved threads could be used
for non antivirus CIFS activity from a CAVA server. Now these threads can only be used for virus
checking activities from a CAVA server and can not be tied up by general access from the CAVA
server.
Service Request
18808826
Fixed in version
5.5.32.0
Impact Level
Severity 1
Symptom
Description
When using the old API from Windows 9x clients, trying to get information (properties/attributes) about
the shares caused memory leak on Data Mover in the function that was getting the information. The
Data Mover panicked with the following DART panic/fault message: >>PANIC in file:
../addrspac.cxx at line: 275 :
Out of memory.
The stack trace showed: 0xc7f3f970: 0x13d600 waitForReboot+0x90
0xc7f3f990: 0x13d7fd fault_dump+0x67
0xc7f3f9b0: 0x13d6fd PANIC+0x29
0xc7f3f9c0: 0x1394d6
0xc7f3f9e0: 0x187eaa _Z11more_memoryiP9free_list+0x4e
0xc7f3fa00: 0x187f60 malloc+0x92
0xc7f3fa30: 0x1886a5
xxx_Znwm+0x15
0xc7f3fa40: 0x1886d2 _Znam+0xc
Fix
Removed unneeded call to the function that retrieved share information to prevent the memory leak.
Service Request
19749864
Fixed in version
5.5.32.0
Version 5.5.32.4
Impact Level
Severity 1
Symptom
Description
After rebooting a NIS server or restarting NIS services on it, Data Movers could no longer perform NIS
lookups intermittently, which caused user access issues. The server log repeated the errors:
NETLIB: 4: NIS Ready
NETLIB: 3: NIS ERROR: Can't bind to server which serves this
domain
Rebooting the Data Mover resolved this issue.
Fix
When first connecting to a NIS server, the Data Mover memorized the UDP port number the NIS server
was using. When the port changed after a NIS restart, the Data Mover learned the new port, but
continued to use the old port number for certain types of NIS calls. The Data Mover now updates its
NIS information more quickly and uses the new information for all lookups.
Service Request
19631040
Fixed in version
5.5.32.0
Impact Level
Severity 1
Symptom
Description
Processing CIFS Negotiate Protocol requests (send when establishing a CIFS session) from
unsupported jCIFS clients led to memory corruption and caused the Data Mover to panic with the
following DART panic/fault message: >>PANIC in file: ../sched.cxx at line:
822 : SYSTEM WATCHDOG. The stack trace showed:
0x3625f34: 0x13d10f waitForReboot+0x8b
0x3625f54: 0x13d317 fault_dump+0x67
0x3625f74: 0x13d217 PANIC+0x2b
0x3625f84: 0x13ed3c system_watchdog+0x7c
0x3625f94: 0x1371ad _t_nmi+0x63
0x3625fe0: 0x140610 Sthread_startThread_internal+0xe
Fix
Code was changed to ensure that negotiate reply generation would not cause this memory corruption,
and prevent the panic.
Service Request
19336676
Fixed in version
5.5.32.0
175
Version 5.5.32.4
Impact Level
Severity 1
Symptom
Description
A Data Mover failed to respond to CIFS requests because it entered an infinite loop in NTEvent Log
backup due to a security.evt file corruption.
Fix
Code now checks the event log record validity when backing up the event file.
Service Request
19882164
Fixed in version
5.5.32.0
Impact Level
Severity 2
Symptom
Description
Windows Vista clients could not delete files marked with the read-only attribute, but could when using
other Windows servers. The difference was how the Data Mover responded to delete attempts. When
a read-only file is attempted to be deleted from Windows, Windows will respond with a
STATUS_CANNOT_DELETE error message, which triggers Vista (and other versions of
Windows) to attempt to clear the read-only attribute and try the delete again. Celerra instead responds
with STATUS_ACCESS_DENIED, which triggers Windows 2000 and XP to attempt to clear the
read-only bit, but not Vista.
Fix
Code now returns STATUS_CANNOT_DELETE when trying to set delete on close attribute on a
ReadOnly file.
Service Request
19679190
Fixed in version
5.5.32.1
Control Station
176
The following entries describe fixed problems related to the Celerra

Control Station:
Impact Level
Severity 1
Symptom
Description
Several commands had trouble interpreting the time zone information properly or had errors saying
"ZoneInfo: /usr/java/lib/zi/ZoneInfoMappings (Permission
denied)". This occurred because the last daylight savings time update put in the zi directories with
read permissions only for user. Group and world also needed read permissions.
Fix
Changed the permissions on the zi directory in the jrePatch.tar.gz.src. Made all of the files readable by
user, group, and world, and writable by user.
Fixed in version
5.5.32.2
Version 5.5.32.4
Impact Level
Severity 2
Symptom
Description
When the storage group attributes failovermode or arraycommpath were set to a value other then 0, the
/nas/sbin/navcli h <sp port> - list command failed with the following error message:
Error: port command failed
This version of Core Software does not support Access Logix.
(This line may not always be displayed)
Fix
Setup_clariion and nas_raid commands have been modified to check these parameters and reset
them to zero when they are run. No specific action is needed by the user.
Fixed in version
5.5.32.0
Impact Level
Severity 2
Symptom
Description
The nas_storage modify security command caused segmentation fault if the

username and/or password were too long and the navicli command exceeded the pre-configured
character limit which was 256. An error similar to the following was displayed on the screen and in the
error log:
Error 2237: Execution failed: Segmentation fault: Operating
system signal.
[CLARAPI_FUNCTION_PROTOTYPES.c_clar_change_password]
[root@smithd nas]# nas_storage -m APM00052004495 -security
-username emc12345678910security -password
emc12345678910security -newpassword emc12345678910security1
Fix
The code has been improved by increasing the limit for the navicli command length to 1024 characters,
thus preventing the segmentation fault. There is a CLARiiON limitation of 32 characters for the
username and the password lengths.
Fixed in version
5.5.32.0
177
Version 5.5.32.4
Impact Level
Severity 2
Symptom
Description
Diskmark was not marking new disks and no reason was obvious from the display.
Fix
Fixed the way warnings are propagated from the transaction layer to the CLI so the warning which was
being generated is displayed on the CLI/Celerra Manager.
Service Request
20154652
Fixed in version
5.5.32.0
178
Impact Level
Severity 2
Symptom
Description
Celerra Network Server version 5.5.24.0 added support for file system extend over 2TB, which
introduced this issue. If there was no high capacity mode (file system size greater than 2TB) switch
initiated on a Data Mover, user could not extend a PFS size beyond 2 TB limits.
To query status of high capacity mode, Control Station (CS) queries the Data Mover. If the Data Mover
was panicked before/during request, a code trace was reported. Refer to the Managing Celerra
Volumes and File Systems technical module for more detailed information.
Fix
Code has been improved to check if Data Mover is in contacted state or not before proceeding with file
system extend to eliminate system code trace.
Fixed in version
5.5.32.0
Impact Level
Severity 1
Symptom
Description
CIFS shares for one Compname (CIFS server) became inaccessible following a Data Mover reboot. At
some point prior to the reboot, the user attempted to create a new CIFS server, specifying a network
interface already used by the existing CIFS server. The interface was moved from the existing CIFS
server to the new one in memory on the Data Mover, but this change was not written to configuration
on-disk. When the Data Mover rebooted, the out-of-date configuration information was loaded.
Fix
Code has been modified to ensure that any changes to CIFS servers interfaces are correctly recorded
in the Data Mover configuration files.
Service Request
18266412
Fixed in version
5.5.32.0
Version 5.5.32.4
:
Impact Level
Severity 3
Symptom
Description
If proactive sparing was active on a disk, the disk status was being displayed as UNKNOWN.
Fix
The WebUI will now display the proper status of COPY_TO_HOT_SPARE for the disk.
Fixed in version
5.5.32.0
Impact Level
Severity 2
Symptom
Description
When mounting a file system involved in replication, if the server_mount command was used to
mount the file system on multiple VDMs or on a Data Mover after it had already been mounted on a
VDM, the command succeeded which caused replication resync to fail.
Fix
When mounting the file system on multiple VDMs, the server_mount command will now return
the following: Error 240: <vdm_name2> : file system <file system
name> unavailable for mount on VDM <vdm_name2>, already
mounted on <vdm_name1>.
When mounting a file system on a Data Mover after it's already mounted on a VDM, the
server_mount command will now return the following: Error 4109: <server_name>
: <filesystem_name> : is mounted on <vdm_name>.
Fixed in version
5.5.32.3
CORE

Data Movers:
Impact Level
Severity 1
Symptom
Description
When doing a directory convert from ASCII to UNICODE, during the file system initialization process, if
the root vnode was accessed prior to the sync threads starting, the following panic occurred: DART
panic/fault message: >>PANIC in file: ../ufsvfsop.cxx at line: 373 :
sync: no progress reducing dirty list. The stack trace showed:
0x9bbc5f40: 0x13b8d0 waitForReboot+0x90
0x9bbc5f60: 0x13bad0 fault_dump+0x5e
0x9bbc5f80: 0x13b9d9 PANIC+0x2b
0x9bbc5f90: 0x5f8319 _ZN21UFS_SyncThreadInitial5startEv+0x13d
0x9bbc5fe0: 0x13edf0 Sthread_startThread_internal+0xe
Fix
The code now checks to make sure the sync threads are instantiated prior to the directory convert code
accessing the root vnode, and the Data Mover no longer panics.
Fixed in version
5.5.32.0
179
Version 5.5.32.4
180
Impact Level
Severity 1
Symptom
Description
Running a custom script to collect Data Mover statistics caused Data Mover to panic with the following
DART panic/fault message: >>PANIC in file: ../sched.cxx at line: 822 :
SYSTEM WATCHDOG. The stack trace showed:
0x3655f34: 0x13d1e0 waitForReboot+0x90
0x3655f54: 0x13d3e9 fault_dump+0x67
0x3655f74: 0x13d2e9 PANIC+0x2b
0x3655f84: 0x13ee0c system_watchdog+0x7c
0x3655f94: 0x137279 _t_nmi+0x63
0x3655fe0: 0x1406f8 Sthread_startThread_internal+0xe
Fix
Watchdog panic occurred due to corruption in the mblk where the mblk was pointing to itself. To avoid
this panic, all debug log messages in the PCI error path have been removed.
Service Request
18143850
Fixed in version
5.5.32.0
Impact Level
Severity 1
Symptom
Description
When doing a directory convert from ASCII to UNICODE, during the file system initialization process, if
the root vnode was accessed prior to the sync threads starting, the following panic occurred: DART
panic/fault message: >>PANIC in file: ../ufsvfsop.cxx at line: 373 :
sync: no progress reducing dirty list. The stack trace showed:
0x9bbc5f40: 0x13b8d0 waitForReboot+0x90
0x9bbc5f60: 0x13bad0 fault_dump+0x5e
0x9bbc5f80: 0x13b9d9 PANIC+0x2b
0x9bbc5f90: 0x5f8319 _ZN21UFS_SyncThreadInitial5startEv+0x13d
0x9bbc5fe0: 0x13edf0 Sthread_startThread_internal+0xe
Fix
The code now checks to make sure the sync threads are instantiated prior to the directory convert code
accessing the root vnode, and the Data Mover no longer panics.
Fixed in version
5.5.32.0
Version 5.5.32.4
Impact Level
Severity 1
Symptom
Description
A Data Mover panicked if volume A was deleted between the "volume hyper" and "volpool start"
commands. The above scenario must coincide with a "volpool stop" command on volume B. The
"volpool stop" command for volume B must complete in between the "volume hyper" and "volpool
start" commands for volume A. DART panic/fault message: *** GP exception. Virt
ADDRESS: 754efc. Err code: 0 ***. The stack trace showed:
0x1701178: 0x13d664 waitForReboot+0x90
0x1701198: 0x13d861 fault_dump+0x67
0x17011b8: 0x137914 handle_fatal_fault+0xa
0x1701200: 0x7549d0 _Z13FindAndCreateP7LVolumePi+0xb4
0xd1d8bb70: 0x7577bb _Z10VolumePoolR6Option+0x345
0xd1d8bbf0: 0x1683b3
_ZN11cfgcmd_desc6invokeEPKcPP11CCMDMessageii+0
xxx0xd1d8bc60: 0x20c926
_Z10xml_cfgcmdP7macDataRP8CCMD_Msg+0x38e
xxx0xd1d8bcb0: 0x208bd0
_ZN17macCCMDCommand_DB6invokeEPKcP7macDataRP8C
Fix
A global memory lock was released prematurely if a logical volume was deleted while allocating
volumes to create a checkpoint Savvol. The code now avoids this issue by holding the global lock until
location and configuration of volumes completes.
Service Request
19731348
Fixed in version
5.5.32.0
181
Version 5.5.32.4
Impact Level
Severity 1
Symptom
Description
After rebooting the Data Mover, user could not start replication. The following error was logged to the
console - "Cannot restart replication - sizes differ; block counts
are not the same". The server_log showed that one disk volume was not created because
both paths were concatenated into one command.
2007-11-03 23:00:03: STORAGE: 3: Basic Volume 234 not
created, invalid option 'disk'.
2007-11-03 23:00:03: ADMIN: 3: Command failed: volume disk
234 c0t15l3 disk_id=234 size=8631volume disk 234 c16t15l3
disk_id=234 size=8631
The basic volume was not created, so the metavolume for the SFS was not fully built - therefore the
size of the PFS and SFS meta volumes did not match, which triggered the error message.
Fix
In this case, replication was affected. Other symptoms such as file systems not being mounted, shares
not being exported, etc. could have resulted from this same issue. Because the Data Mover image
loader parses configuration files in 32K chunks, if the next byte after any 32K chunk in the boot.cfg file
was a line feed character, the boot process did not read it and the line was concatenated to the
previous line. This caused both lines (and subsequently, everything in that particular 32K chunk) to fail
to execute when the Data Mover was booted. Any services depending on these lines/32K chunk failed.
The Data Mover boot process has been fixed and now parses the boot.cfg correctly.
Service Request
19479452
Fixed in version
5.5.32.3
CSA
182

Startup Assistant:
Impact Level
Severity 2
Symptom
Description
When using the CSA, user could only link directly to Powerlink from the welcome screen. The label
implied that users could go directly to the Celerra Tools section of Powerlink.
Fix
The CSA label has been renamed to Powerlink, and CSA now provides instructions to navigate to the
Celerra Tools area if needed.
Fixed in version
5.5.32.0
Version 5.5.32.4
Impact Level
Severity 3
Symptom
Description
A note at the bottom of the CSA "Pre_Configuration" screen stated "Note: To modify your
configuration information, click Back until you reach the
required screen. If you are confident the configuration data
you entered is correct, click Back and then Next to rerun
pre-configuration." Users were confused about the need to click both the "back" and the
"next" buttons to continue.
Fix
Modified the note to read "Note: To modify your configuration

information, click Back until you reach the required screen.
If you are confident the configuration data you entered is
correct, click Next to continue."
Fixed in version
5.5.32.0
Impact Level
Severity 3
Symptom
Description
Users were confused and mistakenly interpreted the Cancel button in CSA to mean "cancel the
current operation" when in fact is meant to close CSA.
Fix
When the user cancels CSA, a confirmation message will now be presented to warn that completing
the cancel operation will result in the closing of the CSA window not the operation.
Fixed in version
5.5.32.0
Impact Level
Severity 2
Symptom
Description
During Celerra installation using CSA, after filling in information on the Control Station initialization
screen and clicking "Next", user received message, "Configuration completed
successfully. Now you can access the Control Station over the
network". Next click "OK". At this point, CSA attempted to log in to the CS to verify the
Celerra model type. During the Control Station setup portion of the CSA initialization, the user
received an error message stating "Error in finding the Celerra Model Type.
Action: Make sure the network connection is good.Try Again!
In this particular case, the root password had been changed.
Fix
Users will now be presented with additional recommended actions in the event that CSA is unable to
connect to a Celerra. The recommended action will be "To verify if the IP has been
set on the Control Station, try to ping the Celerra IP or
manually connect with an SSH application" .
Fixed in version
5.5.32.0
183
Version 5.5.32.4
184
Impact Level
Severity 2
Symptom
Description
While running the Registration Wizard invoked via CSA, on the "Configuration Collection: Automatic"
screen, user received the following: "An instance of log_config is already
running" message. The message was misleading and no further direction was provided to the
user.
Fix
New recommended actions added in the event Registration Wizard attempts to complete auto data
collection and finds if log_config is already in use. The recommended action will look like "The
log_config service is currently in use by another user or
application on the control station, there can only be one
instance running at a time, use the back button to retry the
collection process after waiting approximately 4 minutes for
the current process to complete".
Fixed in version
5.5.32.0
Impact Level
Severity 2
Symptom
Description
On Health Check screen, if the user encountered an error ending in "9999", no error message was
displayed.
Fix
Error messages are now provided for the error codes ending with 9999. Error codes that ends with
9999 are unrecoverable / unknown error. The recommended action is to contact the service provider.
Fixed in version
5.5.32.0
Impact Level
Severity 2
Symptom
Description
In the CSA Initialization screen, the primary and secondary DNS server fields didn't display DNS
servers present on the same subnet as the one the client machine was connected to. If the user
proceeded further, the following error was displayed "Not on same subnet" although the client
machine may be on the same subnet.
Fix
Code changes were made such that in the Initialization (Set up Control Station for accessing over
network) screen, the DNS server that belongs to the same subnet is always displayed as the primary
DNS. If none of the DNS belong to the same subnet, then they are displayed in the order they appear
in the output of ipconfig /all command in the client machine.
Fixed in version
5.5.32.0
Version 5.5.32.4
Impact Level
Severity 2
Symptom
Description
CSA failed to pick the right client network if the machine was configured with multiple network
interfaces (like one or more VMWare Virtual network interfaces, physical NICs, etc). It selected the
first valid ethernet network connection from the 'ipconfig /all' command output after skipping the WLAN
and/or VPN connections. When the user connected the Celerra to the VMWare interface or any other
ethernet interface which comes later in the output stream, the following error was displayed: "Not
on the same subnet".
Fix
CSA has been modified to recognize that there are multiple IP interfaces/NICs that it cannot
differentiate and now displays the possible NICs/IP addresses on a popup screen and lets the user
select the one that is connected to the Celerra.
Fixed in version
5.5.32.0
Impact Level
Severity 2
Symptom
Description
While using the Celerra Startup Assistant, after the Cable check section is completed, if the user
clicked the "Next" button and then clicked the "Back" button, the application hit the following exception:
DEBUG 2007-06-27 20:10:01,509
[com.emc.celerra.startup.framework.WizController]: calling
next() method
ERROR 2007-06-27 20:10:04,962
[com.emc.celerra.startup.healthcheck.screens.WizHealthCheckFo
rm]: Exception occurred in method displayStatus :
'java.lang.ArrayIndexOutOfBoundsException: 2'
DEBUG 2007-06-27 20:11:28,146
[com.emc.celerra.startup.framework.WizController]:
validation code = 2
Fix
The exception was removed. There was a wrong setting for the array size in the health check code.
Setting the correct array size solved the problem.
Fixed in version
5.5.32.0
185
Version 5.5.32.4
186
Impact Level
Severity 2
Symptom
Description
The CSA did not validate the host name and accepted invalid values such as .illegalhostname, jackjill.; a:b?c~d%e . These illegal hostnames can cause problems in the network.
Fix
CSA now performs host name validation on Control Station settings page following RFC1123 and
RFC952
Fixed in version
5.5.32.0
Impact Level
Severity 3
Symptom
Description
The creation of LUNs with size greater than the available size produced no error message and creation
succeeded with the available size. Message should display that there is no space available for
requested LUN size and operation should fail.
Fix
Code has been improved to verify space availability against the requested size during NFS or iSCSI
wizards and in case it is less, operation will fail and appropriate message will be presented to the user.
Fixed in version
5.5.32.0
Impact Level
Severity 1
Symptom
Description
After starting CSA, users received a "duplicate IP address" message. However, no duplicate IP
address problem existed. This problem occurred if the client PC running CSA was in a different subnet
than the Control Station, but the Control Station IP address was pingable from the client PC.
Fix
Users will now receive a message stating "the IP address of your client PC does not match the subnet
of the Celerra" (a requirement for CSA), which is the expected behavior.
Fixed in version
5.5.32.0
Version 5.5.32.4
FileMover
The following entries describe fixed issues related to the Celerra

FileMover:
Impact Level
Severity 2
Symptom
Description
When a DHSM-enabled file system had an http connection with no admin name specified, if an
administrator tried to delete the http connection; the following code trace was reported: Error
2237: Execution failed: admin_name_not_void: Precondition
violated. [CDMS_CONNECTION_ENTRY.full_admin_name]
Fix
There was no void object check in 'fs_dhsm delete action with recall_policy yes' command code flow.
The appropriate void check has been added and the command will now succeed.
Fixed in version
5.5.32.0
Impact Level
Severity 2
Symptom
Description
Offline FileMover files were not reporting Offline bit via CIFS when: 1) DHSM backup mode was set
to passthrough, 2) file access requests were marked to indicate they had backup intent, and 3) the
user accessing files was a Backup Operator. This meant that backup applications could not use the
offline status of files as criteria on whether or not to back up files.
This behavior was introduced in 5.5.27.5. Previously, the offline status was reported in all situations.
This led to certain backup software recording the offline status as part of the file, then attempting to
restore the attribute during file restores. Celerra will drop the offline bit, as after a restore a file is no
longer offline, but when files were restored to other locations (for example, a users local NTFS drive),
the offline attribute was maintained, even though the file was not offline.
Fix
This change reverts back to the pre-5.5.27 behavior. Now, offline attributes of files are reported in all
cases. The original change was made to work around backup software behavior, but prevents
applications from seeing the true offline status of files.
Service Request
19671500
Fixed in version
5.5.32.0
187
Version 5.5.32.4
Impact Level
Severity 2
Symptom
Description
The fs_dhsm -connection <filesystem> -info command displayed "[ Migration: FAIL ]"
status, even though the connection was valid.
Fix
The code now cleans up the status from a previously-used connection to ensure it's not reused the next
time this command is issued.
Service Request
18825520
Fixed in version
5.5.32.0
Install/Config/
Upgrades
188
The following entries describe fixed issues related to installations,

configurations and upgrades:
Impact Level
Severity 2
Symptom
Description
There is an interoperability issue with the CISCO MDS switches when the Data Mover and a
CLARiiON are configured in different VSANS and the Cisco Intra VSAN Routing with Network Address
Translation is enabled in firmware version 3.1(2a) or later. The IVR NAT feature adds significant delay
in PLOGI Accept on the FC SAN causing the Data Mover to fail to boot.
Fix
The Data Mover's BIOS/Post has been updated to resolve this issue. The Data Movers need to be
running a BIOS/Post revision equal or greater than those listed below. For platforms not listed, the
Cisco IVR NAT feature is not supported and the system must be configured with the Data Mover and
Storage Processor in the same VSAN. Refer to the NAS Support Matrix for more detailed information.
System _______Post BIOS
NS 40/20 ______1.46 3.58
NS 80/NSX____ 2.60 5.14
NS 700/704 ____.39 3.56
Fixed in version
5.5.32.0
Version 5.5.32.4
Impact Level
Severity 2
Symptom
Description
When the Pre-Upgrade Health Check (PUHC) determined that the NAS Storage API was missing, it
displayed a recommended action to correct the problem. However, for step 3 it read:
3. Run "tar -xzvf $storage_tgz" to unpack the nasStorageAPI
tarball.
$storage_tgz is incorrect, instead a file name should be displayed.
Fix
The proper file name (i.e. nasStorageAPI-6.0-9.tar.gz) is now displayed.
Fixed in version
5.5.32.0
Impact Level
Severity 2
Symptom
Description
During removal or upgrade of NAS Storage API RPM phase of the NAS code upgrade, the following
error message "Could not delete /nbsnas/etc/Navisphere -- please
delete it manually was displayed , although those directories were removed and RPM was
deleted or upgraded. This issue was introduced in 5.5.31.0 .
Fix
The new code resolves these problems so that erase/install/upgrade of NAS Storage API RPM
successfully completes without generating any of these error messages.
Fixed in version
5.5.32.0
Impact Level
Severity 2
Symptom
Description
When upgrading a NAT (Network Address Translation) setup for public access to the CLARiiON
attached to an NS20/NS40 system, the /nas/sbin/clariion_mgmt
-upgrade_to_proxy_arp command failed with the error: Not an integrated
system even though the Celerra was an integrated system.
Fix
Fixed the clariion_mgmt command to allow for this upgrade to take place successfully.
Fixed in version
5.5.32.3
189
Version 5.5.32.4
iSCSI
190
The following fixed problems are related to Internet Small Computer

Storage Interface (iSCSI):
Impact Level
Severity 2
Symptom
Description
An iSCSI LUN and its snaps became inaccessible after a snap was deleted. The Data Mover had to be
failed over in order to resolve this issue.
Fix
A deadlock occurred when an iSCSI snap deletion process accessed a version file locked by another
thread. When this occurs, the delete process is now deferred until the current locks are released to
avoid the deadlock.
Service Request
19457680
Fixed in version
5.5.32.0
Impact Level
Severity 1
Symptom
Description
Memory leak during iSCSI replication led to a Data Mover panic with the following DART panic/fault
message: >>PANIC in file: ../addrspac.cxx at line: 275 :
Out
of memory. The stack trace showed:
0xeec6eac0: 0x13d1d3 waitForReboot+0x8b
0xeec6eae0: 0x13d3db fault_dump+0x67
0xeec6eb00: 0x13d2db PANIC+0x2b
0xeec6eb10: 0x139184
0xeec6eb30: 0x187636 _Z11more_memoryiP9free_list+0x4e
0xeec6eb50: 0x1876e3 malloc+0x89
0xeec6eb80: 0xaa3852
xxx_ZN12httpd_thread7rprintfEP11per_requestPcz+0x
0xeec6eba0: 0xaa3067
xxx_ZN12httpd_thread16send_http_headerEP11per_req
Fix
The memory leak was resolved by removing unnecessary duplication of the message block in packets
sent from the destination side of iSCSI replication to the source.
Service Request
18966980
Fixed in version
5.5.32.0
Version 5.5.32.4
Impact Level
Severity 1
Symptom
Description
When user continuously invoked the iSCSI replication commands (mostly query commands), the Data
Mover panicked with the following DART panic/fault message: >>PANIC in file:
../include/stream.hxx at line: 1014 : out of msgb. The stack trace
showed:
0x3627c60: 0x13d1d3 waitForReboot+0x8b
0x3627c80: 0x13d3db fault_dump+0x67
0x3627ca0: 0x13d2db PANIC+0x2b
0x3627cb0: 0x16c26f _ZN15Buffer_MsgbPool3getEv+0xa7
0x3627cd0: 0x16a6bf allocb+0x59
0x3627d00: 0x33f6b4 _Z10tcp_outputP5tcpcb+0xc16
0x3627de0: 0x33b1ce
xxx_Z9tcp_inputP13IPInterface_tP4msgbiiPi+0x1c62
0x3627e60: 0x3120bf ip_input+0x565
Fix
Code has been fixed to prevent this out of memory issues on source side when running iSCSI
replication commands.
Service Request
19565908
Fixed in version
5.5.32.0
NDMP/Backup
The following entries describe fixed problems pertaining to Network

Data Management Protocol and/or Backup:
Impact Level
Severity 2
Symptom
Description
While using EDM (EMC Data Manager) and NDMP version 2, user selected a single file for non-DAR
alternate restore. The EDM alternate file level restore was changed to the original path, but the files
were not restored.
Fix
Code has been modified to ensure that files are restored to the correct place when doing single file
non-DAR restores with NDMP v2.
Service Request
19044078
Fixed in version
5.5.32.0
191
Version 5.5.32.4
Impact Level
Severity 1
Symptom
Description
An NDMP VBB backup aborted when run on a checkpoint of an IP Replication target file system.
When IP Replication playback is run, the target file system is unmounted/remounted so clients always
see a consistent view. The unmount of the target file system is the reason a checkpoint needs to be
used for backup.
Fix
The unmount/remount no longer impacts VBB backups running on a checkpoint of the target file
system.
Fixed in version
5.5.32.0
Networking
192
The following entries describe fixed issues related to Networking:
Impact Level
Severity 2
Symptom
Description
After rebooting a Data Mover that had network interfaces marked DOWN, user found that the routers
ARP tables had been updated to point to the rebooted Data Movers DOWN interfaces. The DOWN
interfaces were only configured for use in DR situations, and the IP addresses were active on another,
redundant Data Mover. The logical interface was down, but the physical NIC was connected and up.
This caused a temporary access outage since the ARP tables now pointed to the wrong (secondary)
Data Mover.
Fix
When a Data Mover reboots or fails over, it generates gratuitous ARP messages to inform the network
that its physical location and MAC address may have changed. These messages were being sent
erroneously even if an interface was marked DOWN. Gratuitous ARP messages are no longer sent on
DOWN network interfaces.
Service Request
19582978
Fixed in version
5.5.32.0
Impact Level
Severity 2
Symptom
Description
When a client issued the extended FTP command EPSV ALL , the FTP session hung.
Fix
The code has been improved to comply with RFC 2428 to prevent the FTP session from hanging when
using this extended FTP command.
Fixed in version
5.5.32.0
Version 5.5.32.4
Impact Level
Severity 1
Symptom
Description
User was unable to unmount a file system after an FTP data transfer failed because a connection to
the client could not be established. The Data Mover panicked with the following DART panic/fault
message: >>PANIC in file: ../dskdump.cxx at line: 1437 : User
Panicked Server (Sparse Dump). The stack trace showed:
0xbc8b9ad0: 0x13b7d3 waitForReboot+0x8b
0xbc8b9af0: 0x13b9d2 fault_dump+0x5e
0xbc8b9b10: 0x13b8db PANIC+0x2b
0xbc8b9b20: 0x7f6224 _Z11paniccfgcmdR6Option+0x178
0xbc8b9c60: 0x15b35c _ZN11cfgcmd_desc6invokeEPKcii+0x124
0xbc8b9cd0: 0x1eee6a _Z10xml_cfgcmdP7macData+0x282
0xbc8b9d10: 0x1ebd8c
xxx_ZN13macCommand_DB6invokeEPKcP7macData+0x74
0xbc8b9d40: 0x1ec0a5 _Z13mac_DocParserP7macData+0x19
Fix
Before calling the data transfer loop, code now verifies that the data connection has been accepted to
prevent this condition.
Service Request
19520252
Fixed in version
5.5.32.0
NFS
The following entries describe fixed Network File System problems:
Impact Level
Severity 2
Symptom
Description
Users were able to export file systems for NFS using an alias that was the same as the file system
pathname. Afterwards, the Data Mover showed two entries for the export when NFS clients asked for
the Data Movers export list via the showmount command.
Fix
Export aliases are alternate names for exports, and cannot have the same name as the pathname
being exported. The Celerra would attempt to create the alias anyway, and inadvertently create two
export entries. If users specify an alias name and pathname that are the same, the code now simply
exports the pathname and does not create the alias.
Service Request
20037294
Fixed in version
5.5.32.0
193
Version 5.5.32.4
Replication
194
The following entries describe fixed problems encountered while

using Celerra Replicator, SnapSure or Checkpoints:
Impact Level
Severity 1
Symptom
Description
A differential fs_copy of a checkpoint of a file system that was the source in a replication session failed
with: Error 3105: invalid filesystem specified.
Fix
A check was failing because it assumed the differential fs_copy was being issued in order to start IP
replication. In that case the fromfs ckpt needs to be older than the time when replication was
established. However, in this case the target of the differential fs_copy is different from the target of
the replication session. The check is now only performed when the differential fs_copy is issued to the
target of the replication session.
Service Request
19863544
Fixed in version
5.5.32.0
Impact Level
Severity 1
Symptom
Description
During a replication failover, checkpoints created on the secondary side were stamped with the latest
deltaset. When synchronizing back to the source side after a failback, not all changes were
incorporated, causing file system corruption and the following panic: Assertion failure:
'status == File_OK' on line 538 in file '../ufsacl.cxx'
>>PANIC in file: ../assert.c at line: 26 : assert failed. The
stack trace showed:
0x8fea7c60: 0x14694c waitForReboot+0x90
0x8fea7c80: 0x146b5f fault_dump+0x67
0x8fea7ca0: 0x146a49 PANIC+0x29
0x8fea7cb0: 0x1ce4b1 _assert+0x51
0x8fea7cd0: 0x8363c6 _ZN9UFS_Vnode6setAclEiPvmS0_+0x252
0x8fea7d10: 0x11096e6 _ZN10MGFS_Vnode6setAclEiPvmS0_+0x34
0x8fea7d30: 0x7011bd
_ZN15File_NamingNode5setSDEP15SDStorageHeaderb
0x8fea7e10: 0xbfa793
_ZN11smb_sdShare26share_SD_Administrators_RWEv
Fix
Checkpoints created during replication failover are now stamped with the correct deltaSet number to be
played back. Using this deltaSet number, the resync operation will work correctly between the PFS and
SFS.
Fixed in version
5.5.32.3
Version 5.5.32.4
Impact Level
Severity 1
Symptom
Description
When attempting to start or restart replication, after upgrading to NAS Code version 5.5.27 or later,
messages similar to the following were seen in the server_log:
2007-07-31 08:54:30: VRPL: 3: Replication is not Allowed for
SrcVolume id > 16383 (here srcvol = 51358)
2007-07-31 08:54:30: VRPL: 3: Invalid Argument specified.
2007-07-31 08:54:30: ADMIN: 3: Command failed: replica start
srcvol=51358 savevol=2005 cfgvol=2007
2007-07-31 08:54:30: VRPL: 3: Replication is not Allowed for
SrcVolume id > 16383 (here srcvol = 51358)
Note that the start srcvol (the meta volume that the FS is built on) was 51358 or higher than 16383.
Fix
There was a change introduced in version 5.5.27.5 to prevent Replication (V1) from being started
against a file system with a source volume ID > 16383. This limitation has been removed, refer to the
Configuration Guidelines table in the release notes for specific file system limitation information. Also,
the nas_fs -list command can be used to list all file systems and the source volume_id they
are built on.
Fixed in version
5.5.32.3
Impact Level
Severity 2
Symptom
Description
After performing multiple replication failover and failback attempts, the fs_replicate
resync command failed with a Precondition violated error. Once this issue was
encountered, other commands that accessed file system information, such as the nas_fs command,
also failed with the same error. File system-related pages in the WebUI were also affected.
Fix
The fs_replicate reverse caused a malformed entry in the filesys file of the NAS
database. This didnt cause immediate impact, but caused a later fs_replicate resync process to
further corrupt the entry in the filesys file, which had to be manually repaired.
The fs_replicate reverse and fs_replicate resync commands have been
updated to prevent this kind of NAS database corruption from occurring.
Service Request
16360212
Fixed in version
5.5.32.0
195
Version 5.5.32.4
196
Impact Level
Severity 1
Symptom
Description
There was a small window between async unmount requests for the oldest checkpoint during
inactivation. If the Data Mover received the mount command for the oldest checkpoint in this window, a
CIFS dead lock condition occurred which in turn caused the Data Mover to panic with the following
DART panic/fault message: I/O not progressing LastVol touched not
Known. The stack trace showed:
0x4001e90: 0x13d1db waitForReboot+0x8b
0x4001eb0: 0x13d3e3 fault_dump+0x67
0x4001ed0: 0x13d2e3 PANIC+0x2b
0x4001ee0: 0x703afd _ZN9RawIO_IRP13checkProgressEv+0x107
0x4001fd0: 0x7039ec _ZN17IrpProgressThread5startEv+0x6
Fix
Code has been modified to prevent this deadlock from occurring.
Service Request
19263178
Fixed in version
5.5.32.0
Impact Level
Severity 1
Symptom
Description
After restarting replication, destination file systems were left in an inconsistent state and caused the
Data Mover to panic when accessed. For IP Replicated file systems with checkpoints, during a
checkpoint refresh operation the Control Station stamps the checkpoint entry with the current
replication delta set number.The delta set number is returned from the Data Mover as part of xml
response. Previously, the checkpoint refresh operation stamped the checkpoints with a delta set
number even if replication was inactive. If these checkpoints were selected for an out-of-sync restart
operation, it corrupted the SFS because the delta set represented by that checkpoint had not yet been
replayed on the SFS.
Fix
A checkpoint will only be stamped with the delta set number when replication is active. If replication is
inactive and a checkpoint is refreshed, that checkpoint will not be stamped with a delta set. This fix will
prevent the file system inconsistency issue.
Service Request
20724594
Fixed in version
5.5.32.2
Version 5.5.32.4
Impact Level
Severity 1
Symptom
Description
File system corruption on the secondary side of an IP replication pair caused the Data Mover to panic
with the following DART panic/fault message: readBlock: Bad Block Index . The stack
trace showed:
0xa6b8bce0: 0x13d1e0 waitForReboot+0x90
0xa6b8bd00: 0x13d3e9 fault_dump+0x67
0xa6b8bd20: 0x13d2e9 PANIC+0x2b
0xa6b8bd30: 0x60da21 _ZN14UFS_FileSystem28markCorrupted
xxxFsForUnmount
0xa6b8bdc0: 0x61e0c8 _ZN14UFS_FileSystem9readBlockEliRP
xxx13Buffer_Cac
0xa6b8be30: 0x619f1b
xxx_ZN9UFS_Vnode5breadEjR11File_Statusi+0x167
0xa6b8be80: 0x619f83
xxx_ZN9UFS_Vnode9blockReadEjR11File_Status+0x4b
0xa6b8beb0: 0xb7a4f5
xxx_ZN10DHSM_Vnode9blockReadEjR11File_Status+0x38
The IP Replication fs_replicate -restart command used a "less than or equal to" equation to select the
checkpoint to use as the -fromfs based on the current delta set number on the destination side. This
allowed the restart to select a checkpoint which represented a delta set still in the savvol, and not yet
replayed to the SFS. This meant that some changes were not copied to the SFS, leaving it inconsistent
with the source, and causing the Data Mover to panic.
Fix
The restart command now uses a simple "less than" equation to ensure that the wrong checkpoint is
not selected as the -fromfs.
Service Request
20892866
Fixed in version
5.5.32.0
197
Version 5.5.32.4
SRDF
198
The following entries describe fixed problems pertaining to

Symmetrix Remote Data Facility:
Impact Level
Severity 1
Symptom
Description
The system was configured with a mix of SRDF-protected LUNs from a Symm which were mapped to
the SRDF-protected Data Movers, and some local LUNs from a CLARiiON which were mapped to
non-SRDF Data Movers. Even though they were mapped to separate Data Movers, the combination of
both types of storage (protected and unprotected) caused a problem for symapi and the nas_rdf script.
Fix
The dismark checks performed on the Target system now ignore any local, unprotected, devices which
are listed in the symapi database from the source side. Only SRDF-protected LUNs are checked during
fail over.
Fixed in version
5.5.32.1
Impact Level
Severity 2
Symptom
Description
During the nas_rdf -activate command, the failover of the Data Movers failed with a "replace_volume:
volume is unreachable" error, followed by a list of all the LUNs used by that Data Mover. As a result the
nas_rdf failover was not successful. This happened if the symapi database has been updated (by a
nas_storage command, nas_diskmark command, server_devconfig command, nas_rdf command,
etc) while the SRDF links between the Symmetrix were down.The most typical example is if the rdf
links are dropped for a nas code upgrade. In this situation, the remote information was removed from
the symapi database, and symapi could not determine the correct R1/R2 relationship during the
nas_rdf -activate.
Fix
The storage algorithms on the Control Station have been modified to check the diskmark information
to build the R1/R2 mappings when SymAPI data is not available. This will allow the nas_rdf -activate to
complete.
Service Request
20704954
Fixed in version
5.5.32.2
Version 5.5.32.4
Impact Level
Severity 1
Symptom
Description
When the user issues a fs_timefinder Mirror refresh command for a TimeFinder
snap, the TimeFinder application internally performs this operation in two steps:
1) A Symmetrix control command is issued to establish the mirror
2) A device synchronization command is issued to update the symapi database
Both of these operations require a Symmetrix Gatekeeper device lock. If the second step fails to
acquire lock for the Gatekeeper (e.g. the Gatekeeper was unavailable for some reason), then the first
step is ignored which leaves inconsistency between NAS database and the symapi database: the
snap is in mirror established state, but it is in mirror split state in the NAS database.
When trying to mount the snap in this scenario, the Data Mover panicked with the following DART
panic/fault message: >>PANIC in file: ../BVolumeIrp.cxx at line: 323 :
IO failure on Vol:7, blkNo:16, SKey:0x24. The stack trace showed:
0x3887e80: 0x13d664 waitForReboot+0x90
0x3887ea0: 0x13d861 fault_dump+0x67
0x3887ec0: 0x13d761 PANIC+0x29
0x3887ed0: 0x77669b
xxx_ZN8SgIO_IRP9irpFailedE14BlockIO_Statusy+0x30f
0x3887f20: 0x809736
xxx_ZN14CamDisk_Device8completeEP10ccb_header+0xe6
0x3887f70: 0x789e42 _ZN8Scsi_Xpt16processCallbacksEv+0x3c2
0x3887fd0: 0x789e7a _ZN18Scsi_serverInitial5startEv+0x6
0x3887fe0: 0x140df8 Sthread_startThread_internal+0xe
Fix
This problem occurred due to an un-necessary call to synchronize the symapi database after a
Symmetrix control command to establish a mirror. This call needed a Gatekeeper lock and the request
for which failed even after a successful control command.
Any device synchronization call failure after a successful control command will not change the mirror
status of the snap in NAS database and will cause an inconsistency in the BCV device status between
NAS database and the symapi/backend database.
The fix for this issue is to not make a device synchronization call after the control command call since
the TimeFnder symapi control command automatically synchronizes the symapi database at the end
of the control operation.
Service Request
Number
20558970
Fixed in version
5.5.32.3
199
Version 5.5.32.4
Storage
200
The following entries describe fixed storage-related issues:
Impact Level
Severity 1
Symptom
Description
The Data Mover panicked with the following DART panic/fault message: >>PANIC in file:
../sched.cxx at line: 819 :
0xd2e61904: 0x13d600 waitForReboot+0x90
0xd2e61924: 0x13d7fd fault_dump+0x67
0xd2e61944: 0x13d6fd PANIC+0x29
0xd2e61954: 0x13f459 system_watchdog+0x77
0xd2e61964: 0x137481 _t_nmi+0x63
0xd2e619c0: 0x1886a5 _Znwm+0x15
0xd2e619d0: 0x1886d2 _Znam+0xc
0xd2e619e0: 0x653c75
xxx_ZN20UFS_DirHashInfo_dir3C1EP11Local_Vnode+0xc
Fix
This panic occurred after receiving an invalid Fibre Channel frame, which caused the Data Mover to
attempt to allocate a very large amount of memory. The code now performs a sanity check in this
situation, and will not attempt to allocate unrealistically large amounts of memory.
Service Request
20458312
Fixed in version
5.5.32.0
Impact Level
Severity 1
Symptom
Description
While performing the procedure to repair uncorrectable sector errors, the mapblock utility hung. This
utility identifies files which were affected by the uncorrectable sector.
Fix
The code has been changed to make sure that pointers are not overwritten when mapblock searches
for file names in very large directory structures.
Service Request
19293946
Fixed in version
5.5.32.0
Version 5.5.32.4
UFS
The following entries describe fixed UNIX File System issues:
Impact Level
Severity 1
Symptom
Description
When doing a directory conversion from ASCII to UNICODE, syncThreads used a NULL pointer which
caused the Data Mover to panic with the following DART panic/fault message: *** Page Fault
Interrupt. Virt ADDRESS: 879de3 Err code: 0 Target addr: d2 **.
The stack trace showed:
0x21ef154: 0x146380 waitForReboot+0x90
0x21ef174: 0x146593 fault_dump+0x67
0x21ef194: 0x13c560 handle_fatal_fault+0xa
0x21ef230: 0x7f4ef8
xxx_ZN18UFS_FileSystemKind6initFSEPKcP9IO_Objectii
0xbf39b930: 0x117054f
xxx_ZN19DHSM_FileSystemKind6initFSEPKciR6OptionP
0xbf39b9b0: 0x7afd76
xxx_Z25File_MountCommandInternalR6OptionP18VFS_Fi
0xbf39bc00: 0x7b1ca5
xxx_Z23File_ServerMountCommandR6OptionP18VFS_File
0xbf39bc30: 0x7b25a1 _ZN12filecmd_desc6invokeER6Option+0x6d
Fix
Fixed the syncThread code in the directory convert code path to check for a NULL pointer and handle
this error correctly.
Service Request
Fixed in version
5.5.32.0
201
Version 5.5.31.6
Version 5.5.31.6
Celerra
Management
Applications
202

Impact Level
Severity 2
Symptom
Description
When using the XML API, read/write/total requests were incorrect for the SymmDeviceStatus structure.
Instead of reporting the number of read, write, and total requests, these fields were pointing to a number
of bytes read, number of bytes written, total number of bytes (read+write).
Fix
The read/write/total requests are now returned correctly by XMLAPI.
Fixed in version
5.5.31.0
Impact Level
Severity 1
Symptom
Description
When a Celerra was connected to a DMX-3 backend, some of the devices owned by the Celerra were
not reported correctly by JServer.
Fix
JServer only supports a maximum of 16K devices. Because the DMX-3 can have up to 64K, the list of
available devices owned by the Celerra was being displayed incorrectly. JServer was enhanced to
record only the DMX-3 devices that are actually being used by the Celerra. This allows us to display
the correct devices.
Fixed in version
5.5.31.0
Impact Level
Severity 2
Symptom
Description
In Celerra Manager, when a mount option was added in the Mount Properties page for the file system,
the change wasn't immediately seen in the Expanded Mount view. A refresh was needed to sync the
Expanded view with the new mount options.
Fix
When a mount option is added to a file system mount, it will now be immediately reflected in the
Expanded view of the mounts.
Service Request
17448008
Fixed in version
5.5.31.0
Version 5.5.31.6
Impact Level
Severity 2
Symptom
Description
When starting the registration wizard, the user is prompted for a password. If user entered the same
wrong password string twice, the password entry was disabled and user had to cancel and restart the
registration wizard.
Fix
User can now enter the same wrong password string multiple times without being locked out. User
must still enter the correct password to continue, and can now do so using the back button.
Fixed in version
5.5.31.0
Impact Level
Severity 1
Symptom
Description
User encountered problems related to the root file system filling, which could also result in data loss or
data unavailability. The root file system can fill due to a variety of reasons which include log files,
secmap entries, and others. An event, CFS event ID 4 (Crossed the root filesystem size threshold),
was posted in /nas/log/sys_log, but it was not brought to the administrator's attention and was easily
missed.
Fix
Code now generates an alert for display on the Home Status page of Celerra Manager when this event
occurs. The event will also affect the overall system status in Celerra Manager so the administrator will
be aware that a problem exists.
Fixed in version
5.5.31.0
Impact Level
Severity 3
Symptom
Description
On the Connect Home Properties page, for any disabled transport, both the Primary and Secondary
main fields (i.e. Email-SMTP, FTP Server and Phone Number) were configured. When the user made
the Primary field empty but the corresponding secondary field was not empty, and pressed the Apply
button, a popup dialog was displayed with null rather than a proper message.
Fix
On the Connect Home Properties page for any disabled transport, when the user makes the primary
field empty but the secondary is not empty, and presses the Apply button, a popup dialog will now
display the following: When a transport is set to Disabled, you cannot
apply Optional secondary fields when primary fields are blank.
Fixed in version
5.5.31.0
203
Version 5.5.31.6
204
Impact Level
Severity 2
Symptom
Description
When a user tried to add a new DNS server IP to an existing Data Mover DNS configuration, the
following error was displayed: DNS Server x.x.x.x is already configured
with some other Domain.
Fix
User can now add a new DNS server IP up to a maximum number of three servers.
Service Request
19234010
Fixed in version
5.5.31.0
Impact Level
Severity 1
Symptom
Description
When using the CIFS Server wizard - the domain password was written in clear text into the apache
log. This allowed users with access to that log to obtain the domain password.
Fix
Code was modified to ensure passwords are no longer displayed in clear text in the apache log.
Fixed in version
5.5.31.0
Impact Level
Severity 3
Symptom
Description
Upon entering an invalid email address, if the user then left this field (move control outside the field)
and then hit the APPLY button, an Invalid Field pop-up message was displayed. If the user then hit OK
to accept the pop-up, the Apply was processed with the invalid field.
Fix
Field validations are now called when the user submits the form (hit the APPLY button). If a field is
invalid, the user will not be allowed to submit the operation without correcting the error.
Fixed in version
5.5.31.0
Version 5.5.31.6
Impact Level
Severity 2
Symptom
Description
On the Connect Home Properties page, if double quotes were entered in fields such as Site ID,
Celerra Serial Number, Number to Dial, the fields were not validated and the incorrect values were
passed to the lower layers. As a result, a generic error message invalid field was returned.
Fix
Code has been modified to ensure that when the user enters the double quotes in any of the fields, a
specific pop-up message (e.g., Invalid Celerra serial number) is displayed stating it
is an invalid value.
Fixed in version
5.5.31.0
Impact Level
Severity 2
Symptom
Description
The CNS platform didn't allow the Celerra Serial Number field to be modifiable from the UI when
logged in as root.
Fix
The CNS platform now allows the Celerra Serial Number to be modified from the UI when logged in as
root.
Fixed in version
5.5.31.0
Impact Level
Severity 3
Symptom
Description
When user attempted to create a file system with a name that already existed, the message displayed
was: File system already exists.
Fix
Changed the error message to be the following: The file system you are trying to
create already exists. Specify a file system using a different
name.
Fixed in version
5.5.31.0
205
Version 5.5.31.6
Impact Level
Severity 2
Symptom
Description
If a user entered an e-mail address with an invalid format on the Connect Home Properties page, the
following pop-up was displayed: invalid field. The pop-up did not specify which field was
invalid.
Fix
The pop-up now displays the invalid values, similar to Invalid Recipient email
address(es), and asks the user to correct them.
Fixed in version
5.5.31.0
Impact Level
Severity 3
Symptom
Description
In the CIFS creation screen of the Celerra Startup Assistant, if a user changed the default values, and
then changed focus from that page to a new page. When user returned back to the page where the
values were changed, all fields reverted to their default values, erasing any changes made.
Fix
The changes made to the fields are now retained under this condition.
Fixed in version
5.5.31.0
Impact Level
Severity 2
Symptom
Description
In the Celerra Startup Assistant (CSA), when attempting to create an NFS export, if there was no
remaining available space in the storage pool, the screen hung indefinitely, the available storage space
field displayed a wrong value, and no error message was displayed.
Fix
When attempting to create an NFS export and there is no remaining space in the storage pool, an
error message will display indicating that there is no remaining space in the storage pool.
Fixed in version
5.5.31.0
CIFS
206

problems:
Version 5.5.31.6
Impact Level
Severity 1
Symptom
Description
Some CIFS users could not connect to Celerra in either of two situations:
- If the Data Mover needed to look up the CIFS users group or user SID in a trusted domain that had
xSID filtering (quarantine) enabled
- If the SID of a trusted domain was not returned and the connecting user had user or group
xmembership in ANY trusted domain
This problem only occurred when the Data Mover needed to resolve user SIDs, such as when the user
was connecting to Celerra for the first time, or if the Secmap SID caching feature was turned off.
Fix
These problems were introduced in version 5.5.29 by an enhancement that attempted to identify and
avoid lookups to unreachable trusted domains. This was intended to help improve speed of logins to
Celerra in environments with unreachable trusted domains (i.e. due to a temporary network outage).
The code has been reverted to the pre-5.5.29 behavior until these bugs are addressed.
Fixed in version
5.5.31.0
Impact Level
Severity 1
Symptom
Description
When user tried to access a DFS link in a DFS share located on an un-mounted file system, the Data
Mover panicked with the following DART panic/fault message: *** Watch Point
exception. Virt ADDRESS: 6a08d2 ***. The stack trace showed:
0xe820bcbc: 0x13d844 waitForReboot+0x90
0xe820bcdc: 0x13da41 fault_dump+0x67
0xe820bcfc: 0x137af0 handle_fatal_fault+0xa
0xe820bd60: 0x996aff
xxx_ZN13smb_threadCtx9parsePathEPKcbiRbR11File_St
0xe820bda0: 0x9b101c
xxx_ZN13smb_threadCtx14replyCreateNTXEP4msgb+0x2a
0xe820bf30: 0x914c3d _ZN13smb_threadCtx7processEP4msgbi+0x957
0xe820bfa0: 0x90abb9 _ZN13smb_threadCtx5startEv+0x633
0xe820bfe0: 0x140fd8 Sthread_startThread_internal+0xe
Fix
Corrected a problem parsing DFS paths that passed an invalid reference to file system handling code.
This panic no longer occurs.
Fixed in version
5.5.31.0
207
Version 5.5.31.6
208
Impact Level
Severity 1
Symptom
Description
When a compname account was manually created in AD, the join operation reported success (but it
failed internally). The compname could no longer be used with Kerberos authentication as the join
operation didn't complete.
Fix
Now, the Control Station displays command failed when the join fails. User can check the
server_log to get the failure reason.
Service Request
18274826
Fixed in version
5.5.31.0
Impact Level
Severity 2
Symptom
Description
During a fresh install of an NS20 with SATA disks on the DPE shelf and FC on add-on shelf, the control
LUNs were being created on the FC disks.
Fix
Code now creates the control LUNs on the first 5 drives on the first shelf of Integrated Celerras if the
first 5 drives are not already in use.
Fixed in version
5.5.31.0
Impact Level
Severity 1
Symptom
Description
After a Data Mover made a secure connection to a domain controller, the domain controller returned
ACCESS DENIED errors to NetrLogonSamLogon requests. This prevented NTLM users from
authenticating with the Celerra. The reason for the domain controller behavior is unknown, but
appears to happen if the domain controller has two or more network interfaces, and the Data Mover
has both a Compname and a NetBIOS name defined within the same Windows domain.
Fix
The Data Mover cannot prevent the domain controller problem from occurring, but now recovers from it
automatically by disconnecting and re-establishing a secure channel when this problem occurs.
Fixed in version
5.5.31.0
Version 5.5.31.6
Impact Level
Severity 2
Symptom
Description
Requesting the access right on a file system object or a share for a local user of a standalone server
failed with the following message:
server_cifssupport server_3 -accessright -name Administrator
-compname myStandaloneServer -path /mnt04/test build
server_3 :Error 13160939521: server_3 : The Windows user
Administrator must be defined by its name and its domain.
In this case the name of the standalone server, myStandaloneServer, was not sent to the Data Mover.
That caused the Data Mover to reject the access right request for the user Administrator against the
file system object /mnt4/test. When one or more CIFS server had joined the specified domain, the
same request for a user that was in a domain worked as designed.
Fix
The name of the standalone server that was missing is now sent to the Data Mover. The request can
then be successfully processed.
Fixed in version
5.5.31.0
Impact Level
Severity 1
Symptom
Description
Group Policies are collected and settings are evaluated in a precedence order (lowest to highest),
unless one of two things happen:
A Group Policy is set to "No override" (also known as "Enforced"). In this case, any settings defined
in this Group Policy cannot be changed by settings in higher precedence Group Policies.
An Organizational Unit has "Block Inheritance" set. In this case, all settings collected up to that
point are discarded, except those that are from Group Policies and are set to "No override".
An exception to the precedence order occurs if two or more Group Policies within a single
Organizational Unit are set to "No override". In this case, the precedence order of those Group Policies
determines which setting will be used (and set to "No override"), instead of the "first to be set to No
override" (which would favor the lower precedence Group Policy). For example, if there are two Group
Policies in an Organizational Unit, both set to "No override", and have the same setting defined, then
the setting from the higher precedence Group Policy should be used.
The Data Mover did not correctly determine the setting when two or more Group Policies within a
single Organizational Unit were set to "No override", and had the same settings defined. The Data
Mover incorrectly used the setting from the lowest precedence Group Policy, because it was the first to
be encountered with "No override" set.
Fix
This has been corrected by updating precedence processing within an OU, when no override is set.
Service Request
18958042
Fixed in version
5.5.31.0
209
Version 5.5.31.6
Control Station

Control Station:
Impact Level
Severity 1
Symptom
Description
During an upgrade or fresh install of an NS/NSX standby Control Station, the standby copies files from
/nbsnas (the NAS database file system) to /nas (the NAS cache). In this case, the standby determined
that /nbsnas file system needed to be repaired and attempted to run FSCK against the file system.
However, because /nbsnas was already mounted on the primary Control Station with read/write
access, the attempted repair damaged the file system instead. This resulted in the inability to start the
NAS service after an upgrade, NAS command failures, file system errors in the /var/log/messages file,
and file I/O errors during a subsequent upgrade.
Fix
The NAS upgrade process and Control Station Linux kernel were changed to never attempt repairs of
NAS file systems mounted for read-only access on standby Control Stations.
Fixed in version
5.5.31.0
Impact Level
Severity 2
Symptom
Description
Every time a user aborted a remote fs_copy session using the fs_copy -abort command, the
timestamp of the destination file system was updated. This caused the next fs_copy command to fail
with the following error: operation in progress (not
interruptible)...Error 2237: Execution failed: has_obj :
Precondition violated. [FILE_SYSTEM.unregister_backup]. This
problem was introduced in version 5.5.27.5.
Fix
In a remote copy case, when user manually aborts the copy session, the production file system and
timestamp info are no longer sent to remote side to cleanup, thus preventing the NASDB corruption.
Fixed in version
5.5.31.0
210
Impact Level
Severity 2
Symptom
Description
When executing setup_clarion -setup and a non-secure version of navicli was installed and the
backend was running FLARE version 26, although the RAID 6 templates were available and selectable,
the command failed without any error messages.
Fix
Added a check to verify CLARiiON security is ON prior to presenting RAID 6 templates. If it is OFF, the
following error is returned: WARNING: RAID 6 templates skipped because
security has not been enabled. Contact EMC Customer Service
and refer to PRIMUS emc168569.
Fixed in version
5.5.31.0
Version 5.5.31.6
:
Impact Level
Severity 1
Symptom
Description
When using a Secure Navicli command, if the command was issued and then the connection to the
storage processor went down for any reason, the command hung indefinitely.
Fix
This fix corrects the problem so Secure NaviCLI properly aborts after 1 minute if the communication to
the storage processor is lost.
Fixed in version
5.5.31.2
Impact Level
Severity 1
Symptom
Description
When user had more than one LUN 0 configured on the backend (which is an unsupported
configuration), the Data Mover failed to boot upon the reboot.
Fix
For the NS series systems, extended POST has been changed so that the system will be able to
choose and boot with the correct LUN 0 if it is misconfigured with multiple LUN0s on the backend.
Fixed in version
5.5.31.0
CORE

Data Movers:
Impact Level
Severity 1
Symptom
Description
File systems could not be unmounted after running the server_checkup command against a Data
Mover. This could happen to any file system, but would be more likely to occur with the root file system
of a Data Mover or VDM.
Fix
After running the server_checkup command, in-memory references were not being removed, causing
the file system to appear to be in use and preventing it from being unmounted. This reference is now
removed, and the problem does not reoccur.
Fixed in version
5.5.31.0
211
Version 5.5.31.6
212
Impact Level
Severity 1
Symptom
Description
User upgraded a Cisco Fibre Channel switch to a newer firmware version which had a longer login time
frame of 2+ seconds. This caused the Data Mover fabric login process to timeout.
Fix
Extended the fabric login timeout on the Data Mover by 2.5 seconds to allow for login into the Cisco
VSAN.
Service Request
18077322
Fixed in version
5.5.31.0
Impact Level
Severity 1
Symptom
Description
While an NFS client was deleting a file or a directory and a CIFS client was trying to access the same
object, the Data Mover panicked with the following DART panic/fault message: *** GP
exception. Virt ADDRESS: 527864. Err code: 0 ***. The stack trace
showed:
0x129ba98: 0x13b8cb waitForReboot+0x8b
0x129bab8: 0x13baca fault_dump+0x5e
0x129bad8: 0x135c74 handle_fatal_fault+0xa
0x129bb10: 0x58549f
xxx_ZN9VFS_Vnode21releaseInternalLockedEPv11File_S
0xd8f83de0: 0x585c23
xxx_ZN9VFS_Vnode15releaseInternalEPvi11File_Statu
0xd8f83e10: 0x527e31 _ZN15File_NamingNode7releaseEv+0x27
0xd8f83e30: 0x4dc17c _Z11nfsv3_WriteP4msgb+0x5462
0xd8f83f30: 0x4d10d8 _ZN17nfs_serverInitial5startEv+0x87c
Fix
The file's triggerset is part of the CIFS structure supporting the file. During the NFS delete, the
triggerset is released. The CIFS access continues to use the released triggerset and if that memory is
reassigned, as was the case here, a panic will occur. As of version 5.5.31.6, code no longer releases
the triggerset before the file or the directory is deleted, and the panic no longer occurs.
Service Request
18496178
Fixed in version
5.5.31.0
Version 5.5.31.6
Impact Level
Severity 2
Symptom
Description
The Data Mover showed low or 0% CPU idle, and in this case, NFS users saw slow performance. The
user's application heavily accessed one particular file. The accesses came through multiple clients
and multiple TCP connections on each single client. For each file, client, and TCP connection
combination, the Data Mover creates one entry in a list of open files, which is then hashed for fast
searching through the list. Because of the way that the list was being hashed, having a very large
number of opens of the same file caused high CPU impact when scanning the list of open files.
Fix
The hashing mechanism has been improved to handle this type of workload more efficiently, requiring
less CPU.
Fixed in version
5.5.31.0
FileMover

FileMover:
Impact Level
Severity 2
Symptom
Description
Attempting to delete a DHSM connection using the fs_dhsm command resulted in the error: Error
2237: Execution failed: valid_src: Precondition violated.
[CDMS_CONNECTION_ENTRY.string_extract]. Only certain connections were
affected. Fs_dhsm c <fs_name> -list output listed the affected
connection twice.
Fix
This problem occurred with DHSM connections that had been told to migrate all files back to primary
storage before deleting the connection. When a failure occurred during that migration, the "Error 2237"
was displayed. This problem has been corrected. Also, when detailed information is viewed for the
connection with the fs_dhsm command, the migration failure status is now displayed.
Fixed in version
5.5.31.0
Install/Config/
Upgrades

213
Version 5.5.31.6
214
Impact Level
Severity 2
Symptom
Description
During a NAS code upgrade from 5.5.21 or earlier to 5.5.21 or later, the Pre-Upgrade Health Check
(PUHC) displayed the following warning:
Storage System: Check if FLARE is supported
Symptom: Backend Storage Requirements Check Failed: invalid
requirements file -/tmp/be::check_if_FLARE_is_supported13539
Unable to determine if the Storage System software is
qualified with this version of NAS.
This warning did not impact the NAS code upgrade.
Fix
When the PUHC runs, it validates that a supported version of FLARE is running on the attached
CLARiiONs. This is called a requirements check.
If a new requirements check was added to a NAS release, PUHC also needed to be updated to
prevent the check from running on older releases that did not support the check. In some earlier
releases, PUHC was not updated, and this error occurred. PUHC now automatically detects
differences between the requirements checks in the new release and the old release, preventing this
error from occurring again.
Service Request
20155968
Fixed in version
5.5.31.0
Impact Level
Severity 2
Symptom
Description
When doing a NAS code installation, if the install process detected exactly 5 FC disks (which is a valid
configuration) the following message was displayed, which might confuse the installer:
Found 5 unused ATA disks of size <size>. ****
System RAID Group will be setup on disks 0_0_0 0_0_1 0_0_2
0_0_3 0_0_4
Do you wish to continue [yes or no]?:
This occurred whether drives were Fibre Channel or ATA. However, the installation continued normally.
Fix
This message is now displayed only if less than 5 FC disks are detected, which is an invalid
configuration.
Fixed in version
5.5.31.0
Version 5.5.31.6
Impact Level
Severity 2
Symptom
Description
The upgrade code did not check for sufficient free space on the Control Station root file system, and
the NAS code 5.5 upgrade failed due to lack of disk space.
Fix
The Pre-Upgrade HealthCheck (PUHC) was changed to make sure the Control Station root fs is less
than 71% full. If it is 71% or more full, the HealthCheck will prevent the upgrade from continuing.
Service Request
18274202
Fixed in version
5.5.31.0
Impact Level
Severity 1
Symptom
Description
Prior to 5.5.31.0, the system didnt check whether disk volumes were of supported RAID types prior to
marking them for use by Celerra. Additionally if disk volumes were of supported RAID type but were
somehow modified to an unsupported type using standard CLARiiON tools there could be unintended
side affects including failed upgrades with : Error 5008: server_2 :
clar_r5_performance : storage pool has members which do not
have consistent disktype.
Fix
Beginning with version 5.5.31.0, disk volumes built from unsupported RAID configurations or meta
LUNs will no longer be marked for use by Celerra. When detected during diskmark, a health check
warning will be generated but the disk will not be added to the Celerra database.
In addition heath checks have been added and will fail with an error for any existing disk volume which
meets the following criteria.
Is built from a Meta LUN
Is built on an unsupported RAID configuration AND
Is used in a system defined AVM pool
Is not in-use
The following error is displayed when the nas_storage -check -all command fails:
Error 5017: storage health check failed
APM00070204288 d13, metaLUN, doesn't match any storage
profile.
Other disk volumes built from unsupported RAID configurations will continue to function and will not
affect upgrades.
Service Request
17219002
Fixed in version
5.5.31.0
215
Version 5.5.31.6
216
Impact Level
Severity 1
Symptom
Description
NAS file systems could not be mounted to a secondary Control Station in read-only mode if they were
already mounted to the primary Control Station. This prevented the secondary Control Station from
being upgraded. The Pre-Upgrade Health Check failed with the following error:
Symptom: Command failed
* Command: mount /nbsnas -o ro
* Command output: NAS Partitions cannot be mounted on CS 1
(this CS)
when it is the not the current Primary CS
* Command exit code: 1
Fix
This problem was introduced in version 5.5.29 by a change to prevent NAS file systems from being
mounted in read/write mode, allowing NAS database corruption to occur. This change erroneously
blocked read-only mounts as well. NAS file systems can now be mounted read-only on a Control
Station even they are already mounted by the other Control Station.
Fixed in version
5.5.31.0
Impact Level
Severity 2
Symptom
Description
After upgrading to NAS version 5.5.27.5 or higher, user began seeing messages similar to the following
in the server_log: LOCK: 4: Statd_MonitorHost::suppr: Client <nfs
client name>, addr <ip address> isn't monitored. In an environment with
NFS clients that unlock the same lock twice, the error flooded the server_log.
Fix
The message was added to help debug situations where NFS clients unlocked files improperly. If a
client attempted to unlock a file that it did not have a lock for, this message was logged. However, there
are many NFS clients that decide to unlock a file twice, which triggers this message to be logged.This
behavior doesn't cause a problem, and the error message is informational only. We moved the error
message from a warning severity error to an information level error, and it will no longer be logged by
default.
Fixed in version
5.5.31.0
Version 5.5.31.6
Impact Level
Severity 2
Symptom
Description
A Data Mover's root file system had been extended with a slice of LUN 0 - from the root_disk_reserve
area. This is the disk area used by the Data Mover to save panic dumps. When the Data Mover
panicked, it overwrote data from the Data Mover root file system when writing the dump, making the
root_fs unusable. Since NAS version 5.4, this area is marked as in use so it cannot be reused for file
systems. NAS code upgrades to NAS 5.4 or NAS 5.5 already check if this area is in used and report an
error - but the message is printed in the middle of the upgrade log and may not be seen by personnel
performing the upgrade.
Fix
The PUHC/PAHC identify this condition and prevent an upgrade from continuing. Customer Support
needs to be involved to resolve the situation before the upgrade can be completed.
Service Request
18030094
Fixed in version
5.5.31.0
iSCSI

Impact Level
Severity 2
Symptom
Description
While restoring a destination snap to the source LUN, user modified the destination LUN's attributes to
read-write. However, since the restore expects the destination LUN to be always read-only, iSCSI
replication failed with the following error: cfgModifyLun failed. Replication task
is running on LUN #, protected attr cannot be modified. Error
4020: server_# : failed to complete command.
Fix
Code was improved to always check if any replication task is running on the LUN before modifying its
attributes.
Fixed in version
5.5.31.0
NDMP/Backup

217
Version 5.5.31.6
218
Impact Level
Severity 2
Symptom
Description
While performing an NDMP/PAX backup of a file system with many hardlinks or database files, the
Data Mover panicked with the following DART panic/fault message: >>PANIC in file:
../addrspac.cxx at line: 275 : Out of memory. The stack trace showed:
0xa0255cc0: 0x13d1db waitForReboot+0x8b
0xa0255ce0: 0x13d3e3 fault_dump+0x67
0xa0255d00: 0x13d2e3 PANIC+0x2b
0xa0255d10: 0x13918c
0xa0255d30: 0x187876 _Z11more_memoryiP9free_list+0x4e
0xa0255d50: 0x187923 malloc+0x89
0xa0255d80: 0x451287
xxx_Z12kg_unseal_v1P13_krb5_contextPjP20_krb5_gss
0xa0255e40: 0x4516ca
xxx_Z9kg_unsealP13_krb5_contextPjPvP22gss_buffer_
Fix
A hardlink table was not deleted properly when there was an error/abort during backup. Fixed the
hardlink table memory leak in NDMP. The hardlink table is now deleted properly when there is an
error/abort during backup.
Service Request
19398930
Fixed in version
5.5.31.0
Impact Level
Severity 1
Symptom
Description
During the file creation stage of VBB, two threads attempted to create the same file simultaneously.
This caused the VBB data retrieval stage to fail because it detected that the same set of blocks were
being used twice which caused the Data Mover to panic with the following DART panic/fault message:
Assertion failure: 'currFileData.dblk.startBlk >=
tapeRecStart' on line 565 in file '../vbbRestoreFileData.cxx'
>>PANIC in file: ../assert.c at line: 26 : assert failed
Fix
During the file creation stage of VBB, the code now ensures (prior to creating a file) that it doesn't
already exist. If it exists, a link to the file is created instead.
Fixed in version
5.5.31.0
Version 5.5.31.6
Networking
The following entries describe fixed issues related to Networking:
Impact Level
Severity 2
Symptom
Description
When closing an httpd connection, there is a condition which does not clean up all the memory
allocated to this connection. Over time this caused a memory leak which led to a Data Mover panic
with the following DART panic/fault message:
>>PANIC in file:
../addrspac.cxx at line: 275 :
Out of memory. The stack trace
showed:
0x13d60aa0: 0x145660 waitForReboot+0x90
0x13d60ac0: 0x145873 fault_dump+0x67
0x13d60ae0: 0x14575d PANIC+0x29
0x13d60af0: 0x13d7a6 _ZN19Memory_AddressSpace10allocPages
xxxEi12Memory
0x13d60b10: 0x1af562 _Z11more_memoryiP9free_list+0xde
0x13d60b40: 0x1af787 malloc+0x16f
0x13d60b80: 0xdf59d8
0x13d60ba0: 0xdf4eed
Fix
When an httpd connection is closed, code now ensures that all memory allocated for this process is
also freed.
Fixed in version
5.5.31.0
Impact Level
Severity 1
Symptom
Description
Users saw slow performance when reading data from the Data Mover over a network with a low
congestion point. The network was frequently overwhelmed and dropped data, causing performance
loss.
Fix
Corrected a condition where the Data Mover did not shrink its TCP congestion window at all after
network congestion and packet drops occurred. This caused network congestion to occur more
frequently, which in turn resulted in packet drops and slower performance.
Service Request
18101906
Fixed in version
5.5.31.0
219
Version 5.5.31.6
Impact Level
Severity 2
Symptom
Description
In an environment with a large number of packet drops, user saw slow network throughput when
reading data from a Data Mover.
Fix
Corrected a condition where the Data Mover did not perform fast SACK/FACK recovery, and instead
performed a slow TCP timeout. This caused recovery from network packet loss to occur slower than
necessary. Also corrected a problem that caused the TCP congestion window to shrink
unnecessarily, which slowed throughput.
Service Request
18101906
Fixed in version
5.5.31.0
NFS
220
Impact Level
Severity 1
Symptom
Description
When mounting multiple file systems from a large number of clients (close to 1000 in this case)
simultaneously, some mounts either failed or hung.
Fix
Along with other changes in previous releases to increase the number of mount threads from 3 to 64,
code now prevents too many TCP connections from being opened which caused the hang condition.
Service Request
19123664
Fixed in version
5.5.31.0
Impact Level
Severity 1
Symptom
Description
An NFS deadlock condition occurred when using NFSv4. The Data Mover became unresponsive if an
NFSv4 client became unreachable after it wrote some data in the server cache (for example, option
UNSTABLE4 is set with the WRITE operation, or heavy write load on the Data Mover).
Fix
Corrected a deadlock that occurred when the server cleaned up the resources held by an expired
NFSv4 client.
Fixed in version
5.5.31.0
Version 5.5.31.6
Impact Level
Severity 1
Symptom
Description
1477 : bad SD storage size. The stack trace showed:
0xc2013820: 0x145678 waitForReboot+0x90
0xc2013840: 0x14588b fault_dump+0x67
0xc2013860: 0x145775 PANIC+0x29
0xc2013870: 0x677df7
xxx_Z16nfs4_setModeBitsP15File_NamingNodetP17Acce
0xc2013a30: 0x672bf9
xxx_ZN16nfsv4_attributes7SetAttrEP8nfs4_ctxP15Fil
0xc2013ae0: 0x646b2c
xxx_Z12nfs4_setAttrRP4msgbS1_RPhR8nfs4_ctx+0x302
0xc2013d20: 0x63ec9b _Z23nfs4_convertNfsv3Statusm+0x8e5
0xc2013e90: 0x61aa94 _ZN17nfs_serverInitial5startEv+0x13d8
Fix
Fixed a wrong computation that was causing this panic when building ACL for synchronization.
Fixed in version
5.5.31.0
Impact Level
Severity 1
Symptom
Description
NFS clients saw hangs and RPC timeouts when attempting to access the Data Mover. No error
messages were logged in the Data Mover logs, but a Data Mover failover resolved the issue.
Investigation showed that the problem was triggered by an LDAP server that was unresponsive. This
problem only happened when the LDAP server did not acknowledge receiving data at a TCP level.
Fix
As LDAP requests were generated to be sent to the LDAP server, these requests are buffered. They
cannot be removed from the buffer until the LDAP server acknowledges receiving them. However, if the
buffer filled and could not receive more LDAP requests, LDAP did not return an error, instead the
request hung. LDAP now returns an error to the calling module, which allows it to take the appropriate
action. The action taken and message logged differs depending on the calling module (NFS, CIFS
GPO, CIFS domain join, etc).
Service Request
17823176
Fixed in version
5.5.31.0
221
Version 5.5.31.6
Impact Level
Severity 1
Symptom
Description
After setting NFSv4 ACEs with inheritance using the Solaris 10 chmod command, ACEs were not
inherited. Using the setfacl to set ACLs (POSIX ACLs) inherited as expected.
Fix
This happened because the chmod command set both NFSv4 ACE information and Unix MODE bits.
The ACL was created based on MODE information. Files and directories now retain inherited ACEs
when they are created, and Unix MODE information is used to create ACEs for User, Group, and
Others unless defined otherwise by inherited ACEs.
Service Request
19234122
Fixed in version
5.5.31.0
Replication
222

Impact Level
Severity 1
Symptom
Description
IP Replication delta sets could not be transferred to the target side due to a network issue. As writes
continued on the PFS, the Replication SavVol on the source became full and the block numbers of
changed blocks continued to be tracked. When the network issue was resolved, delta sets were again
able to be transferred to the target which freed space in the source SavVol. At this point a delta set
needed to be created which contains all the changed blocks. However, because of the large number
of changed blocks, the operation to create a delta set caused a Data Mover panic with the following
DART panic/fault message:>>PANIC: I/O not progressing LastVol touched
LF197 Kind 6 (ptr=0xd18e4c04). The stack trace showed:
0xf6aade90: 0x13d1db waitForReboot+0x8b
0xf6aadeb0: 0x13d3e3 fault_dump+0x67
0xf6aaded0: 0x13d2e3 PANIC+0x2b
0xf6aadee0: 0x703afd _ZN9RawIO_IRP13checkProgressEv+0x107
0xf6aadfd0: 0x7039ec _ZN17IrpProgressThread5startEv+0x6
0xf6aadfe0: 0x1406f4 Sthread_startThread_internal+0xe
Fix
The amount of changes to the PFS that can be tracked while source flow control is enabled has been
increased to 8GB. When this limit is reached the replication session will become inactive. When this
occurs, a message is logged to the server_log and sys_log, and the WebUI shows an alert.
Service Request
17684968
Fixed in version
5.5.31.0
Version 5.5.31.6
Impact Level
Severity 1
Symptom
Description
A race condition occurred while replication sessions were recovered in asynchronous mode and file
system HighCapacity mode for IP replication (FS mounts > 2TB allowed) was set to FALSE. The replica
recover command was invoked, followed by the file system mounts. Mount for file systems over 2TB
failed if the mount command finished too quickly before replication recover read the value for
HighCapacity mode from disk. Failure in mount of FS resulted in failover failure.
Fix
The fix changes the value of HighCapacity mode from FALSE to TRUE and prevents the race condition
from occurring.
Service Request
19241558
Fixed in version
5.5.31.0
SRDF

Impact Level
Severity 1
Symptom
Description
The nas_rdf restore command failed with: CRITICAL FAULT: Unable to ping
source site control station at 1.
This caused a NASdb lock and slot_recover process to hang, which cleared after the Control Station
was rebooted. If these were not cleared, other NAS processes may have been affected, including
command failures and Data Mover failover problems.
Fix
This problem was introduced in 5.5.27.5 as part of improvements to prevent two Data Movers from
being configured the same way (same file systems, same IP addresses) at the same time.
In this case a log maintained during failover was being written when the failover process was started,
but not failed after RDF failover completed. This problem has been corrected.
Service Request
18910294
Fixed in version
5.5.31.0
223
Version 5.5.31.6
Impact Level
Severity 2
Symptom
Description
While running the nas_rdf activate command, the following messages were displayed:
An RDF 'Update R1' operation execution is in progress for
device
'DEV001' in group '1R2_500_2'. Please wait...
Unable to acquire the Symmetrix device lock
The Unable to acquire lock message is informational only, and will always happen when the R1
system is shut down during RDF activation (failover). It does not cause impact, but caused users to
believe that RDF failover had encountered errors.
Fix
This informational message is no longer displayed to the user when performing RDF failover. It
continues to be logged to the RDF log for debugging purposes.
Service Request
15163466
Fixed in version
5.5.31.0
Storage
224
Impact Level
Severity 1
Symptom
Description
After receiving an error or incorrectly formatted Fibre Channel frame from the storage array, the Data
Mover panicked with the following DART panic/fault message: >>PANIC in
file: ../dart/dart_debug_util.cxx at line: 29 :
fc_osw_dart_panic. The stack trace showed:
0x103ef2c: 0x13d600 waitForReboot+0x90
0x103ef4c: 0x13d7fd fault_dump+0x67
0x103ef6c: 0x13d6fd PANIC+0x29
0x103ef7c: 0x18872d _assert+0x51
0x103ef9c: 0x787450
xxx_ZN14ScsiBus_Device13sim_finishHBAEP14Scsi_IORe
0x103efcc: 0x787223
xxx_ZN14ScsiBus_Device10sim_finishEP14Scsi_IOReque
0x103efec: 0x79e717 _Z14FCP_completeIOPv+0x3c9
0x103f01c: 0x7f24a2 cpd_osw_finish_srb+0xd2
Fix
A number of different error situations could allow the Data Mover Fibre Channel state machine to enter
a state from which it could not recover. When encountering any of these situations, the Data Mover
now performs an HBA reset, which allows it to recover.
Fixed in version
5.5.31.0
Version 5.5.31.6
Impact Level
Severity 2
Symptom
Description
The nas_diskmark command failed to find all available paths to RAID 6 storage on the first execution.
Fix
Code now recognizes RAID 6 specific info in inquiry data.
Fixed in version
5.5.31.0
Tools
The following entries describe fixed problems pertaining to tools

such as EMCOPY, EMCACL, LGDUP, SHAREDUP:
Impact Level
Severity 2
Symptom Description
Starting Virus Checker from MMC did not save entry in the Celerra database, therefore the virus
checker service was not started after a Data Mover reboot.
Fix
Virus Checker now starts/stops simultaneously with the start/stop of CIFS.
Fixed in Version
5.5.31.0
UFS
Impact Level
Severity 2
Symptom
Description
When attempting to fix a corrupted ACL header, FSCK detected the corruption and reported it as fixed,
however it did not fix the corruption. Running FSCK an additional time found the error again.
: UFS: 3: Aclchk fsid 465: Incorrect next ptr slot 20: 20,
actual -30
: UFS: 6: Aclchk fsid 465: Merging free chunk slot 52 with
chunk slot 51, next free= 0, new size= 0xffffffff
: FSTOOLS: 6: Aclchk fsid 465 - phase2: 100% complete
Fix
This ACL header corruption is now repaired properly by FSCK.
Fixed in version
5.5.31.0
225
Version 5.5.31.6
226
Impact Level
Severity 1
Symptom
Description
If the server was rebooted via the server_cpu -reboot command during FileMover or CDMS migration, a
race condition led to a Data Mover panic with the following DART panic/fault message: ***
Divide exception. Virt ADDRESS: de3e5e ***. The stack trace showed:
0x170670c: 0x13d810 waitForReboot+0x90
0x170672c: 0x13da0d fault_dump+0x67
0x170674c: 0x137abc handle_fatal_fault+0xa
0x17067a0: 0xde3d99 __udivdi3+0x29
0xeb693a00: 0x621a3d _ZN9UFS_Vnode4bmapElRlPiS1_+0x233
0xeb693a60: 0x62514d
xxx_ZN9UFS_Vnode5breadEjR11File_Statusi+0xc1
0xeb693ab0: 0x625267
xxx_ZN9UFS_Vnode9blockReadEjR11File_Status+0x4d
0xeb693af0: 0xb970f4
xxx_ZN10MGFS_Vnode9blockReadEjR11File_Status+0x48
Fix
The race condition was fixed - thus preventing the memory corruption and panics.
Fixed in version
5.5.31.0
Impact Level
Severity 1
Symptom
Description
The Data Mover panicked with the following: PANIC in file: ../ufsvfsop.cxx at
line: 374 : sync: no progress reducing dirty list.
Fix
While attempting to flush journaling information to disk, if a NULL pointer was found in the list of
metadata to be flushed to disk, and the Data Mover stopped writing journaling changes to disk. This
resulted in the Data Mover panic. All data is now flushed properly, correcting this problem.
Fixed in version
5.5.31.0
Version 5.5.30.4
Version 5.5.30.4
Celerra
Management
Applications

Impact Level
Severity 2
Symptom
Description
All UPSMonitor events, including those at severity level INFO (6), were posted as alerts on the Home
Status page of Celerra Manager. Also, a more severe posting from UPSMonitor did not update the
overall system status so the status showed as 'OK' when there were UPSMonitor warnings or errors.
Fix
Added severity filter to alert configuration file for UPSMonitor events. Only events with severity level
Warning or above will be posted as alerts on the Home Status page of Celerra Manager. Also updated
alert status when UPSMonitor alerts occur to ensure severity is reflected in overall system status.
Fixed in version
5.5.30.0
Impact Level
Severity 3
Symptom
Description
When using special characters in the community name when specifying an SNMP trap address for any
of the four types of notifications, an "Invalid destination" or other invalid entry pop-up was displayed.
Fix
Changed the input validation to allow special characters ~ ! @ $ % ^ * + = { } : ? _ # Characters A-Z,

a-z, 0-9, . and - were already allowed. (Note: These are shown with a spaces for readability, but no
spaces are allowed.)
The hostname/IP restrictions didn't change, hostnames can only be A-Z, a-z, 0-9, - (dash or minus
sign), and . (period) as a delimiter in a domain-style name. The following special characters are still
not accepted due to restrictions with Linux and with the notification implementation: & ( ) | [ ] \ " ' ; < >
, / space
Service Request
17677790
Fixed in version
5.5.30.0
227
Version 5.5.30.4
228
Impact Level
Severity 2
Symptom
Description
In the Mount list screen of Celerra Manager, the mounts which were not accessible to the user logged
in were displayed. This is just a read only view and the user cant access the details of those mounts.
When the user tried to visit the properties page of those mounts, the properties screen was blank.
Fix
The fix/new behavior addresses this failover gracefully. i.e. the mount properties page displays a
warning message that the mount properties are not available. Also, the page is disabled and only the
back button is available for navigation.
Fixed in version
5.5.30.0
Impact Level
Severity 2
Symptom
Description
The schedule name's value was set to schedule ID property. This triggered an invalid action response
from Celerra Manager when listing task files. Users were unable to access task functionality through
Celerra Manager. The following error was displayed: "Invalid action response from
Celerra Manager agent for task Query tasks All..."
Fix
This issue relates to renaming of a schedule. The Checkpoint creation task which gets fired after
renaming the schedule had a wrong value for 'schedule=' field. This field was changed so that the
schedule name property now contains the correct scheduleName.
Service Request
18137730
Fixed in version
5.5.30.0
Impact Level
Severity 2
Symptom
Description
After a NAS code upgrade from 5.4.22-2 to 5.5.25.2., user was unable to monitor any tasks. The
following error was displayed: Invalid action : response from celerra
manager agent for task query task all.
Fix
Added a check in the upgrade_54schedules.sh script to ensure that schedule file names match the
"Name" parameter of the /nas/tasks/schedule file before upgrading the schedule.
Service Request
17744032
Fixed in version
5.5.30.0
Version 5.5.30.4
CIFS

problems:
Impact Level
Severity 1
Symptom
Description
A system watchdog panic occurred when more than one user quota was being set at the same time
and quotas was turned OFF. DART panic/fault message: >>PANIC in file:
../sched.cxx at line: 802 : SYSTEM WATCHDOG. The stack trace showed:
0x3625f34: 0x13d047 waitForReboot+0x8b
0x3625f54: 0x13d24f fault_dump+0x67
0x3625f74: 0x13d14f PANIC+0x2b
0x3625f84: 0x13eb5a system_watchdog+0x7c
0x3625f94: 0x1370e5 _t_nmi+0x63
Fix
The code has been modified to serialize quota updates and prevent the race condition that caused this
panic.
Service Request
15991252
Fixed in version
5.5.30.0
:
Impact Level
Severity 1
Symptom
Description
In a mixed environment, user needed to control CIFS user access to files/directories using only UNIX
permissions on specific file systems. The cifs.acl.checkacl parameter applies to all file
systems mounted on a Data Mover and cannot be limited to individual file systems.
Fix
A new parameter, cifs.acl.unixCheckACL, has been added. This parameter enables or

disables CIFS ACL checking for any file system that is mounted with the UNIX access policy on the
Data Mover. It has no affect on file systems mounted with other access policies.
By default, the param is set to 1 - this means that CIFS ACL checking is ENABLED.
Setting the param to 0 will DISABLE CIFS ACL checking on any file system mounted with the
ACCESSPOLICY=UNIX.
The result of disabling the param is that all CIFS access rights are controlled by the UNIX mode bits.
Fixed in version
5.5.30.0
229
Version 5.5.30.4
230
Impact Level
Severity 1
Symptom
Description
Distributed File System (DFS) referrals from Celerra use NetBIOS name syntax. Depending on the
DNS configuration on the client, it may have been unable to resolve server names returned by the DFS
get referral query.
Fix
A value has been added to the Celerra CIFS server registry that can be changed to make the referrals
contain the Fully Qualified Domain Name (FQDN) rather than NetBIOS.
HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Services\Dfs
Value Name: DfsDnsConfig
Data Type: REG_DWORD
Value Data: 0 or 1
If you set the data value to 1, all roots added to the DFS tree use an FQDN. A value of 0 (the default)
specifies that NetBIOS names are used.
Refer to the Managing Celerra for the Windows Environment technical module for information on the
CIFS server registry.
Service Request
18157720
Fixed in version
5.5.30.0
Impact Level
Severity 1
Symptom
Description
CIFS clients could not connect to Celerra when the following security policies where set:
NTLMv2 is not allowed (LMcompatibility level is 0 or 2)
128-bit encryption is required
Kerberos is not allowed (or the client is unable to use it)
This caused clients to attempt to use NTLMv1 or NTLMSSP with 128-bit encryption, which the Celerra
indicated it did not support.
Fix
The Data Mover indicated that it did not support 128-bit encryption unless NTLMv2 support was
allowed, even though it was supported. The Data Mover now properly indicates that it supports
128-bit encryption at all times.
Service Request
18218122
Fixed in version
5.5.30.0
Version 5.5.30.4
:
Impact Level
Severity 2
Symptom
Description
When the Data Mover receives a CIFS request to open a file with the disposition FILE_SUPERCEDE,
if the file did not exists, the Data Mover did not create it. An open request with the FILE_SUPERCEDE
disposition should create a non-existent file.
Fix
Fixed code to correctly process the FILE_SUPERCEDE disposition option.
Service Request
18555796
Fixed in version
5.5.30.0
:
Impact Level
Severity 1
Symptom
Description
When the command server_cifssupport <datamover> -cred uid <uid>

-build was issued shortly after having restarted CIFS with the command server_setup, the
Data Mover panicked with the following DART panic/fault message: *** Page Fault
Interrupt. Virt ADDRESS: d706be Err code: 0 Target addr: 1 **
0x228ef58: 0x146950 waitForReboot+0x90
0x228ef78: 0x146b63 fault_dump+0x67
0x228ef98: 0x13c9a0 handle_fatal_fault+0xa
0x228efd0: 0xde779a
xxx_ZN12xmlAuditUser24buildNtCredFromWindowsIdE11l
0xa4ff9aa0: 0xdeaf38
xxx_ZN12xmlAuditUser7getUserEP14XmlElementNodebPK
0xa4ff9b20: 0xdeb3e2
0xa4ff9bc0: 0xdec9fa
0xa4ff9d40: 0x2908d4
xxx_ZN17macCCMDCommand_DB6invokeEPKcP7macDataRP8C
Fix
The panic occurred because there is a window of time where the internal CIFS structures are not fully
initialized. The Data Mover no longer panics under this condition, but the command fails when issued
too early after starting CIFS. The error returned if the command fails is "The NT credential
for the user <name> uid <uid> cannot be built".
Fixed in version
5.5.30.0
231
Version 5.5.30.4
Impact Level
Severity 2
Symptom
Description
MAC OS X client was unable to see the correct file system size when it exceeded 4TB. This problem
occurred when the Quotas were set on the file system and the policy was based on FileSize.
Fix
This problem only occurred with CIFS clients that use an older SMB_INFO_ALLOCATION call. In
this case, fields which are limited to 32-bit were being overflowed. The Data Mover now utilizes fields
so that they are not overflowed, and can report file system sizes to these clients of up to 32TB per
filesystem. Clients using newer CIFS calls do not have this limitation.
Service Request
18662998
Fixed in version
5.5.30.0
Control Station
232

Control Station:
Impact Level
Severity 1
Symptom
Description
A Data Mover failover was initiated from the command line and then an attempt was made to abort it
(using Ctrl-C) before it finished. This resulted in the particular step executing at the time to be aborted
but not the entire failover process. The incomplete failover caused an inconsistency within the
checkpoint SavVol of one file system leading to a panic when mounting the checkpoints. DART
panic/fault message:
>>PANIC in file: ../PagedBlockMap.cxx at line: 2061 :
duplicate blockmap. The stack trace showed:
0x364fce0: 0x13d1db waitForReboot+0x8b;
0x364fd00: 0x13d3e3 fault_dump+0x67;
0x364fd20: 0x13d2e3 PANIC+0x2b;
xxx0x364fd30:
0x7644f4_ZN9BtreeSnap11insertEntryEP10btreecmn_tjjP15Bt;
0x364fd60: 0x763170 _ZN15Btree_Insert_CB4doneEv+0x624;
0x364fdc0: 0x762b07
xxx_ZN9BtreeSnap8insertCBEP15Btree_Insert_CB+0x45.
Fix
The Data Mover failover process now ensures that step to build a minimum configuration on the primary
(so it becomes a standby) is complete before continuing. If this step fails the failover process will fail
leaving both servers in the same role (primary or standby) as when the failover was initiated.
Service Request
18246360
Fixed in version
5.5.30.0
Version 5.5.30.4
Impact Level
Severity 2
Symptom
Description
There were two interfaces specified for CIFS running on a VDM. The server_ifconfig
command was used to set the state of one interface to "DOWN". CIFS on that VDM failed to start
again after a reboot of the server on which the VDM was loaded.
Fix
The fix prevents server_ifconfig from setting an interface to "DOWN" when it is being used
by CIFS.
Service Request
15958942
Fixed in version
5.5.30.0
Impact Level
Severity 2
Symptom
Description
In a mixed array configuration with Symmetrix as the primary and CLARiiON as the secondary, the
nas_storage -modify -security command failed to change the cached credentials for
the secondary CLARiiON array.
Fix
Corrected problem where storage system ID was only being checked to see if the primary was a
CLARiiON. Code now checks to see if the primary is a Symmetrix also.
Fixed in version
5.5.30.0
Impact Level
Severity 2
Symptom
Description
The nas_server -delete command failed with the following: Execution failed:
valid_address: Precondition violated. This occurred when a user tried to delete
a VDM whose root file system had previously been copied using fs_copy to another file system.
Although the command appeared to fail, the VDM was deleted successfully and there were no side
effects.
Fix
Disabled the DB table refresh during the middle of commit, to avoid reloading a stale table from disk
while the committed data has already been written to new files.
Fixed in version
5.5.30.0
233
Version 5.5.30.4
234
Impact Level
Severity 2
Symptom
Description
The nas_storage command failed with Error 2224: invalid operation if the
boot array was a Symmetrix.
Fix
The nas_storage command has been corrected to properly handle creating a secondary
CLARiiON and initializing security when the boot array is a Symmetrix.
Fixed in version
5.5.30.1
Impact Level
Severity 2
Symptom
Description
CIFS shares that contained a leading dot in the share name ( /.<mount>) were not visible to
CLI/WebUI.
Fix
Code no longer filters shares with leading dots.
Fixed in version
5.5.30.0
Impact Level
Severity 1
Symptom
Description
After the server_http command was run by a user as root, and the command crashed, users
were no longer able to run server_http as other users, including nasadmin. When the
server_http command was issued by non-root users, it failed with the following: Error
2237: server_3 : Execution failed: Permission denied: Routine
failure. [FILE.file_open].
Fix
When running the server_http command by the root user, temporary files were created in the
/var/tmp directory that were not cleaned up when the command crashed. These files had permissions
of the root user and could not be accessed by other users. This prevented the server_http
command from running as the command attempted to access these files. The server_http
command no longer uses temp files which prevents this problem from happening.
Fixed in version
5.5.30.0
Version 5.5.30.4
:
Impact Level
Severity 2
Symptom
Description
NAS database history grew to very large sizes, unnecessarily, which filled the /nas file system on the
Control Station. This caused Celerra Manager/Monitor statistics gathering to stop when the file
system became more than 67% full, and eventually caused the file system to be 100% full.
Fix
The Control Station maintains a database (SCCS) which logs the history of each change to the NAS
database. When the nasdb_backup process runs (by default, once per hour), it causes several files
(up to 10 per Data Mover) to be regenerated. The contents of these files dont change, but since the
files are rewritten, their timestamps change.
Since the timestamps of the files changed, SCCS then logged an entry showing that each file changed
(but that zero lines changed). This caused SCCS databases to grow significantly faster than
necessary, consuming space on the /nas file system. Code was modified so that SCCS no longer
saves database updates when the *contents* of these files do not change.
Fixed in version
5.5.30.0
FileMover

FileMover:
Impact Level
Severity 1
Symptom
Description
While using the fs_dhsm command to make a FileMover connection, if the address of the
secondary (NFS server) was omitted, the Data Mover panicked with the following DART panic/fault
message: Assertion failure: 'addr != 0' on line 231 in file
'../client.cxx' >>PANIC in file: ../assert.c at line: 23 :
assert failed. The stack trace showed:
0x176af680: 0x143998 waitForReboot+0x90
0x176af6a0: 0x143b98 fault_dump+0x5e
0x176af6c0: 0x143aa1 PANIC+0x2b
0x176af6d0: 0x1a1497 _assert+0x53
0x176af6f0: 0x40aba5 _ZN6CLIENTC2Emmmmmll+0x1bb
0x176af750: 0xdcb018 _Z18kernel_clnt_createPKcmmS0_mt+0x17c
0x176af7a0: 0xdc273a
xxx_ZN21MGFS_NFSv3_Connection13connectLockedER14R
0x176afa00: 0xdc1fc2
xxx_ZN21MGFS_NFSv3_Connection7connectER14RemoteDi
Fix
Added more validation to the NFS server specification to ensure that a 0x0 address is not passed to
the RPC client library.
Fixed in version
5.5.30.0
235
Version 5.5.30.4
236
Impact Level
Severity 2
Symptom
Description
When deleting DHSM connections, the server log filled with messages like "LIB: 3:
freemalloc: try to free a pointer not stored". While the message is
harmless (memory is freed), the server log wrapped or rolled over, potentially losing important
information.
Fix
Fixed DHSM to call correct free method to ensure this message is no longer printed in the server log.
Service Request
17630030
Fixed in version
5.5.30.0
Impact Level
Severity 1
Symptom
Description
The Data Mover panicked after receiving an HTTP request from a DHSM policy engine which does not
respect HTTP protocol. In this case, the content-length header value contained a comma which is a
non-digit character. DART panic/fault message: *** GP exception. Virt ADDRESS:
d50941. Err code: 0 ***. The stack trace showed:
0x16f4e38: 0x13d600 waitForReboot+0x90
0x16f4e58: 0x13d7fd fault_dump+0x67
0x16f4e78: 0x1378b0 handle_fatal_fault+0xa
0x16f4ec0: 0xd509aa XML_GetFeatureList+0xf49a
0xf56a9ab0: 0xd50b81 XML_GetFeatureList+0xf671
0xf56a9af0: 0xd517be XmlParseXmlDecl+0x53
0xf56a9b30: 0xd433e8 XML_GetFeatureList+0x1ed8
0xf56a9b90: 0xd43ba2 XML_GetFeatureList+0x2692
Fix
We added code to avoid the panic. We will now close the TCP connection and return
HTTPSTATUS_BAD_REQUEST when a bad content-length value is specified in HTTP header.
Service Request
18683934
Fixed in version
5.5.30.0
Version 5.5.30.4
Impact Level
Severity 2
Symptom
Description
Non-administrator user was unable to modify the username/password of an existing DHSM connection.
Fix
Administrator privileges are no longer required to modify a DHSM connection.
Service Request
17773234
Fixed in version
5.5.30.0
Impact Level
Severity 2
Symptom
Description
Attempts to delete a FileMover connection failed. This occurred if the file system contained zero byte
files that had been migrated offline. The server log contained messages similar to the following:
Command failed: mgfs action=verify fsid=19 cid=0.
Fix
Zero byte files were not being recalled which then prevented FileMover deletion from completing.
Code now allows zero byte files to be recalled.
Service Request
18604250
Fixed in version
5.5.30.0
Install/Config/
Upgrades

Impact Level
Severity 2
Symptom
Description
On an integrated NS system, disk 0_0_13 failed but the Celerra did not generate a dial home for the
failure.
Fix
In this case, Naviagent had previously failed. Naviagent is now controlled by nas_mcd so that if it
terminates abnormally, it will be restarted. If Naviagent is running, we will detect backend disk errors.
This will make dial homes for issues from the backend much more reliable.
Service Request
17084706
Fixed in version
5.5.30.0
237
Version 5.5.30.4
Impact Level
Severity 2
Symptom
Description
During the installation of a dual Control Station's NS80, the installer did not prompt for whether this
was a dual Control Station install. This prevented the second Control Station from being installed.
Fix
The prompt for the dual Control Station was added to NS80 installations.
Fixed in version
5.5.30.0
Impact Level
Severity 2
Symptom
Description
If user had a customized /etc/motd (message-of-the-day) file, during an upgrade to 5.5.28.0, the file
was overwritten with the standard /etc/motd file.
Fix
Code no longer overwrites /etc/motd during an upgrade.
Service Request
18607542
Fixed in version
5.5.30.0
iSCSI
238

Impact Level
Severity 2
Symptom
Description
If an iSCSI snap creation using Replication Manager (RM/SE) failed, an orphan snap was left on the file
system where the iSCSI LUN resided, using up file system space and potentially preventing new snaps
because of space limitations.
Fix
Since iSCSI snap delete and create are serialized in UFS layer, and snap delete may take a long time
for a large file or a snap with large delta, a "wait" was added when creating a snap. RM/SE now fails the
snap create operation when there is async snap delete pending from the Data Mover (to ensure there
is no orphan snap left on the file system).
Service Request
17025672
Fixed in version
5.5.30.0
Version 5.5.30.4
Impact Level
Severity 1
Symptom
Description
When two Data Movers were involved in replication and one was upgraded to release 5.5.27.5 or
above and the other was still at a version below 5.5.27.5, a misleading message about the destination
LUN being out of disk space was displayed. This caused the user to take incorrect actions to resolve
an issue that didn't exist.
Fix
A mismatch between the data structures of the different releases caused this problem. This issue has
been resolved by reordering this data structure.
Fixed in version
5.5.30.0
Impact Level
Severity 1
Symptom
Description
In a rare misconfiguration situation where two hosts had the same iSCSI initiator node name, both
hosts entered an infinite loop of logins, and were interrupted by the other host. The server log was filled
with messages "Session reinstatement: new sess=". The Data Mover panicked due
to a memory leak that occurred during each login with the following DART panic/fault message:
>>PANIC in file: ../addrspac.cxx at line: 275 : Out of memory.
0x361fe00: 0x13d05b waitForReboot+0x8b
0x361fe20: 0x13d263 fault_dump+0x67
0x361fe40: 0x13d163 PANIC+0x2b
0x361fe50: 0x13900c
xxx_ZN19Memory_AddressSpace10allocPagesEi12Memory_
0x361fe70: 0x1651f5
xxx_ZN20Buffer_Uncached_Pool10addBuffersEmim+0x97
0x361feb0: 0x16510e
xxx_ZN20Buffer_Uncached_Pool14getUncachedBufE15Str
0x361fed0: 0x163f4e allocb_enet+0x2e
0x361fef0: 0x258487
xxx_Z16BCM_queueNextPktP16_LM_DEVICE_BLOCKh+0x33
Fix
Fixed the memory leak that occurred during the login process. The Data Mover will no longer panic in
this case. However, this fix does not address the login loop that could occur with this configuration
problem as this is iSCSI initiator behavior.
Service Request
18216236
Fixed in version
5.5.30.0
239
Version 5.5.30.4
Impact Level
Severity 2
Symptom
Description
When creating or refreshing iSCSI snapshots, iSCSI I/O became blocked for longer than necessary.
Fix
This behavior did not cause any noticeable user impact, but caused slight delays to I/O while snapshots
were created. A more efficient mechanism for performing locks while snapshots was identified, and
I/Os are now delayed for shorter times.
Service Request
18157032
Fixed in version
5.5.30.0
Impact Level
Severity 1
Symptom
Description
While running iSCSI replication, the Data Mover panicked with the following DART panic/fault message:
Call to abort function. The stack trace showed:
0xc0a09c88: 0x13d664 waitForReboot+0x90
0xc0a09ca8: 0x13d861 fault_dump+0x67
0xc0a09cc8: 0x13d761 PANIC+0x29
0xc0a09cd8: 0x1386ed abort+0x5
0xc0a09d00: 0xdc8e1c _ZSt9terminatev+0xe
0xc0a09d10: 0xdca7ae __cxa_pure_virtual+0x6
0xc0a09d20: 0xb0e4c1 _ZN15http_connection10releaseCnxEv+0x11
0xc0a09d40: 0xb18b8b _ZN15dicBasicSessionD0Ev+0x3b
Fix
Code was still using an http connection object after it was released. This was fixed by changing the
global variable that the replication process was using to a local variable so that more than one process
can access it simultaneously.
Fixed in version
5.5.30.0
NDMP/Backup
240

Version 5.5.30.4
Impact Level
Severity 1
Symptom
Description
An NDMP backup or restore was aborted by the user at a point where a tape change was required.
The job was not aborted, and the server_pax command showed active backup threads.
Fix
The abort command requires a lock that is held by the tape seek routine. In this case, the offset being
sought was larger than 0x7fffff * 64512. This seek never completed due to a separate bug which was
addressed in release 5.5.28.0, so the lock was never released. This caused the abort command to
hang as well. Under this condition, the lock is now released and the abort command completes
successfully.
Service Request
18288974
Customer
Premier Farnell PLC
Fixed in version
5.5.30.0
Impact Level
Severity 1
Symptom
Description
VBB Full Destructive Restore failed when restores were larger than 2 TB.
Fix
Code was changed from 32-bit block sizes to 64 bit blocks to allow for >2 TB restores.
Service Request
18605122
Customer
Scottish RE
Fixed in version
5.5.30.0
Impact Level
Severity 2
Symptom
Description
A volume level NDMP backup (VLC) of a file system larger than 2 TB did not complete. The stats seen
from server_pax and the backup application showed that the amount backed up was greater
than the size of the file system.
Fix
We now support VLC backups up to the maximum file system size which is 16 TB.
Service Request
16980420
Fixed in version
5.5.30.0
241
Version 5.5.30.4
Impact Level
Severity 2
Symptom
Description
While using NetVault with NAS version 5.5 and VBB, during a file restore of files with names greater
than 9 bytes including extension, Celerra erroneously added a letter into the original file name. For
example, for 100MB.dat0 it added "0", and for config_fileC it added "C", etc.
Fix
When the non-I18N to I18N file name conversion occurred during the restore process, an extra
character was erroneously added to file names longer than 9 bytes. This has been fixed.
Fixed in version
5.5.30.0
Impact Level
Severity 1
Symptom
Description
When unmounting a file system while an NMDP backup/restore session was running, the Data Mover
panicked with the following DART panic/fault message: >>PANIC: I/O not
progressing LastVol touched not Known.
0xdbc03e90: 0x13d1db waitForReboot+0x8b;
0xdbc03eb0: 0x13d3e3 fault_dump+0x67;
0xdbc03ed0: 0x13d2e3 PANIC+0x2b;
0xdbc03ee0: 0x703afd _ZN9RawIO_IRP13checkProgressEv+0x107;
0xdbc03fd0: 0x7039ec; ZN17IrpProgressThread5startEv+0x6;
0xdbc03fe0: 0x1406f4 Sthread_startThread_internal+0xe
Fix
If umount is called on the backup/restore file system or its PFS during a tape change process, the
code now aborts the NDMP session instead of panicking. There are no error messages generated if
the abort operation doesn't succeed. However, the backup software window, where you normally
check the status of backup/restore, contains the status of NDMP sessions.
Service Request
18502916
Fixed in version
5.5.30.0
Networking
242
The following entries describe fixed Networking problems:
Version 5.5.30.4
Impact Level
Severity 2
Symptom
Description
I/O errors occurred while doing recalls using HTTP connections with authentication enabled. This only
occurred when connecting to an APACHE-based secondary server that required authentication.
Fix
In this case, certain versions of APACHE responded with an HTTP error "206 partial
content" instead of "401 auth required". The Data Mover now treats this unexpected
response properly and file recalls work as expected.
Fixed in version
5.5.30.0
Impact Level
Severity 2
Symptom
Description
After rebooting or stopping/starting NFS services, HP/UX clients were intermittently unable to lock files
on one or more Data Movers. Rebooting the client again or rebooting the Data Mover resolved the
issue.
Fix
The Data Mover was using the wrong UDP port to send lock replies. HP/UX clients use asynchronous
locking, which require the Data Mover to learn the port that the client listens for lock requests and
replies on. Prior to NAS 5.5, the Data Mover asked the client for this information and cache it for two
minutes.
Starting with 5.5, the Data Mover only cleared this information if:
1. The client sent the Data Mover a reboot notification (which should happen if the client owned
locks), or
2. If the Data Mover contacted the client using a UDP port and is told that the port is no longer in use.
Rebooting the client could cause the UDP port it used to change, and in some situations the above two
triggers would never occur, causing the Data Mover to cache the information forever. The Data Mover
now clears its port cache if either of the two conditions are met OR after two minutes.
Service Request
18119530
Fixed in version
5.5.30.0
243
Version 5.5.30.4
Impact Level
Severity 1
Symptom
Description
../sched.cxx at line: 819 : SYSTEM WATCHDOG
0x3851f24: 0x13d600 waitForReboot+0x90
0x3851f44: 0x13d7fd fault_dump+0x67
0x3851f64: 0x13d6fd PANIC+0x29
0x3851f74: 0x13f459 system_watchdog+0x77
0x3851f84: 0x137481 _t_nmi+0x63
0x3851fe0: 0x140d94 Sthread_startThread_internal+0xe
Fix
Each time a correctable memory error occurred, a process to handle this event caused the Data
Mover's clock to be delayed by 400ms. After several of these errors happened in a short period of
time, the Data Mover's watchdog timer fired and panicked the Data Mover because it believed that an
internal process was hung. The 400ms pause no longer occurs for a correctable ECC memory error.
Service Request
18782670
Fixed in version
5.5.30.0
NFS
244
Impact Level
Severity 1
Symptom
Description
After re-exporting an NFS share using the name option (to create an alias for the share), users could
no longer access the export. Exporting the share without specifying a name, or completely
unexporting and re-exporting the share worked as expected.
Fix
Creating an NFS alias causes the datamover to create a symbolic link to the destination
share/directory, which is what clients access when mounting the alias name. If a path was already
exported when the user tried to create the alias (including if the alias already existed, but was being
re-exported), then the symbolic link would be deleted and/or not re-created, which prevented access.
This problem was introduced in 5.5.27.5, and has now been corrected.
Service Request
18559934
Fixed in version
5.5.30.0
Version 5.5.30.4
Impact Level
Severity 2
Symptom
Description
NFS lookups failed intermittently for non-UTF8 international clients using multibyte characters when
the last byte was not a complete character. This problem was seen only with names whose length
were a multiple of 4 bytes.
Fix
Code has been modified to ensure lookups show all files, and properly handle non-UTF8 names in this
situation.
Service Request
18358234
Fixed in version
5.5.30.0
Replication

Impact Level
Severity 2
Symptom
Description
When SavVol needs to be extended, the Control Station and Data Mover communicate passing ckptID
and savVol ID. The ckptID is the most recent active ckpt ID. However, between time the decision is
made to extend the SavVol and time when actual extension takes place, the most recent active ckpt ID
can change. When this occurred, a message was logged in the sys_log similar to: invalid ckpt
id 281, finding ckpt with SavVol id 360.
Fix
This message is informational only, therefore this issue was fixed by changing the severity and text of
the message. This message is now logged with severity LOG_INFO. The new message displayed in
the sys_log is: ckpt auto-extension: ckpt <FS_ID> not found,
checking the most recent ckpt on SavVol <ID>.
Service Request
17786176
Fixed in version
5.5.30.0
Impact Level
Severity 2
Symptom
Description
A Data Mover returned local time to the Control Station for fs_replicate -info -v even
though GMT was set on the Data Mover. It should have returned GMT timezone information.
Fix
Time from the Data Mover to the Control Station during replication info is now reported in GMT
timezone instead of local time if GMT timezone is set.
Fixed in version
5.5.30.0
245
Version 5.5.30.4
Storage
Impact Level
Severity 1
Symptom
Description
Environmental issues on the backend fibre loop that resulted in a "Link Down" condition caused the
driver to enter an unexpected state and then to the following panic. DART panic/fault
message: >>PANIC in file: ../dart/dart_debug_util.cxx at
line: 29 : fc_osw_dart_panic
0xf7d250: 0x13d047 waitForReboot+0x8b
0xf7d270: 0x13d24f fault_dump+0x67
0xf7d290: 0x13d14f PANIC+0x2b
0xf7d2a0: 0x7bb3ce fc_tl_handle_outbound_completion+0x49e
0xf7d2d0: 0x7b250d fc_tl_isr+0xbd
0xf7d300: 0x7aca6f cpd_osw_isr+0x6f
0xf7d330: 0x756561 _Z26fcpInterruptServiceRoutinePv+0x31
0xf7d340: 0x144148 _ZN3ISR6invokeEv+0x30
Fix
Code now resolves error states before trying to access context information presented with the fibre
channel error frame.
Service Request
18293732
Fixed in version
5.5.30.0
UFS
246
Impact Level
Severity 1
Symptom
Description
When deleting more than 500 files (of size 3MB+) on two or more file systems, the Data Mover
panicked with the following DART panic/fault message: >>PANIC in file:
../ufsvfsop.cxx at line: 277 : sync: no progress reducing
dirty list. The stack trace showed:
0x95383f30: 0x146560 waitForReboot+0x90
0x95383f50: 0x146773 fault_dump+0x67
0x95383f70: 0x14665d PANIC+0x29
0x95383f80: 0x82d9f9 _ZN21UFS_SyncThreadInitial5startEv+0x481
0x95383fe0: 0x14bdd8 Sthread_startThread_internal+0xe
Fix
While deleting or truncating more than 500 large files of 3MB+ each on more than one file system, the
process that was deleting files from one file system blocked the delete/truncation activity on the others.
This led to a Data Mover panic when the system detected that delete/truncation activity was not
progressing. This problem has been fixed.
Fixed in version
5.5.30.0
Version 5.5.30.4
Impact Level
Severity 1
Symptom
Description
An attempt to unmount a read/write mounted file system did not complete. This can happen when
Alternate Data Stream (ADS) references exist on Celerra system files.
Fix
In order for an unmount to complete, the number of file system references must fall to the number of
expected internal references. Now the references to the file system due to ADS on Celerra system
files are accounted for during unmount.
Service Request
17450584
Fixed in version
5.5.30.0
VDM
The following entries describe fixed issues related to Virtual Data

Movers:
Impact Level
Severity 2
Symptom
Description
If a user created/deleted a VDM while CIFS was stopped, the Data Mover panicked with the following
DART panic/fault message: Assertion failure:
'"DQueue::remove()"next_p->prev_p == this' on line 274 in file
'/home/cala56x/WS/5_6_x/5_6_x/Dart/server/src/include/dqueue.
hxx' >>PANIC in file: ../assert.c at line: 26 : assert failed.
0xdff75d40: 0x1463f4 waitForReboot+0x90
0xdff75d60: 0x146607 fault_dump+0x67
0xdff75d80: 0x1464f1 PANIC+0x29
0xdff75d90: 0x1cdb09 _assert+0x51
0xdff75db0: 0x497bb7
xxx_ZN11Lock_client10deleteLockER10DQueueLinkI10O
0xdff75dd0: 0x7cb18a _ZN10ObjectLockD1Ev+0x1e
0xdff75df0: 0x7d468e
xxx_ZN14File_LocksData10unlockFileEP11objLockData
0xdff75e20: 0x768125 _ZN9NTLogFile10unlockFileEv+0x6b
Fix
This problem occurred because the creation/deletion of VDM corrupted the list of open files of the local
lock client. Reworked the Lock_localClient class so that the lists of locked items are protected by
dedicated mutexes instead of the Lock_client mutex, which was held only for NFS clients. The
Lock_localClient object is used when the Data Mover itself internally sets locks on files.
Fixed in version
5.5.30.0
247
Version 5.5.29.1
Version 5.5.29.1
Celerra
Management
Applications

Impact Level
Severity 2
Symptom
Description
With a high load of 4096 mounted file systems on server_2 and server_3 with more than 2000 shares,
it took a long time to display the Replication New Screen.
Fix
Processing of this page was improved to reduce time needed to display pages with a large number of
shares.
Fixed in version
5.5.29.0
CFS
Impact Level
Severity 2
Symptom
Description
No files were scanned when a full virus scan of a file system mounted under an NMFS (Nested Mount
File System) was initiated from the Control Station. This is done using the server_viruschk
-fsscan <file system name> -create command.
Fix
The path to files in the component file systems of an NMFS are now correctly sent to the virus checker
engine. This ensures the files are correctly scanned.
Service Request
17483308
Fixed in version
5.5.29.0
CIFS
248
The following entries describe fixed Celerra File System problems:

problems:
Version 5.5.29.1
:
Impact Level
Severity 1
Symptom
Description
When a user ran the command server_cifssupport <mover> -accessright on a

share to check user rights, the Data Mover panicked with the following DART panic/fault message:
Assertion failure: 'n->back == n' on line 111 in file
'../vnodelst.cxx' >>PANIC in file: ../assert.c at line: 26 :
0xdda77820: 0x13d600 waitForReboot+0x90
0xdda77840: 0x13d7fd fault_dump+0x67
0xdda77860: 0x13d6fd PANIC+0x29
0xdda77870: 0x18872d _assert+0x51
0xdda77890: 0x5ea0e3
xxx_ZN13VFS_VnodeList17insertAtEndLockedEP9VFS_Vn
0xdda778b0: 0x5eb9dd _ZN9VFS_Vnode16releaseInactive2Ev+0x13d
0xdda778e0: 0x5eb88e
xxx_ZN9VFS_Vnode15releaseInactiveEPvi11File_Statu
0xdda77910: 0x5ebfcf
Fix
Code attempted to release an object from memory twice which resulted in this panic. The request is
now correctly executed, and the Data Mover no longer panics.
Fixed in version
5.5.29.0
:
Impact Level
Severity 2
Symptom
Description
User reported 10 second delays when accessing files from CIFS clients while CAVA was enabled.
This problem only occurred if CAVA was configured for Scan on Read.
Fix
Once a file is opened by a client and a CIFS client sends a read request, the CAVA server is instructed
to open and scan the file. When this happened, the Data Mover sent a notification to the CIFS client
that another system was attempting to access the file. Some versions of Windows did not respond in
this case, causing a deadlock until the Data Mover timed out (10 seconds later). In the meantime,
access to the file and scanning of the file were delayed. Celerra no longer sends these notifications
when the CAVA user (a special case) attempts to scan the file no longer experiences timeouts.
Service Request
17683466
Fixed in version
5.5.29.0
249
Version 5.5.29.1
:
Impact Level
Severity 1
Symptom
Description
User saw delays forming new CIFS connections. When a user that was a member of a group in a
CIFS domain that was unreachable tried to connect, the Data Mover requested information about the
group from the local domain controller which timed out attempting to reach the unreachable domain.
Fix
If the domain controller indicates that a trusted domain is quarantined (offline or unreachable), the
Data Mover will no longer attempt to lookup SIDs related to this domain.
Service Request
17424500
Fixed in version
5.5.29.0
250
Impact Level
Severity 2
Symptom
Description
The server_cifssupport -pingDC command did not work correctly with CIFS servers
built on VDMs. The VDM was not used to lookup the requested domain controller.
Fix
The server_cifssupport -pingDC command is now fully supported with VDMs.
Fixed in version
5.5.29.0
Impact Level
Severity 1
Symptom
Description
User could not access CIFS shares. The DHSM secondary server was in a hung state. After a reboot,
retrieval of archived files was working again.
Fix
Redesigned the locking mechanism used in CIFS client on the Data Mover to resolve this issue.
Service Request
17977564
Fixed in version
5.5.29.0
Version 5.5.29.1
Impact Level
Severity 2
Symptom
Description
When a Celerra CIFS Server is a member of a Windows 2003 domain, Kerberos cannot be used to
connect to a DNS alias of the server. This is because the alias name is sent in the ticket rather than the
real server name. The client sees the error "duplicate name on the network" when
attempting to access the server using the alias.
Fix
A new parameter, cifs LanmanServer.disableNameChecking, is available which

controls whether the name sent in the Kerberos ticket needs to match the server name. When the
param is on (value 1), the name sent in the Kerberos ticket does not need to match the real server
name, and the client is allowed to connect with a DNS alias. This parameter is off (value 0) by default,
which means the client is only allowed to connect with Kerberos when using the primary computer
name.
Service Request
17757418
Fixed in version
5.5.29.0
Impact Level
Severity 2
Symptom
Description
CIFS clients lost access to Celerra shares due to resource issues when there were enough events
requiring Celerra to send messages and the clients were not configured to receive them. The Celerra
server log indicated that the clients were unable to receive popup messages (messenger service was
disabled). 007-03-01 10:55:50: SMB: 3: Invalid reply from
<clientname> for NETBIOS Session Setup
Fix
Clients that do not accept Celerra's attempt to create a NetBIOS session (needed in order to send a
popup message) are now remembered. Sending messages to these clients is no longer attempted
and access is no longer lost because of this resource issue.
Clients can be reconfigured to accept NB sessions by disconnecting and reconnecting, or rebooting
the Data Mover.
Service Request
17769450
Fixed in version
5.5.29.0
251
Version 5.5.29.1
252
Impact Level
Severity 2
Symptom
Description
An Administrator was unable to add the account NT AUTHORITY\NETWORK to a local group on a

Celerra CIFS server. The Windows client saw the error message: The following error
occurred while attempting to save properties for group <<local
group name>> on computer <<datamover>>: A new member could not
be added to a local group because the member has the wrong
account type.
Fix
The NT AUTHORITY accounts can now be added to local groups on Celerra CIFS servers.
Service Request
18190668
Fixed in version
5.5.29.0
Impact Level
Severity 1
Symptom
Description
A Data Mover could not communicate with a domain controller that was separated from the Data
Mover by a firewall. The frames were dropped by the firewall because the Data Mover set the Last
Frag bit to 0 and the First Frag bit to 0. Windows based systems will set one bit, or the other, or both,
but will not reject a packet if both are not set. However, the Cisco firewall expected bits to be set and
blocked the request.
Fix
The code now sets the FirstFrag and LastFrag bits to avoid packets from being dropped by some
firewalls.
Service Request
18017442
Fixed in version
5.5.29.0
Impact Level
Severity 2
Symptom
Description
LAN backup of CIFS share failed if file extension filtering was enabled on a Data Mover and user tried
to backup and restore a file marked as "denied" by the file filtering.
Fix
File filtering is now disabled for backup/restore files to prevent this condition.
Service Request
16674998
Fixed in version
5.5.29.0
Version 5.5.29.1
Impact Level
Severity 2
Symptom
Description
When restarting CIFS after an error due to a UDP (User Datagram Protocol) stream leak, the Data
Mover panicked with the following DART panic/fault message:
Assertion failure: SIDCache::globalSidCache == NULL' on line
796 in file '../cifs_thrd.cxx' >>PANIC in file: ../assert.c at
line: 26 :
0x13cdba40: 0x145660 waitForReboot+0x90
0x13cdba60: 0x145873 fault_dump+0x67
0x13cdba80: 0x14575d PANIC+0x29
0x13cdba90: 0x1b0cf1 _assert+0x51
0x13cdbab0: 0xb85a2c _ZN10smb_server12cifsStartAllEi+0xa2
0x13cdbaf0: 0xaf5798
xxx_ZN10smb_server17parseStartCommandER6OptionRPc
Fix
A UDP stream is now preallocated when CIFS is started to prevent this Data Mover panic.
Fixed in version
5.5.29.0
Control Station

Control Station:
Impact Level
Severity 2
Symptom
Description
An attempt to display the "Mounts" tab on the "Checkpoints" page in Celerra Manager failed with the
following error: "The process for task 2466 has terminated abnormally.
It will not be restarted."
Fix
This problem occurred because the accesspolicy mount option did not have the required value and the
validation routine did not detect that. Improved validation of "primary" and "accesspolicy" mount options
to ensure that they have proper values specified.
Fixed in version
5.5.29.0
Impact Level
Severity 1
Symptom
Description
During a CNS based system upgrade, the Data Movers were not rebooted. When the Data Movers
were booted, after the upgrade, it was possible (depending on how long the Data Mover took to reboot
- 4+ minutes) for the internal interfaces to get swapped.
253
Version 5.5.29.1
254
Fix
Added checks to make sure that code to swap internal interface is attempted for CNS cabinet only if
the Data Mover is in a ready or contacted state.
Service Request
18086880
Fixed in version
5.5.29.0
Impact Level
Severity 2
Symptom
Description
An attempt to modify NetBIOS name of a standalone CIFS server failed with the following: Error
2213: server_2 : <netbios_name> : invalid name specified.
Fix
Code has been fixed to allow the modification of NetBIOS name of the standalone CIFS server.
Fixed in version
5.5.29.0
Impact Level
Severity 1
Symptom
Description
Nas partitions could be mounted read/write on both the primary and secondary Control Stations. This
caused corruption to the NAS file systems as both Control Stations simultaneously attempted to write
to the same file system.
Fix
Introduced protective code to prevent users from mounting a NAS partition read/write on a secondary
Control Station while a primary Control Station is active.
Service Request
15703410
Fixed in version
5.5.29.0
Impact Level
Severity 1
Symptom
Description
If an HLU (host LUN number) was reused by a new LUN, the details of the previous LUN using that
HLU was still cached in the Data Mover. The message "Device cXtYlZ Serial Number
has changed" was logged in the server_log. The file system creation or file system extension
command that attempted to use this HLU failed, leaving the system in an inconsistent state.
Fix
The system now rescans the bus to update the cached details of the HLUs when a serial number
mismatch is detected.
Service Request
17537628
Fixed in version
5.5.29.0
Version 5.5.29.1
:
Impact Level
Severity 2
Symptom
Description
CLI command "server_df ALL" failed with error message "Execution failed:
Segmentation violation".
Fix
This problem occurred if a query from a Control Station to a Data Mover for "df" output from a VDM
failed. Code has been fixed to post the error appropriate to the failure instead of producing an Eiffel
dump.
Service Request
17701110
Fixed in version
5.5.29.0
:
Impact Level
Severity 1
Symptom
Description
Following an in-family upgrade within 5.5, NAS database files relating to volume and file system
structure became zero-byte and needed to be restored from a nasdb backup.
Fix
This problem occurred when the /nas/var/.row_locking_db_version file was deleted prior to the
upgrade (this should not occur in normal operation). In this situation, database files are not changed,
and an error is logged to the upgrade log indicating the problem.
Service Request
17570370
Fixed in version
5.5.29.0
255
Version 5.5.29.1
CORE
256

Data Movers::
Impact Level
Severity 1
Symptom
Description
NFS threads were blocked for more than 6 minutes, which caused a Data Mover panic with the
following DART panic/fault message:
>>PANIC: System deadlocked
Service:NFSD Pool:NFSD_Exec BLOCKED for 417 seconds. The stack trace
showed:
0x61add90: 0x1465c8 waitForReboot+0x90
0x61addb0: 0x1467db fault_dump+0x67
0x61addd0: 0x1466c5 PANIC+0x29
0x61adde0: 0x15884e
xxx_ZN12cThreadsPool23processBlockedConditionER12c
0x61ade20: 0x157437
xxx_ZN12cThreadsPool9parsePoolER12cSvcParseCtxP14X
0x61ade60: 0x156858
xxx_ZN15cThreadsService17parseServicePoolsER12cSvc
0x61adeb0: 0x1565bb
xxx_ZN19cThreadsServicesMgr14_parseServicesER12cSv
0x61adf00: 0x154bac
xxx_ZN19cThreadsServicesMgr13parseServicesER12cSvc
Fix
When performing file system operations, a race condition allowed threads (such as NFS, in this case)
to fall into an infinite loop. Code has been modified to eliminate the condition and ensure the threads
are no longer blocked in an infinite loop.
Fixed in version
5.5.29.0
Impact Level
Severity 2
Symptom
Description
A file system crossed the file system size notification threshold. When the file system was extended,
the message filesystem size threshold dropped below... was never posted
to the sys_log.
Fix
When a file system usage drops below the configured threshold value due to file system extension, an
event similar to "threshold dropped below" is now posted in the sys_log.
Service Request
17424560
Fixed in version
5.5.29.0
Version 5.5.29.1
Impact Level
Severity 1
Symptom
Description
When user invoked the server_kerberos -ccache -flush command, the Data Mover panicked with the
*** Page Fault Interrupt. Virt ADDRESS: 538f22 Err code: 0
Target addr: 909090a4 **.
0x1bb43f4: 0x144d9b waitForReboot+0x8b
0x1bb4414: 0x144fa3 fault_dump+0x67
0x1bb4434: 0x13ac86 handle_fatal_fault+0xa
0x1bb4480: 0x54544a krb5_ser_ccache_init+0x82
0xafab3640: 0x54569d creds_match+0x45
0xafab3660: 0x546b7d
xxx_Z22krb5_mcc_remove_lockedP13_krb5_contextP12_
0xafab36b0: 0x54653c
xxx_Z14krb5_mcc_storeP13_krb5_contextP12_krb5_cca
0xafab36e0: 0x53cf56 krb5_get_credentials+0x19c
Fix
The Data Mover panicked when memory that had been freed was re-accessed. This problem no longer
occurs.
Fixed in version
5.5.29.0
Impact Level
Severity 1
Symptom
Description
The Data Mover panicked with the DART panic/fault message:

>>PANIC in file: ../include/stream.hxx at line: 1015 : out of
msgb.
0x65ca0: 0x13b8cb waitForReboot+0x8b
0x65cc0: 0x13baca fault_dump+0x5e
0x65ce0: 0x13b9d3 PANIC+0x2b
0x65cf0: 0x1642b9 _ZN15Buffer_MsgbPool3getEv+0xa7
0x65d10: 0x16274d allocb+0x59
0x65d40: 0x3137b2 _Z10tcp_outputP5tcpcb+0x7e4
0x65e20: 0x30f906
_Z9tcp_inputP13IPInterface_tP4msgbiiPi+0x1c34
0x65ea0: 0x2ecec6 ip_input+0x470
Fix
The problem was caused by a memory leak which was fixed by removing memory allocation during an
error path where it was not used and freed before.
Service Request
17887976
Fixed in version
5.5.29.0
257
Version 5.5.29.1
258
Impact Level
Severity 1
Symptom
Description
The Data Mover panicked with the following DART panic/fault message: *** Watch Point
exception. Virt ADDRESS: 16850f ***. The stack trace showed:
0xe9ddfd1c: 0x13d047 waitForReboot+0x8b
0xe9ddfd3c: 0x13d24f fault_dump+0x67
0xe9ddfd5c: 0x1373ee handle_fatal_fault+0xa
0xe9ddfda0: 0x387946
xxx_ZN6CLIENT15async_call_procElmPFiP4XDR_PvEPclP
0xe9ddfde0: 0x6d5f82
xxx_ZN20revokeHandleListInfo4sendEP7session+0x24a
0xe9ddfe50: 0x6ce072
xxx_ZN17fileSystemWrapper16revokeHandleListE22rev
0xe9ddff00: 0x6d3b0d
xxx_ZN7session16revokeHandleListE22revokeHandleLi
0xe9ddff30: 0x6cf3d0
xxx_ZN11hashElement16revokeHandleListE22revokeHan
Fix
Added a mutex to protect initialization of the clientHandle. In rare cases, when the first two
notifications for an MPFS client needed to be sent at the same time, a race condition occurred when
one thread was using clientHandle while another was still initializing it.
Fixed in version
5.5.29.0
Impact Level
Severity 1
Symptom
Description
When Data Mover memory dumps were generated, they were incomplete in some cases. In this
instance, incomplete dumps were caused by a stack overflow by the dump processing thread. This
prevented analysis of the memory dump while troubleshooting Data Mover issues.
Fix
By making a data structure static, stack overflows are prevented in the memory dump code path. This
ensures that memory dumps are complete.
Fixed in version
5.5.29.0
Version 5.5.29.1
FileMover

FileMover:
Impact Level
Severity 1
Symptom
Description
User reported about 10MB/sec aggregate recall rate when using FileMover with an NFS secondary
server selecting the full recall method. One problem that was identified was that the FileMover NFS
client always read 8K per read operation.
Fix
Optimized the FileMover v3 NFS client to allow the recall rate for full migration to approach 30MB/sec
over a GigE connection to an external NFS server. Loopback connections to a different file system on
the same Data Mover may achieve recall rates of 70MB/sec.
Fixed in version
5.5.29.0
Impact Level
Severity 1
Symptom
Description
User reported poor performance when retrieving archived files from Centera via a Legato Disk Xtender
solution. DHSM was configured to use HTTP to communicate with the secondary storage.
Fix
The HTTP read size has been increased from 32K to 256K for full and partial read policy methods to
improve performance.
Service Request
17836866
Fixed in version
5.5.29.0
HighRoad (MPFS)
The following entries describe fixed issues related to the MultiPath

File System (MPFS):
Impact Level
Severity 2
Symptom
Description
Attempts to unmount a file system timed out if a CIFS client using MPFS was actively connected to the
file system.
Fix
This problem was introduced in version 5.5.25.2. Code has been modified to ensure file systems can
be unmounted successfully under this condition.
Fixed in version
5.5.29.0
259
Version 5.5.29.1
Install/Config/
Upgrades
260

Impact Level
Severity 2
Symptom
Description
When performing an upgrade to Celerra version 5.5 on a CNS-14 platform, several problems were
encountered:
If Pre-Upgrade Health Check detected errors, it did not provide the option to abort the upgrade and
simply restarted the health check (without warning) when the error was acknowledged.
If Pre-Upgrade Health Check reported warnings and the user chose to stop the upgrade, the
upgrade restarted anyway.
The Pre-Upgrade Health Check did not display the complete error or warning messages.
Fix
These three issues have been addressed:

If an error is detected when running Pre-Upgrade Health Check (meaning upgrade cannot continue)
the error is displayed to the user and the upgrade exits without making changes. The upgrade can
be restarted once the problem detected by the health check is resolved.
If the user chooses to abort the upgrade after receiving a warning, the upgrade exits properly.
The entire error and warning messages are displayed when they occur.
Service Request
17554118
Fixed in version
5.5.29.0
Impact Level
Severity 2
Symptom
Description
During the installation of a secondary Control Station, or during a Control Station recovery procedure, a
message like the following was displayed:
cp: cannot stat `/nas_standby/sbin/upgrade_to_dual_cs'.
Fix
This error message was displayed indicating that the upgrade_to_dual_cs script was missing, however,
it is not needed on a secondary Control Station. This condition is normal and the error will no longer be
reported.
Fixed in version
5.5.29.0
Version 5.5.29.1
iSCSI

Impact Level
Severity 1
Symptom
Description
User lost access to two iSCSI LUNs following an upgrade from NAS 5.4 to 5.5.
Fix
In releases prior to 5.x, a problem existed when iSCSI opaque files that contained the attributes for
iSCSI snaps became larger than 8KB. During an upgrade to 5.5, iSCSI LUN upgrade could not
complete due to this condition. Code now repairs the opaque file to the right size allowing the whole
iSCSI LUN upgrade to complete.
Service Request
18160416
Fixed in version
5.5.29.0
Impact Level
Severity 3
Symptom
Description
While transferring an iSCSI snapshot (part of the replication process), the Data Mover returned
progress information above 100 percent and error code DpRequest_Busy.
Fix
The percentage has be limited to 100%. This was a replication internal calculation error and has been
corrected.
Fixed in version
5.5.29.0
Impact Level
Severity 1
Symptom
Description
While performing iSCSI snapshot deletion, a race condition caused an iSCSI replication session to be
stuck in the sync phase. The session could not proceed nor be aborted; only rebooting the Data Mover
released this deadlock.
Fix
This race condition has been eliminated which will prevent this problem.
Fixed in version
5.5.29.0
261
Version 5.5.29.1
NDMP/Backup

Impact Level
Severity 2
Symptom
Description
After creating the NDMP account on a Data Mover and giving it a 15-character password, NDMP
operations did not complete. The user list showed that the gecos field (user information field) of the
password database for the user was missing. This prevented the NDMP module from verifying the
password and caused all the NDMP operations to fail. Passwords of 14 characters or less were not
impacted by this issue.
Fix
The NDMP account entry is now created properly if the account is assigned a 15 character password.
Note: Passwords must be between 6 and 15 characters, and cannot contain white spaces, tab
characters, or any of the following symbols: / \ : ; , = * + | [ ] ? < >. The password limitation is not
NDMP specific, and applies to any user account passwords created by server_user.
Fixed in version
5.5.28.0
NFS
262
Impact Level
Severity 1
Symptom
Description
The Data Mover panicked with the following DART panic/fault message: *** Page Fault
Interrupt. Virt ADDRESS: 173943 Err code: 0 Target addr:
feeb0b13 ** . The stack trace showed:
0x1695604: 0x13d3e3 fault_dump+0x67
0x1695624: 0x137582 handle_fatal_fault+0xa
0x1695660: 0x170b25 AsyncStream+0x3df
0xead574c0: 0x16ed66 CloseStream+0x1a8
0xead57510: 0x39fb72 _ZN6CLIENT14closeRpcStreamERPv+0x66
0xead57530: 0x3a046a _ZN6CLIENT16check_connectionEv+0x288
0xead57590: 0x39fde8
xxx_ZN6CLIENT12buildRequestElmP11Rpc_RequestP17Ac
Fix
The Data Mover panicked because the callback failed to a client which requested NFSv4 delegation of
a file. This code path has been fixed.
Service Request
18117800
Fixed in version
5.5.29.0
Version 5.5.29.1
Replication

Impact Level
Severity 2
Symptom
Description
A pending checkpoint schedule should not be able to be paused, but user was able to pause it using
the nas_ckpt_schedule -pause command.
Fix
The following message is now returned when an attempt is made to pause a pending checkpoint
schedule "Error, the schedule cannot be paused".
Fixed in version
5.5.29.0
Storage
Impact Level
Severity 1
Symptom
Description
After an upgrade to 5.5, some file systems could not be mounted. The following messages were
logged in the server_log:
2007-02-11 07:06:32: STORAGE: 3: DiskMark has changed
Fields:old,new Marker:2,2 Target:0,0 Lun:1,1
sym_serial_no:xxxxxx sym_model_no:0,0 MakerType:2,175
celerraMarker:EMC Celerra,EMC Celerra
inquiryData:860F112345677890600F0/,1234567/
old_cel_id:0xxxx/,/ task_flag1:0ZZ?HIjP,0ZZ?HIjP
task_flag2:0ZZ?HIjP,0ZZ?HIjP cel_idAa,/
2007-02-11 07:06:32: STORAGE: 3: 1: Volume Basic-Volume 1
xxxc0t0l1 disk-id 2,disk -mark has changed
2007-02-11 07:06:32: ADMIN: 4: Command succeeded: volume
xxxdisk 2 c0t0l1 disk_id=2 size=4153
xxxdisk 2 c16t0l1 disk_id=2 size=4153
xxxslice 112 2097152 640 8192 2
Fix
Diskmark protection measures introduced in NAS 5.5 incorrectly detected a conflict between an old
diskmark (pre NAS 4.0 diskmark) and the current diskmark. Code now verifies the contemporary
diskmark against the physical configuration.
Service Request
17514246
Fixed in version
5.5.29.0
263
Version 5.5.29.1
Impact Level
Severity 1
Symptom
Description
While performing Replicator or SnapSure operations, the Data Mover panicked with the following
DART panic/fault message: *** GP exception. Virt ADDRESS: 6f8c8a. Err
code: 0 ***. The stack trace showed:
0x16523d8: 0x13d10f waitForReboot+0x8b
0x1652418: 0x1374b6 handle_fatal_fault+0xa
0x1652440: 0x752741 _ZN8SgIO_IRP7irpDoneEv+0x57
0x363df20: 0x7e003c _ZN14CamDisk_Device8completeEP10ccb_heade
0x363df70: 0x760bff _ZN8Scsi_Xpt16processCallbacksEv+0x3b9
0x363dfd0: 0x760c38 _ZN18Scsi_serverInitial5startEv+0x6
0x363dfe0: 0x140610 Sthread_startThread_internal+0xe
Fix
A function that processed structures recursively caused a stack overflow that led to this panic. The
function is now called non-recursively and prevents this problem.
Service Request
16767102
Fixed in version
5.5.29.0
UFS
264
Impact Level
Severity 1
Symptom
Description
../vfsnode.cxx at line: 861 : unexpected node state in
release(). The stack trace showed:
0xde699700: 0x13b7ef waitForReboot+0x8b
0xde699720: 0x13b9ee fault_dump+0x5e
0xde699740: 0x13b8f7 PANIC+0x2b
0xde699750: 0x58034a
xxx_ZN9VFS_Vnode20releaseInvalidLockedEv+0x122
0xde699780: 0x58070f
0xde6997b0: 0x521599 _ZN15File_NamingNode7releaseEv+0x27
0xde6997d0: 0x77eb38
xxx_ZN15Uthread_Initial16lookupDirOfAFileEPKcRP18
0xde699850: 0x781db3 _ZN15Uthread_Initial6accessEPKci+0x27
Fix
Eliminated a race condition that caused the code to attempt to release memory that is already
released.
Service Request
17731700
Fixed in version
5.5.29.0
Version 5.5.29.1
Impact Level
Severity 2
Symptom
Description
In a directory with thousands of hard links, an attempt to create a new hard links failed with
"Input/output error". This error was due to a name generation collision while creating DOS
8.3 names. This occurred due to the similarity of the full file names, the pattern of characters in the file
names and the fact that all files were hardlinks that pointed to the same inode
Fix
The Inode number was used to generate the DOS 8.3 name for a file and since all files shared the
same inode, this problem was more likely to occur. The DOS 8.3 name creation mechanism has been
improved to be independent of the inode number.
Service Request
17771202
Fixed in version
5.5.29.0
Impact Level
Severity 1
Symptom
Description
Removing several snaps in a one VLU while creating a file system caused file system corruption.
Errors similar to the following were reported in the /var/log/messages file of the Linux client which had
the iSCSI LUN mounted:
kernel: EXT2-fs warning: maximal mount count reached, running
xxxe2fsck is recommended
kernel: SELinux: initialized (dev sdc, type ext2), uses xattr
kernel: scsi2: ERROR on channel 0, id 0, lun 1, CDB: Write
xxx(10) 00 00 aa 17 b8 00 00 70 00
kernel: Info fld=0x0, Current sdc: sense key Medium Error ...
kernel: Additional sense: I/O process terminated
kernel: end_request: I/O error, dev sdc, sector 11173448
kernel: Buffer I/O error on device sdc, logical block 1396488
...
Fix
In this circumstance, a race condition was encountered that allowed iSCSI LUN corruption to occur.
This condition has been eliminated.
Fixed in version
5.5.29.0
265
Version 5.5.29.1
266
Impact Level
Severity 2
Symptom
Description
During periods of very high data change rate, on systems with slow back-end speeds, a Data Mover
../ufsvfsop.cxx at line: 150 : watchdog: no progress reducing
dirty list.
0x361ff30: 0x13d10f waitForReboot+0x8b
0x361ff50: 0x13d317 fault_dump+0x67
0x361ff70: 0x13d217 PANIC+0x2b
0x361ff80: 0x6502da _Z17swatchdogCallbackP5timer5tmarg+0x102
0x361ffb0: 0x13cb65 _Z13RTtimer_schediPv+0x77
0x361ffc0: 0x13e31c _ZN15SchedRT_Initial5startEv+0xf6
0x361ffe0: 0x140610 Sthread_startThread_internal+0xe
Fix
Ufslog sync threads were unable to keep up with the incoming metadata change. When the UFS log is
full, it blocks the nfs/cifs threads at the UFS layer. This results in NFS/client time-outs and causes
watchdog panics. Increased the default number of sync threads from 16 to 32 to help prevent this
condition.
Fixed in version
5.5.29.0
Version 5.5.29.1
Impact Level
Severity 1
Symptom
Description
A file system extend was started while a CIFS client was in the process of updating the security.evt file.
After the extend finished, the update to the security.evt file could not complete because of a locking
problem. The Data Mover panicked with the following: DART panic/fault message: *** Page
Fault Interrupt. Virt ADDRESS: 1982af Err code: 0 Target addr:
f8000000 **. The stack trace showed:
0x1690590: 0x13d1db waitForReboot+0x8b
0x16905b0: 0x13d3e3 fault_dump+0x67
0x16905d0: 0x137582 handle_fatal_fault+0xa
0x1690620: 0x59b2ad
xxx_ZN16File_WriteSource11updateBlockEP13Buffer_Ca
0xeb083980: 0x59b212
xxx_ZN16File_WriteSource11createBlockEmjRmR11File
0xeb0839b0: 0x59cf07
xxx_ZN13File_OpenFile9writeDataEymR11File_StatusP
0xeb083a30: 0x59e47f
xxx_ZN13File_OpenFile5writeEPhmR11File_Statusyi+0
0xeb083aa0: 0x5bd396
xxx_ZN9NTLogFile10WriteEventEP15_EVENTLOGRECORD+0
Fix
Event log updates and file system extention operations now complete successfully and the Data Mover
no longer panics under this condition. The size member of the class was not updated in cases where
writeLogHeader function returned an error. Code now properly updates the size member of this class
to eliminate the locking issue.
Service Request
17627208
Fixed in version
5.5.29.0
267
Version 5.5.28.1
Version 5.5.28.1
Celerra
Management
Applications
Impact Level
Severity 1
Symptom
Description
User created a new file system that needed to be replicated. When the GUI was used, the IP address
of the network interface he needed to use didn't show up in the replication portion of the GUI and the
interface was unusable. The interface did show up under CIFS and networking part of the GUI and
could be used when setting up Replicator from the CLI, however.
Fix
This problem occurred if an interface was configured to be used by a CIFS server in this case, the
interface did not show up for use for replication in the GUI. This problem has been resolved.
Service Request
17697574
Fixed in version
5.5.28.1
Impact Level
Severity 2
Symptom
Description
WebUI pages were not displaying properly. The following messages were logged in the
/nas/log/symapi file: "emcClarLoadMirror on Clar <CLARiion name> failed
with error 4651".
Fix
The problem occurred due to an unrecognized CLARiiON disk state. Additional disk state information
was added to prevent this condition.
Service Request
17009698
Fixed in version
5.5.28.0
CIFS
268


problems:
Version 5.5.28.1
Impact Level
Severity 2
Symptom
Description
When Access Based Enumeration was used, symbolic links that a user did not have permission to
access were visible. Note: This problem did not allow un-intended users to access the files.
Fix
Symbolic links that a user does not have access to follow are no longer visible when Access Based
Enumeration is used.
Service Request
16318766
Fixed in version
5.5.28.0
Impact Level
Severity 2
Symptom
Description
Windows systems experienced 10 second delays when deleting a file opened by other clients. This
bug was fixed with Windows XP SP2 and hotfix 811492 or W2K3 SP1.
Fix
Code now includes a workaround to avoid this problem on unpatched CIFS clients.
Service Request
17346968
Fixed in version
5.5.28.0
Impact Level
Severity 1
Symptom
Description
Administrator access of the security log (security.evt) affected CIFS user operations which require
security auditing. In particular when auditing of successful access was utilized, users saw an
interruption in access if the administrator was seeking through the security log or backing it up.
Fix
Administrative use (seeking through the log and backup) of the CIFS security log file no longer
interrupts client access.
Service Request
17172292
Fixed in version
5.5.28.0
269
Version 5.5.28.1
270
Impact Level
Severity 2
Symptom
Description
If the file system quotas grace period was set to something other than -1(Unlimited), when the soft
quota was exceeded, the grace period was applied. When the grace period expired, user could not
write data and the error "Soft quota exceeded for too long" was logged in the
server log. At this point if the grace period was then set to -1 (Unlimited), the user could write until the
hard limit was reached. However, the "Soft quota exceeded for too long"
messages continued to be logged.
Fix
The message "Soft quota exceeded for too long: is no longer logged after the
grace period is set to -1 (Unlimited).
Service Request
13202952
Fixed in version
5.5.28.0
Impact Level
Severity 1
Symptom
Description
A thread stack memory corruption with a single 0 byte occurred when the local group database was
edited manually. The Data Mover panicked with the following DART panic/fault message: ***
Invalid Opcode exception. Virt ADDRESS: c ***.
0x1629078: 0x13d05b waitForReboot+0x8b
0x1629098: 0x13d263 fault_dump+0x67
0x16290b8: 0x137402 handle_fatal_fault+0xa
0x1629140: 0x1ce3a4 _ZN3vdm9setVdmCtxEv+0x4a
0xb3c5ffa0: 0x1c8777 _ZN11vdm_service9setVdmCtxEv+0x15
0xb3c5ffb0: 0x8bbde2 _ZN13smb_threadCtx5startEv+0x306
0xb3c5ffe0: 0x14038c Sthread_startThread_internal+0xe
Fix
The fix recognizes syntax errors in the local groups and does not attempt to parse the database.
Individual users may experience permissions or access issues due to the incorrect edits, but the Data
Mover will not panic. Database entries containing syntax errors will need to be corrected by EMC
support personnel. Files without syntax errors are not affected.
Service Request
17276756
Fixed in version
5.5.28.0
Version 5.5.28.1
Impact Level
Severity 1
Symptom
Description
No answer from server when querying quotas. A CIFS thread entered an infinite loop on quota query.
Fix
This happened when an NFS client created a .etc directory on a share that was not exported from the
root level of the file system. This caused the Data Mover to attempt to locate quota information from
this directory. This problem has been fixed.
Service Request
17409222
Fixed in version
5.5.28.0
Impact Level
Severity 2
Symptom
Description
Existing local groups or local user entries were removed from the local groups database of a CIFS
server when setting a privilege to an external account which had the same RID as an existing local
group (the RID is the last 32-bit value of a SID identifying a user/group within a domain or server).
This had several consequences:
1. enumeration of local group and local users became impossible through the MMC until CIFS service
was stopped/restarted. Restarting CIFS fixes the enumeration problem, but permanently destroys
the local group and leads to other problems as described in b).
2. users experienced logon/access issues if the removed local group entry was used to grant
permissions to files and directories.
Fix
The fix changes the mechanism for addressing objects in the local group database from using the RID
to using the full SID. It may be possible to restore access levels from a NAS database backup if the
backup database has the correct entries and can be restored.
Service Request
17393358
Fixed in version
5.5.28.0
271
Version 5.5.28.1
Impact Level
Severity 2
Symptom
Description
When the security log is stored on the Data Mover root file system, it is limited to 512KB. When the log
location was moved to a user file system, the restriction should have been lifted. However, when the
log location was changed, if the size of the log was set by a GPO, the correct log size was not seen until
the Data Mover was rebooted.
Fix
When the security log location is moved to a user file system, the GPO setting for the log size is now
used. If this GPO is not set, the log size from the Registry is used (this is the same as before).
Service Request
17238312
Fixed in version
5.5.28.0
Impact Level
Severity 2
Symptom
Description
When using non-ASCII characters for client names, server was unable to return the list of open
sessions using srvmgr or MMC. When user tried to check the number of sessions on CIFS server
through computer management->shared folders->Sessions, the following error (pop up message) was
returned: System error occurred during checking session, Error
1726:remote procedure call failed.
Fix
Server now handles international names correctly.
Service Request
16929862
Fixed in version
5.5.28.0
Control Station
272

Control Station:
Impact Level
Severity 1
Symptom
Description
Celerra Monitor did not load. Data Mover events sent to the system log contained unprintable and XML
incompatible character that caused parsing issues which affected Celerra Monitor's initiation code.
Fix
Code was changed so that malformed events are no longer posted to the sys_log. They are still logged
to the server_log, however.
Fixed in version
5.5.28.0
Version 5.5.28.1
Impact Level
Severity 3
Symptom
Description
Messages indicating that Control Station services were started or stopped were not logged correctly to
the /nas/log/sys_log if a network service name contained a parenthesis. The following error was
posted in /nas/log/webui/apl_tm.log :
2007-02-01 13:25:25.735
xxxwebui-N:nasadmin@123.456.789.999:201:25745:S:26530-Modify
xxxNetwork Service FTP (data transfer, control):server_10.
xxxsh: -c: line 1: syntax error near unexpected token `(d'
xxxsh: -c: line 1: `/nas/sbin/postevent -s 6 -f 24 -i 11
xxxNetwork service FTP (data transfer, control) has been
xxxstopped on node server_10, pending a Data Mover reboot, by
xxxUID 201.'
2007-02-01 13:25:39.926
xxxwebui-N:nasadmin@10.4.10.111:201:25745:E:26530-Modify
xxxNetwork Service FTP (data transfer,
xxxcontrol):server_10.(complete=true,failed=false)
This included changes for the following network services:
xFTP (data transfer, control)
xCIFS (NETBIOS name, NETBIOS datagram, NETBIOS session, CIFS)
xNFS file locking (lockd, statd)
The log entry should have been posted in /nas/log/sys_log without a syntax error to log service
changes.
Fix
The Control Station now supports network service names containing parenthesis.
Fixed in version
5.5.28.0
CORE

Data Movers:
Impact Level
Severity 2
Symptom
Description
User was able to set incorrect settings on the Data Mover network interface. In this case, the interface
allowed the user to configure "auto" for speed and full for duplex. Speed and duplex settings must
either be both auto or both set to a manual value. This caused the interface not to be configured
properly on the primary Data Mover and prevented Data Mover failover from completing successfully.
Fix
Code now enforces both settings to be either fully auto-negotiating or set to a manual value.
Service Request
15347594
Fixed in version
5.5.28.0
273
Version 5.5.28.1
Impact Level
Severity 2
Symptom
Description
During a broadcast storm causing a lot of outbound traffic (ARP replies), the Data Mover panicked with
the following DART panic/fault message: >>PANIC in file:
../include/stream.hxx at line: 944 : out of msgb. The stack trace
showed:
0x67e40: 0x13b7ef waitForReboot+0x8b
0x67e60: 0x13b9ee fault_dump+0x5e
0x67e80: 0x13b8f7 PANIC+0x2b
0x67e90: 0x15f245 _ZN15Buffer_MsgbPool3getEv+0xa7
0x67eb0: 0x15d6d9 allocb+0x59
0x67ee0: 0x22b963 _Z7bcmreadPvP4msgbjjj+0x1b
0x67f20: 0x23a6f9
xxx_Z16BCM_queueNextPktP16_LM_DEVICE_BLOCKh+0x415
0x67f60: 0x235b3f
xxx_Z20LM_ServiceInterruptsP16_LM_DEVICE_BLOCK+0xfd
Fix
This condition occurred only when using broadcom devices. Code now properly frees all secondary
ring message blocks when a broadcom device is reset.
Service Request
15938916
Fixed in version
5.5.28.0
FileMover
274

FileMover:
Impact Level
Severity 2
Symptom
Description
While using FileMover to archive files that had already reached the hard quota set on a tree, the
creation of the stub file caused the Data Mover to return the DHSM QuotaExceeded error to the
FileMover and the archive to fail.
Fix
Changed quotas and FileMover to not count opaque (offline file) data against the user block quotas.
Code no longer includes stub files as part of the quota limit.
Service Request
16712292
Fixed in version
5.5.28.0
Version 5.5.28.1
Install/Config/
Upgrades

Impact Level
Severity 1
Symptom
Description
If user set up WebUI event notifications with multiple email addresses, an upgrade to any 5.5 version
resulted in the cwm_notify.cfg file being corrupted and all notifications stopped working after the
upgrade.
Fix
Fixed upgrade code to prevent corruption in the configuration file, and also fixed the file if it has been
broken by previous upgrades.
Fixed in version
5.5.28.0
Impact Level
Severity 2
Symptom
Description
NAS services on the Control Station have to be completely started before the ProActive Health Checks
can be performed. This was denoted by just showing a progression of Control Station reason codes
with minimal context. This problem led to confusion about the activity the health check was actually
performing.
Fix
While the installation procedure is waiting for the services to come up, a message is now displayed
explaining what is happening along with the progression of the reason codes for the Control Station.
Fixed in version
5.5.28.0
NDMP/Backup

Impact Level
Severity 1
Symptom
Description
When using NDMP to get a list of file systems on a particular Data Mover, the list included all file
systems, checkpoints, AND the following two entries: - '/' and '/.etc_common' .
Fix
Code now filters out the internal file systems ("/" and "/.etc_common").
Fixed in version
5.5.28.0
275
Version 5.5.28.1
276
Impact Level
Severity 1
Symptom
Description
If NDMP backups failed very early in the backup process, the Data Mover panicked with the following
DART panic/fault message: *** GP exception. Virt ADDRESS: 80ee8e. Err
code: 0 ***. The stack trace showed:
0x1652348: 0x13d10f waitForReboot+0x8b
0x1652368: 0x13d317 fault_dump+0x67
0x16523d0: 0x8320bb _ZN14pax_cmd_thread7syswarnEiiPcz+0xcd
0xc1f6beb0: 0x82458a _ZN14pax_cmd_thread4copyEv+0x11a
0xc1f6bf60: 0x8139fd _ZN14pax_cmd_thread8appStartEv+0x5cf
0xc1f6bfd0: 0x7e1efd _ZN15Uthread_Initial5startEv+0x11
0xc1f6bfe0: 0x140610 Sthread_startThread_internal+0xe
Fix
If NDMP backup or restore failures occurred before PAX threads were activated, this panic was
triggered due to variables that were not being deleted or cleared. The code now resets all these
variables properly.
Service Request
17627030
Fixed in version
5.5.28.0
Impact Level
Severity 1
Symptom
Description
While an NDMP backup was in progress, the Data Mover panicked with the following DART panic/fault
message: *** GP exception. Virt ADDRESS: 171635. Err code: 0 ***.
0x1696128: 0x137582 handle_fatal_fault+0xa
0x1696160: 0x1738eb _Z10STRserviceiPv+0xaf
0x3651fc0: 0x13e3e8 _ZN15SchedRT_Initial5startEv+0xf6
Fix
A race condition occurred if multiple threads attempted to close the same network socket at the same
time. The code now allows only one thread to close the socket, and the thread used to perform the
close is decided on by the session handler.
Service Request
17650124
Fixed in version
5.5.28.0
Version 5.5.28.1
Impact Level
Severity 1
Symptom
Description
When an NDMP path was longer than 255 characters, in this case 325 characters, the Data Mover
panicked with the following DART panic/fault message: *** Invalid Opcode exception.
Virt ADDRESS: c ***. The stack trace showed:
0x1602358: 0x13d047 waitForReboot+0x8b
0x1602378: 0x13d24f fault_dump+0x67
0x1602398: 0x1373ee handle_fatal_fault+0xa
0x1602420: 0x1cb6fc _Z17getCurrentVdmNamev+0x44
0xf147d380: 0x1aa3b9
xxx_ZN9logObject11suppressMsgE11logFacility11logS
0xf147d3b0: 0x1aae3d
xxx_ZN5logIO6logmsgE11logFacility11logSeverityPKc
0xf147d3e0: 0x7f614e _ZN14pax_cmd_thread8ar_closeEi+0x1b0
0xf147d4a0: 0x80127f _ZN14pax_cmd_thread10wr_archiveEi+0x575
Fix
The long path was overflowing a static buffer in Dart. Fixed strcpy and related buffer overflows;
correctly set backup status to abort when the backup is aborted; and distinguish the normal backup
completion from error/abort exit in CIFS code.
Service Request
17428682
Fixed in version
5.5.28.0
Impact Level
Severity 2
Symptom
Description
A race condition occurred between the tape device PRLI (Process Log In), and the Celerra Fibre driver.
This can be related to the number, and type of devices on the Arbitrated Loop. Working configurations
are not likely to fail, as the timing remains constant.
Fix
The PRLI race condition seen with FC-AL tape drives has been resolved.
Fixed in version
5.5.28.0
277
Version 5.5.28.1
278
Impact Level
Severity 1
Symptom
Description
../ufsvfsop.cxx at line: 149 : watchdog: no progress reducing
dirty list.
0x77f30: 0x13b8c3 waitForReboot+0x8b
0x77f50: 0x13bac2 fault_dump+0x5e
0x77f70: 0x13b9cb PANIC+0x2b
0x77f80: 0x5f2c26 _Z17swatchdogCallbackP5timer5tmarg+0x102
0x77fb0: 0x13b31b _Z13RTtimer_schediPv+0x77
0x77fc0: 0x13cac6 _ZN15SchedRT_Initial5startEv+0xf6
0x77fe0: 0x13ede0 Sthread_startThread_internal+0xe
Fix
This was due to a race condition that occurred when a checkpoint became inactive while a
server_archive copy was in progress. This problem no longer occurs.
Service Request
16639446
Fixed in version
5.5.28.0.
Impact Level
Severity 2
Symptom
Description
NDMP DAR restore failed if it involved tape seek of more that 0x7fffff (8388607) records. The following
error was displayed in the server_log: NDMP: 3: DAR restore positionTape:
seek failed at offset 4666371436963
Fix
The problem occurred due to a SCSI limitation. Implemented tape seek independent of the SCSI
limitation to prevent the restore failure.
Service Request
17272514
Fixed in version
5.5.28.0.
Version 5.5.28.1
Networking
Impact Level
Severity 2
Symptom
Description
Users were unable to delete or rename files with non-ASCii but UTF-8 encoded names via FTP.
Fix
This occurred because the Data Mover attempted to convert files with unicode names into UTF-8
twice. Now they are converted only once.
Service Request
17225454
Fixed in version
5.5.28.0.
NFS
Impact Level
Severity 2
Symptom
Description
If NFSV4 read delegations were open on the file system, the file system could not be unmounted. The
server_umount command hung and the following error was logged in the server_log:
Service:NFSD Pool:NFSD_Exec BLOCKED for 334010 seconds
Fix
Fixed the deadlock caused by File_LocksData and Delegation mutexes being taken in the wrong
order.
Fixed in version
5.5.28.0
Impact Level
Severity 1
Symptom
Description
When clients were holding too many application locks, server available vnode resources decreased
and eventually led to vnode cache full situation.
Fix
Code now sends the following event to the administrator (1 event every 5 minutes) if the count of
nodes referenced by file locks is above the limit defined by param lockd.vnodepercent. Refer
to the Parameters Guide for detailed information about the vnodepercent parameter.
Fixed in version
5.5.28.0
279
Version 5.5.28.1
Replication
280

Impact Level
Severity 1
Symptom
Description
After configuring Celerra Replication and rebooting, the Data Mover panicked with the following DART
panic/fault message: *** Invalid Opcode exception. Virt ADDRESS: c***.
0x1626c08: 0x13d05b waitForReboot+0x8b
0x1626c28: 0x13d263 fault_dump+0x67
0x1626c48: 0x137402 handle_fatal_fault+0xa
0x1626c80: 0xb73fe4 _ZN9DpRequest7executeEv+0xa0
0xed46ff80: 0xbb8cf3
xxx_ZN16DP_RepDicService7executeEP17dicIncomingPa
0xed46ffc0: 0xbb8bf7 _ZN15DP_RepDicThread5startEv+0x2b
0xed46ffe0: 0x14038c Sthread_startThread_internal+0xe
Fix
The panic occurred when the destination Data Mover attempted to configure the same replication
session against two different IP addresses on the destination Data Mover. Normally this should never
happen, and in this case occurred due to incorrect editing of the NAS database. In the event that the
same replication session is configured against two different IP addresses, the Data Mover no longer
panics. An error message is now logged in the server log, and the second attempt to configure the
replication fails. The first replication setup will succeed.
Service Request
17089756
Fixed in version
5.5.28.0
Impact Level
Severity 1
Symptom
Description
Replication could not be restarted after suspending it prior to an out-of-family NAS code upgrade. The
format of the names of Replication restart checkpoints was changed between the NAS 5.3 and the
NAS 5.4 and above code families. The name format was changed to allow failed-over file system pairs
to be clearly identifiable in the WebUI.
Fix
The NAS 5.4 and 5.5 code families now recognize the 5.3 restart checkpoint name formats, and
restart IP replication if it was suspended before a NAS code upgrade from a version with the older
name format. This problem was only seen if Replication was suspended for the duration of the NAS
code upgrade.
Service Request
15998261
Fixed in version
5.5.28.0
Version 5.5.28.1
SRDF

Impact Level
Severity 2
Symptom
Description
User issued the command fs_timefinder -Mirror off on server13_fs1, but it seemed to
be stalling at 10921 MB remaining. Meanwhile, user tried a nas_fsck on its mirrorfs (CK5069) and
the command returned: "Error 2237 Precondition violated".
Fix
FSCK operation is now rejected on a Mirror file system with the following: Error: "This
operation in not allowed on Mirrored file system (${FS_NAME})".
Fixed in version
5.5.28.0
Storage
Impact Level
Severity 1
Symptom
Description
User was planning to migrate from NS600 to NS40 with CX3-80 backend. During failover tests of the
backend SPs, user pulled SP-A. Three LUNS did not failover and the Data Mover panicked with the
following DART panic/fault message: >>PANIC in file: ../BVolumeIrp.cxx at
line: 323 : IO failure on Vol:151, blkNo:16, SKey:0x54. The stack
trace showed:
0x368de80: 0x13d1e0 waitForReboot+0x90
0x368dea0: 0x13d3e9 fault_dump+0x67
0x368dec0: 0x13d2e9 PANIC+0x2b
0x368ded0: 0x75f991
xxx_ZN8SgIO_IRP9irpFailedE14BlockIO_Statusy+0x309
0x368df20: 0x7f182e
xxx_ZN14CamDisk_Device8completeEP10ccb_header+0xec
0x368df70: 0x77227f _ZN8Scsi_Xpt16processCallbacksEv+0x3b9
0x368dfd0: 0x7722b8 _ZN18Scsi_serverInitial5startEv+0x6
0x368dfe0: 0x1406f8 Sthread_startThread_internal+0xe
In the server_log we see LUN trespass messages with LUNs numbered greater than 65000:
03-23 13:37:59: STORAGE: 4: Trespass lun 65480 to SPA
Fix
When the HLU number was 128 or larger (0x80, high order bit set), sign extension caused the number
to be stored as 0xFF80. The code now uses unsigned values, which prevents this issue. Note, ALU
numbers larger than 128 are not an issue.
Service Request
17872124
Fixed in version
5.5.28.0
281
Version 5.5.28.1
Impact Level
Severity 2
Symptom
Description
A NAS installation failed, and the installation setup screen showed: Setting up Storage
Groups.... Failed. Navisphere Event log message showed: Create
Initiator : Exception: No IP Prop Object defined for port.
Fix
FLARE on the CX3-10 does not support adding initiators without an IP address. This was a Navi CLI
problem which was addressed by updating the Navi CLI release on the Control Station to version
6.24.1.5.0.
Fixed in version
5.5.28.0
UFS
282
Impact Level
Severity 1
Symptom
Description
When a user tried to create a new file system using an NBPI (number of bytes per inode) option of
81920000, the Data Mover panicked with the following DART panic/fault message:
*** Divide exception. Virt ADDRESS: 6c8f68 ***. The stack trace showed
0x1627d4c: 0x13d05b waitForReboot+0x8b
0x1627d6c: 0x13d263 fault_dump+0x67
0x1627d8c: 0x137402 handle_fatal_fault+0xa
0x1627f80: 0x6c7759
xxx_Z15buildFileSystemP9IO_ObjectPKcR6Option+0x149
0xc26a9970: 0x5f2294
xxx_ZN18UFS_FileSystemKind18handleOtherCommandEPK
0xc26a9b90: 0xb15a74
xxx_ZN19MGFS_FileSystemKind18handleOtherCommandEP
0xc26a9bb0: 0x5ca7be
xxx_Z17File_BuildCommandR6OptionP18VFS_FileSystem
0xc26a9bd0: 0x5cd3bd _ZN12filecmd_desc6invokeER6Option+0x6d
Fix
Added upper boundary check to NBPI value while creating file systems to eliminate panic. The max
NBPI allowed is 64274432. If NBPI is more than 64274432, an error message similar to the following
is logged in the server_log:
2007-04-04 18:43:15: FSTOOLS: 3: Maximum nbpi (number of
xxxbytes per inode) supported is 64274432 bytes
2007-04-04 18:43:15: ADMIN: 3: Command failed: file build
xxxuxfs 306=306 s=8192000 i=81920000 w=0
Service Request
15777049
Fixed in version
5.5.28.0
Version 5.5.27.5
Version 5.5.27.5
Celerra
Management
Applications

Impact Level
Severity 1
Symptom
Description
Attempts to create a new network device using Celerra Manager caused the browser (both Internet
Explorer 6.0 and Netscape 7.1) to hang. This occurred when first opening the page or after changing
the Data Mover or device type.
Fix
The issue was resolved and the page is fully functional.
Fixed in version
5.5.27.1
Impact Level
Severity 2
Symptom
Description
The configuration file was improperly formatted when multiple NTP servers were configured via the
GUI.
Fix
Corrected the way multiple NTP server entries are handled.
Service Request
14786517
Fixed in version
5.5.27.1
Impact Level
Severity 2
Symptom
Description
Internal Control Station management processes used a lot of Control Station CPU and memory which
triggered a Control Station failover to standby.
Fix
The process was requesting configuration data for an entire system which utilized memory
inefficiently. Information is now requested per Data Mover and Virtual Data Mover which consumes
less memory.
Service Request
14786517
Fixed in version
5.5.27.0
283
Version 5.5.27.5
Impact Level
Severity 2
Symptom
Description
When the underlying log collection script could not create its configuration file, if it had not yet been
created (log collection has not been run) or if the configuration file was deleted, the following error
message was displayed on the Support > Log Collection page:
The process for task 7607 has terminated abnormally. Root file
system is full. Task will not be restarted.
Fix
Modified the UI status message to read: "Make sure the nasadmin user has
read/write access to the /nas/site directory and that the
directory has sufficient free disk space (2KB)."
Fixed in version
5.5.27.0
Impact Level
Severity 2
Symptom
Description
When user tried to delete an available destination file system mounted on a custom mountpoint, the
following error message was displayed on the UI screen: Replication->Available Destinations->New
(FS): Delete available destination <FSNAME>. <MOVER_NAME> :
<MountPoint_Path>: does not exist.
Fix
Code has been modified to ensure users can delete the Available Destination Filesystem mounted on
a custom mountpoint.
Fixed in version
5.5.27.0
CIFS
284

problems:
Impact Level
Severity 1
Symptom
Description
The C$ administrative share of a CIFS server was mapped and data was being moved between file
systems. Due to a race condition, this resulted in a loss of CIFS access.
Fix
Moves/renames across file systems are now properly handled when performed from C$.
Service Request
15864077
Fixed in version
5.5.27.0
Version 5.5.27.5
Impact Level
Severity 1
Symptom
Description
Operations hung when a Linux 2.6.18-1 client attempted to access files on a CIFS-mounted file
system on the Data Mover.
Fix
When the Data Mover responded to NT Create AndX (file open/create) CIFS calls, it returned
extended responses even if not requested by the client. Older Linux kernels did not have a problem
with this, but the response is unexpected by newer Linux releases. The Data Mover now only returns
extended responses if requested by the CIFS client.
Fixed in version
5.5.27.0
Impact Level
Severity 1
Symptom
Description
While trying to connect to an Active Directory, the Data Mover panicked with the following DART
panic/fault message: Assertion failure: 'false' on line 121 in file
'../ber/BerOctetString.cxx' >>PANIC in file: ../assert.c at
line: 23 : assert failed. The stack trace showed:
0xcdf16990: 0x13d05b waitForReboot+0x8b
0xcdf169b0: 0x13d263 fault_dump+0x67
0xcdf169d0: 0x13d163 PANIC+0x2b
0xcdf169e0: 0x17f77f _assert+0x53
0xcdf16a00: 0x3c791d
xxx_ZN14BerOctetStringC1ER11InputStreami+0x149
0xcdf16a70: 0x3fd2e8
xxx_ZN17LdapClientDecoder6decodeER11InputStreamP9
0xcdf16aa0: 0x3ca9dd
xxx_ZN9BerTaggedC1ER11InputStreamR6BerTagiR13BerT
0xcdf16ad0: 0x3c56c6
xxx_ZN10BerElement4readER11InputStreamR13BerTagDe
Fix
This was caused when incorrectly formatted data was received in response to a LDAP lookup to a
domain controller. Error messages indicating this condition are now logged to the server_log and the
Data Mover no longer panics.
Service Request
16153984
Fixed in version
5.5.27.0
285
Version 5.5.27.5
286
Impact Level
Severity 1
Symptom
Description
When trying to make changes to local group permissions, the Data Mover panicked with the following
DART panic/fault message: >>PANIC in file: ../sched.cxx at line: 802 :
0x3623f34: 0x13d047 waitForReboot+0x8b
0x3623f54: 0x13d24f fault_dump+0x67
0x3623f74: 0x13d14f PANIC+0x2b
0x3623f84: 0x13eb5a system_watchdog+0x7c
0x3623f94: 0x1370e5 _t_nmi+0x63
Fix
This panic occurred while trying to set privileges on a local group that had the same name as a CIFS
server on the Data Mover. This has been fixed in this release.
Service Request
15190866
Fixed in version
5.5.27.0
Impact Level
Severity 2
Symptom
Description
MPDTranslation failed with InvalidArgument errors from function getAllNames(). Upon exit, it
suggested running dirConvert, which administrators found confusing since it is an internal name. The
external process is using the uc_config -convert command.
Fix
Error message was changed to suggest running the uc_config -convert command.
Service Request
13306319, 13284707
Fixed in version
5.5.27.0
Impact Level
Severity 2
Symptom
Description
A CIFS user didn't have write access when accessing a DFS path if they didn't have write access to
the DFS root share.
Fix
The user now only needs read access to the DFS root share in order to be redirected. Their access is
determined by checking the permissions on the target share and file system.
Service Request
14565754
Fixed in version
5.5.27.0
Version 5.5.27.5
Impact Level
Severity 2
Symptom
Description
The creation of 100,000 local groups took a long time (over 50 hours) compared to 90 minutes on a
Windows host. The creation process became exponentially slower, 11,000 groups were created after
one hour.
Fix
Added a new hash mechanism to speed up local item access using their names which brings
performance in line with Windows performance.
Service Request
15141747
Fixed in version
5.5.27.0
Impact Level
Severity 2
Symptom
Description
After migrating files from a Windows server to Celerra, if a file was accessed with WinXP first and then
the same file was accessed with Win98, the application crashed. Conversely, if the file was first
accessed with Win98 after the migration and then the same file was accessed using WinXP, the
application did not crash.
Fix
We now mask the extended DOS attributes when replying to a search request from either a Win98 or
WinXP application.
Service Request
16052977
Fixed in version
5.5.27.0
Impact Level
Severity 1
Symptom
Description
The following messages were logged in the server log: UFS: 3: foundNode(valid),
with treeId x different than passed treeId x. . This does not indicate a
problem as the tree IDs actually matched.
Fix
The error message was logged erroneously, and is no longer logged when the tree IDs match.
Service Request
13373409
Fixed in version
5.5.27.0
287
Version 5.5.27.5
288
Impact Level
Severity 2
Symptom
Description
The tree quota report command timed out when the param cifs.lookup.secmapOnly was
set to 0 and external Usermapper was used for mappings.
Fix
The tree quota report does not need the SID to UID/GID mappings so we no longer make the query
which was causing the timeout.
Service Request
16730608
Fixed in version
5.5.27.0
Impact Level
Severity 2
Symptom
Description
User could not backup security credentials using NDMP. This occurred when a CIFS user was
connected to the Data Mover as a member of the Domain Administrators group and the other was a
member of the Administrators groups.
Fix
In this case, backup operator privileges required by the NDMP user to back up security credentials
were not being assigned properly to the NDMP user. This has been fixed.
Service Request
15060607
Fixed in version
5.5.27.0
Impact Level
Severity 2
Symptom
Description
When a file/directory was created by a CIFS client using the TransNTCreate command, if a Security
Descriptor (SD) was specified in the command it was stored as is. However, when only the owner was
specified, the SD was created with owner/group but with no DACL.
Fix
When the DACL is not specified in the TransNTCreate request, it is now built based on inheritable
permissions on the parent and added to the Security Descriptor.
Service Request
15899164
Fixed in version
5.5.27.0
Version 5.5.27.5
Impact Level
Severity 1
Symptom
Description
The Data Mover panicked with the following DART panic/fault message: *** Page Fault
Interrupt. Virt ADDRESS: b500b3 Err code: 0 Target addr:
ffe567e8 ** . The stack trace showed:
0xfd2e5c: 0x14b36d _waitForReboot+0x7d
0xfd2e6c: 0x14b56e _fault_dump+0x56
0xfd2e80: 0x139142 _t_machine_check+0x36
0xfd2ea4: 0xb50106
xxx___13cMsrpc_PolicyUlP13cMsrpc_PolicyUlP17cMsrpc_
0xaeecdd90: 0xb07a08
xxx_samrOpenAlias__19cMsRPC_SamrSimulateP13smb_th
0xaeecde04: 0xb00f64
xxx_pipeFunctions__19cMsRPC_SamrSimulateP13smb_th
0xaeecde3c: 0xb4103f
xxx_pipeProcess__13cPipeSimulateP13smb_threadCtxP
0xaeecde74: 0xb411f2
xxx_pipeTrans2__13cPipeSimulateP13smb_threadCtxPP
Fix
Code now protects MSRPC objects in memory from being released before they're no longer used,
which prevents this condition from occurring.
Service Request
16773384
Fixed in version
5.5.27.0
Impact Level
Severity 1
Symptom
Description
After an administrator tried to delete a CIFS server, commands from the Control Station to the Data
Mover timed out.
Fix
The delete operation was attempted at the same time as a GPO update was being run. A mutex is
used to prevent multiple access to the GPO cache file. In this scenario the mutex wasn't being
unlocked in the error path resulting from the second attempt to access the GPO cache file.
Service Request
16655992
Fixed in version
5.5.27.0
289
Version 5.5.27.5
290
Impact Level
Severity 2
Symptom
Description
When using LDAP (iPlanet/SunOne) for UID/GID resolution, user authentication was slow if the user
belonged to a group with a large number of members.
Fix
As the other members of the large group are not needed for user authentication, the complete list is no
longer requested from the LDAP server.
Service Request
16614218
Fixed in version
5.5.27.0
Impact Level
Severity 2
Symptom
Description
When the internal Usermapper database is being exported new mappings can not be created. The
export is done hourly as part of the nasdb backup process. Typically the export is fast so there is no
impact. However, if the system is busy and Usermapper database is large it can result in users that
have not connected before being denied access.
Fix
A new mechanism is used when exporting the Usermapper database which does not prevent new
mappings from being created.
Fixed in version
5.5.27.0
Impact Level
Severity 1
Symptom
Description
User reported CIFS outage on a single Data Mover. The following errors were reported in the
server_log:
2006-12-09 12:54:51: LGDB: 3: Syntax error line 3067 decoding
xxx0x33d9:0x0:\
:\ xxx%<domain>\_<group name>|2|33900|<SID>
2006-12-09 12:54:51: LGDB: 3: server(10) account(18)
xxxlocalgroups(813) localusers(8)
2006-12-09 12:54:51: LGDB: 3: Database seems corrupted, try
xxxto use backup instead till fixed
Fix
This problem occurred because a local group database entry was created with an empty name. Code
now prevents any database entry from being created without a name string.
Service Request
16851838
Fixed in version
5.5.27.0
Version 5.5.27.5
Impact Level
Severity 1
Symptom
Description
The Data Mover panicked with the following DART panic/fault message: *** GP exception.
Virt ADDRESS: 396116. Err code: 0 ***. The stack trace showed:
0x12981c8: 0x13b7ef waitForReboot+0x8b
0x1298208: 0x135bb8 handle_fatal_fault+0xa
0x1298250: 0x3b36af _ZN12LdapResponseaSERKS_+0x47
0xd0819e30: 0x3b3e05
xxx_ZN11LdapService7connectEP10LdapClient+0x18d
0xd0819ec0: 0x3aa720
xxx_ZN10LdapClient7connectERK6StringiS2_S2_+0x214 xxx
0xd0819f40: 0x3d38a7
xxx_ZN16LdapDomainSunOne19connectToLdapDomainEv+0
0xd0819fc0: 0x3d3ab1
xxx_ZN16LdapDomainSunOne21reconnectToLdapDomainEv
Fix
User had configured the Data Mover to perform hostname lookups to an iPlanet server. This caused a
deadlock when non-iPlanet services (CIFS) needed to perform hostname lookups. Resource locking
mechanisms were changed to prevent this deadlock.
Service Request
16915314
Fixed in version
5.5.27.0
Impact Level
Severity 1
Symptom
Description
If user specified a localpath without a leading "/", instead of failing the command, the Data Mover
panicked with the following: >>PANIC in file: ../assert.c at line: 26 : \n
assert failed.
Fix
Code has been modified to avoid this NULL pointer assert.
Fixed in version
5.5.27.0
Impact Level
Severity 2
Symptom
Description
When a CIFS user maps to a share of a tree quota and they have user quota set on the tree, they see
the size of whole tree rather than their user quota.
291
Version 5.5.27.5
292
Fix
When a user quota is set on the tree level, the CIFS user will see this as the size instead of the whole
tree.
Note: In the event that the size of the tree is smaller than the user's quota on the tree, they will see the
size of the tree quota since that is what will limit their usage.
Service Request
17212910
Fixed in version
5.5.27.0
Impact Level
Severity 1
Symptom
Description
User modified the ap_req_chksum_type in the /Library/Preferences/edu.mit.Kerberos file of the OS X

10.4 (and rebooted). When OS X client tried to access a CIFS server, the Data Mover panicked with
the following: >>PANIC in file: ../malloc.cxx at line: 213 : \n
malloc() - incorrect size - not a positive value.
Fix
Fixed the problem that caused a Data Mover to panic after changes to the Kerberos preference file
when an OS X client was trying to access a CIFS server.
Fixed in version
5.5.27.0
Impact Level
Severity 2
Symptom
Description
NTMigrate only migrated NT4 domain global groups. Windows 2000/2003 domain local groups were
ignored.
Fix
Ntmigrate is now able to migrate Windows 2000/2003 domain local groups following the same
migrations rules as those used in an NT4 environment.
Fixed in version
5.5.27.0
Version 5.5.27.5
Impact Level
Severity 1
Symptom
Description
The Data Mover sends LDAP bind and search requests to Windows domain controllers and other
(e.g., iPlanet) LDAP Servers. The search results are returned in a (BER) encoded format. If the results
are corrupted in the network (e.g., a bit is flipped, or a byte is nulled out), errors can occur while
decoding the results. Depending on where the error occurs in the byte stream, a panic (assert) could
occur when an error is detected during decoding. The Data Mover panicked with the following DART
panic/fault message: Assertion failure: 'false' on line 139 in file
'../ber/BerCharacterString.cxx' >>PANIC in file: ../assert.c
at line: 23 : assert failed. The stack trace showed:
0xd8d04dd0: 0x13b7ef waitForReboot+0x8b
0xd8d04df0: 0x13b9ee fault_dump+0x5e
0xd8d04e10: 0x13b8f7 PANIC+0x2b
0xd8d04e20: 0x17540f _assert+0x53
0xd8d04e40: 0x388190
xxx_ZN18BerCharacterStringC2ER11InputStreami+0x16
0xd8d04ee0: 0x38f347 _ZN10BerUTCTimeC1ER11InputStreami+0x1d
0xd8d04f70: 0x3896fd
xxx_ZN10BerElement4readER11InputStreamR13BerTagDe
0xd8d04fb0: 0x38ba20
xxx_ZN14BerOctetStringC1ER11InputStreami+0x110
Fix
Instead of panicking, the Data Mover now handles each error as gracefully as possible. Error handling
depends on where (in the response message) the error was detected.
Service Request
16889478
Fixed in version
5.5.27.0
Impact Level
Severity 1
Symptom
Description
The Data Mover panicked with the following >>PANIC in file: ../sched.cxx at
line: 722 : SYSTEM WATCHDOG. The stack trace showed:
0x67f34: 0x13b7ef waitForReboot+0x8b
0x67f54: 0x13b9ee fault_dump+0x5e
0x67f74: 0x13b8f7 PANIC+0x2b
0x67f84: 0x13d2de system_watchdog+0x7c
0x67f94: 0x1358d1 _t_nmi+0x63
0x67fe0: 0x13eb30 Sthread_startThread_internal+0xe
Fix
A malloc bucket was overwritten with an INVALID CIFS request or reply. Code now protects from this
kind of panic/fault condition. It reports the name and type of client that sent the invalid request in the
server_log.
Service Request
16007008
Fixed in version
5.5.27.0
293
Version 5.5.27.5
Impact Level
Severity 1
Symptom
Description
../sched.cxx at line: 802 : SYSTEM WATCHDOG. The stack trace showed:
0x3623f34: 0x13d05b waitForReboot+0x8b
0x3623f74: 0x13d163 PANIC+0x2b
0x3623f84: 0x13eb6e system_watchdog+0x7c
0x3623f94: 0x1370f9 _t_nmi+0x63
0x3623fe0: 0x14038c Sthread_startThread_internal+0xe
Fix
When the user does not use the same credential as applibnt and * is used as a filter, the code now
follows the correct path, and the Data Mover no longer panics.
Service Request
16922576
Fixed in version
5.5.27.0
Impact Level
Severity 2
Symptom
Description
User configured the "shutdown" option for viruschecker as shutdown=cifs in a single CAVA server
environment. This caused CIFS to become permanently disabled until manual intervention to restart
CIFS was taken.
Fix
If viruschecking is configured with only one CAVA server and the "shutdown=cifs" option is specified,
code now forces shutdown=no, and the discrepancy is logged and highlighted in the server_viruschk
outputs. An error similar to shutdown=cifs invalid for single CAVA/AV
server configuration, reverting to shutdown=no. Configure
another CAVA/AV server to use shutdown=cifs was logged in the server log
and sys_log
If this occurs while using the GUI, the following error is displayed: shutdown=cifs invalid
for single CAVA/AV server configuration, reverting to
shutdown=no.
294
Service Request
14363675
Fixed in version
5.5.27.0
Version 5.5.27.5
Control Station

Control Station:
Impact Level
Severity 1
Symptom
Description
When the NASDB backup is run, many commands are run to provide a good backup copy of the NAS
data base. These commands were logged to the cmd_log, and added confusion when
troubleshooting an issue, or looking for a command issued by an administrator.
Fix
The cmd_log now only contains information pertaining to commands run by users. Commands run by
the automatic backup process are no longer recorded in the cmd_log.
Service Request
13973299
Fixed in version
5.5.27.0
Impact Level
Severity 2
Symptom
Description
The nas_fs -xtend command failed with the error "sizes are not the same". The
extended size was written to NAS database, but the file system size was not updated per the output of
the nas_fs -size command.
Fix
Code now compares the Control Station and Data Mover volume information before attempting a FS
extend. If the component volumes for the hyper are not the same on Control Station and Data Mover,
an error is reported and the FS is not extended. If the number of missing volumes on Data Mover are
2 or more, an error is reported. If there is only 1 volume missing on the Data Mover, user is able to
extend with this missing volume. If user tries to extend by size or by using any other volume, he will get
an error message.
Service Request
14725974
Fixed in version
5.5.27.0
295
Version 5.5.27.5
Impact Level
Severity 1
Symptom
Description
A specific series of events for extending, snapping, and refreshing a BCV file system corrupted the
Control Station database, if the file system had checkpoints.
The steps are
1. perm unmount the R2 snap file system
2. delete the R1 snap file system
3. delete the imported file system
4. extend the file system on the R1 box
5. import and snap again on the R2 box.
The problem may manifest itself in several ways.
1) The snap may fail.
2) The snap works but won't mount
3) The snap works and mounts, but refresh fails.
296
Fix
When importing a file system, code now removes references to backups of the file system. Backups are
not imported so references to them in the database are invalid.
Service Request
15773245
Fixed in version
5.5.27.0
Impact Level
Severity 1
Symptom
Description
The Data Mover was rebooted twice when user issued the server_cpu server_2
-reboot -monitor command. This did not occur if the -monitor option was not used. This
problem only occurred on NS or NSX systems with a single Data Mover.
Fix
The problem occurred because state information about the Data Mover was not flushed prior to the
Data Mover reboot. After the Data Mover restarted, the Control Station believed that the Data Mover
still needed to be rebooted. Code now saves the state of the Data Mover properly.
Service Request
16685060
Fixed in version
5.5.27.0
Version 5.5.27.5
Impact Level
Severity 2
Symptom
Description
If /nas/server/slot_x/export or export.shares file contains an invalid entry, the server_export

server_x all command failed, and did not export the valid entries that followed the invalid
entry.
Fix
Modified the code to log an error when an invalid export entry is encountered but also continue to
process export entries.
Service Request
15409808
Fixed in version
5.5.27.0
Impact Level
Severity 3
Symptom
Description
Need to change the severity of AC Power loss dial home
Fix
User experienced an 8 hour outage as a result of AC power loss in his environment. The outage was
compounded by the fact that the dial-home for the A/C power loss was ranked as Informational.
Service Request
15394610
Fixed in version
5.5.27.0
Impact Level
Severity 3
Symptom
Description
An attempt to export CIFS shares failed from the command line or GUI if the share name contained a
space or the path being exported contained a space in its name. This problem only occurred when
using VDMs.
Fix
Corrected the problem that prevented share or path names with a space from being exported.
Service Request
14790769
Fixed in version
5.5.27.0
297
Version 5.5.27.5
298
Impact Level
Severity 1
Symptom
Description
Celerra Monitor did not load. Data Mover events sent to the system log contained unprintable and XML
incompatible character that caused parsing issues which affected Celerra Monitor's initiation code.
Fix
Code was changed so that malformed events are no longer posted to the sys_log. They are still logged
to the server_log, however.
Fixed in version
5.5.27.0
Impact Level
Severity 2
Symptom
Description
The fs_ckpt -Create command failed with Error 5005: InvalidArgument. The server_log showed
various Logical Volumes not found and "invalid VolPool argument". The issue was a backend disk
failure, but the error returned did not indicate that.
Fix
Code now checks the error before sending checkpoint build command. If there are any errors in the
volume access commands, the following error message is returned: volume access
command failed.
Fixed in version
5.5.27.0
Impact Level
Severity 2
Symptom
Description
NSX Control Station did not recognize the newest SNMP traps from the UPS system, indicating that
A/C power was restored. By not recognizing the new trap, the Control Station presumed that one of
two A/C power feeds to the UPS was faulted. When the second feed to the UPS was lost, the Control
Station began emergency shutdown even though the UPS had a valid A/C power source.
Fix
The new SNMP trap is defined on the Control Station, ensuring that when A/C power is restored to the
UPS the Control Station records the event properly.
Service Request
16546392
Fixed in version
5.5.27.0
Version 5.5.27.5
Impact Level
Severity 2
Symptom
Description
The standby Data Mover log showed "NAS_DB: Error 2" message once per month. There were
no adverse effects to the user's system.
Fix
User had 2 instances of the NAS_DB script causing conflict and the script was attempting to run on
the standby Data Mover. Changed the script to ensure only one instance of script is running and no
attempt is made to run it on the standby Data Mover.
Service Request
16574052
Fixed in version
5.5.27.0
Impact Level
Severity 2
Symptom
Description
While running iSCSI replication, nas_cel -modify ... or nas_cel -delete ... CLI
commands failed with the following error message: Error 2237: Execution failed:
not_off: Precondition violated. [HASH_TABLE.forth].
Fix
A condition caused the pointer in the global file system table to point to the wrong entry. This condition
has been fixed, and CLI commands complete successfully.
Fixed in version
5.5.27.0
CORE

Data Movers:
Impact Level
Severity 1
Symptom
Description
User created a data unavailability event by setting up duplicate IP addresses on Data Movers on two
different Celerras. This problem went undetected for a period of time and caused intermittent and
extended data unavailability.
Fix
When the Data Mover detected that one of its IP addresses was duplicated on another host, it
reported this condition in its server_log. This was detected whether the other host was a Celerra or
not. This condition is now reported as an alert in the WebUI.
Service Request
14383269
Fixed in version
5.5.27.0
299
Version 5.5.27.5
300
Impact Level
Severity 2
Symptom
Description
While trying to create a new connection on an offline path, the Data Mover panicked with the following
DART panic/fault message: >>PANIC in file: ../dskdump.cxx at line:
1450 : InputOutput error. The stack trace showed:
0x3615d20: 0x13d10f waitForReboot+0x8b
0x3615d40: 0x13d317 fault_dump+0x67
0x3615d60: 0x13d217 PANIC+0x2b
0x3615d70: 0x853b64 _Z11paniccfgcmdR6Option+0x178
0x3615eb0: 0x167f62
xxx_ZN11cfgcmd_desc6invokeEPKcPP11CCMDMessageii+0x
0x3615f20: 0x168233
xxx_ZN11cfgcmd_desc8takeFileEP18Stdio_AbstractFile
0x3615f70: 0x1741ae _Z12start_configv+0x18c
0x3615fc0: 0x1785ee Config_StartSystem+0x6
Fix
When creating local path, a check has been added to prevent users from creating a new connection on
an existing offline directory.
Service Request
16397628
Fixed in version
5.5.27.0
Impact Level
Severity 1
Symptom
Description
../malloc.cxx at line: 327 : Memory corrupted. The stack trace showed:
0x54d63d60: 0x13d10f waitForReboot+0x8b
0x54d63d80: 0x13d317 fault_dump+0x67
0x54d63da0: 0x13d217 PANIC+0x2b
0x54d63db0: 0x1876f7 free+0x69
0x54d63de0: 0x187d10 _ZdlPv+0x10
0x54d63df0: 0xc0b583 _ZN14UFSDinodeToolsD1Ev+0x2d
0x54d63e00: 0xc0f8b7
xxx_ZN11UFSMetadata12ProcessFilesEmP6dinodeP8AclC
0x54d63ef0: 0xc1216f
xxx_ZN15ProcessFileTask12processInodeEP6dinodemRi
Fix
Fixed a double-free condition that was found in a code path.
Service Request
16661006
Fixed in version
5.5.27.0
Version 5.5.27.5
Impact Level
Severity 1
Symptom
Description
While using FS copy as part of user's standard backups which runs daily, the frequency of the
mount/umount operation during the backup process caused numerous opening and closing of
volumes. This led to a race condition which caused the Data Mover to panic with the following DART
panic/fault message: *** Invalid Opcode exception. Virt ADDRESS: c
***. The stack trace showed:
0x12924b8: 0x13b8c3 waitForReboot+0x8b
0x12924d8: 0x13bac2 fault_dump+0x5e
0x12924f8: 0x135c6c handle_fatal_fault+0xa
0x1292570: 0x69ff76 _ZN7LVolume21displayUserListLockedEv+0x4e
0xd0da97d0: 0x6a01b9 _ZN7LVolume13handleOverlapEv+0x25
0xd0da97e0: 0x6a0142 _ZN7LVolume15handleReferenceEv+0x3a
0xd0da9800: 0x6aac0c
xxx_ZN7LVolume4openE13IO_ObjectKindR11VolumeErrno
0xd0da9860: 0x6a81bf
xxx_ZN14LVolume_Config19locateAndOpenVolumeEPKcPP
Fix
Code was fixed to take care of the race condition that occurred between open and close of volume. A
function was changed to secure the process for its entirety.
Service Request
15705185
Fixed in version
5.5.27.0
Impact Level
Severity 1
Symptom
Description
When an NMAP (Network Mapper) network scan was run against the Data Mover, commands run
from the Control Station to this Data Mover hung. This was due to the httpd thread being blocked on
ports 5080 and 5081.
Fix
Changed AcceptPushStream such that when processing an incoming connection and that connection
is reset, an error will be returned instead of blocking.
Service Request
15645839
Fixed in version
5.5.27.0
301
Version 5.5.27.5
302
Impact Level
Severity 1
Symptom
Description
User created 500 1GB files, established checkpoint replication and deleted all files. The Data Mover
panicked with the following DART panic/fault message:
>>PANIC in file: ../ufsvfsop.cxx at line: 149 : watchdog: no
progress reducing dirty list. The stack trace showed:
0x6df30: 0x13b7ef waitForReboot+0x8b
0x6df50: 0x13b9ee fault_dump+0x5e
0x6df70: 0x13b8f7 PANIC+0x2b
0x6df80: 0x5f12d2 _Z17swatchdogCallbackP5timer5tmarg+0x102
0x6dfb0: 0x13b247 _Z13RTtimer_schediPv+0x77
0x6dfc0: 0x13c9ca _ZN15SchedRT_Initial5startEv+0xf6
0x6dfe0: 0x13eb30 Sthread_startThread_internal+0xe.
Fix
The deadlock condition that existed within the file system layer during peak read/write activity when a
quota file was being modified has been eliminated.
Service Request
15548004
Fixed in version
5.5.27.0
Impact Level
Severity 1
Symptom
Description
When an NMAP (Network Mapper) network scan was run against the Data Mover, commands run
from the Control Station to this Data Mover hung. This was due to the httpd thread being blocked on
ports 5080 and 5081.
Fix
Changed AcceptPushStream such that when processing an incoming connection and that connection
is reset, an error will be returned instead of blocking.
Service Request
15645839
Fixed in version
5.5.27.0
Version 5.5.27.5
Impact Level
Severity 2
Symptom
Description
The system dialed home with the error: Celerra Manager:

IndicationDistributer : Invalid XML End.
Fix
This problem occurred when the file system ID for certain web GUI objects was determined to be
zero.This condition was reported as an error. This is a normal condition and is no longer reported as
an error.
Service Request
14651316
Fixed in version
5.5.27.0
Impact Level
Severity 2
Symptom
Description
fs_copy commands failed intermittently with the following error: Error 4001: unable to
connect to host:
Fix
This command failed if another process in the Data Mover was performing an ICMP ping at the same
time this command was executed. For example, a user running the server_ping command or the
CIFS domain controller selection process could cause this conflict. Code was changed to allow
multiple ICMP pings to be run in parallel.
Service Request
13764931
Fixed in version
5.5.27.0
Impact Level
Severity 2
Symptom
Description
Worm files whose retention period had expired could not be deleted from Windows clients if they were
RW. RO files could be deleted as per CWORM semantics.
Fix
The Data Mover now resets the RO bit on a worm-expired file which enables the deletion of the file.
Fixed in version
5.5.27.0
303
Version 5.5.27.5
Impact Level
Severity 1
Symptom
Description
../dart/dart_debug_util.cxx at line: 29 : fc_osw_dart_panic .
0xf96a10: 0x13d05b waitForReboot+0x8b
0xf96a30: 0x13d263 fault_dump+0x67
0xf96a50: 0x13d163 PANIC+0x2b
0xf96a60: 0x7ae8c1 fc_dvm_process_link_up_notification+0x51
0xf96a90: 0x7ae08e fc_dvm_handle_event+0x16e
0xf96b00: 0x7c0e45 fc_tpm_process_link_up+0x195
0xf96b80: 0x7c85cf fc_tl_set_link_up_state+0x5f
0xf96ba0: 0x7c898b fc_tl_handle_link_up+0x2bb.
This panic occurred due to a fabric abort.
Fix
Code was fixed so that a panic doesnt occur during a fabric abort. The driver is now being forced into
a kill or rescan and wait until the connection comes up again.
Service Request
15700571
Fixed in version
5.5.27.0
FileMover
304

FileMover:
Impact Level
Severity 1
Symptom
Description
User was unable to delete and/or modify DHSM connections. The following error was displayed in the
command log:
2006-11-15 19:42:08.620 db:201:11148:E: fs_dhsm -c fs16
-delete 0: failed to complete command.
Fix
This occurred when a connection marked for deletion was referenced for other DHSM commands
because of a reference counter leak. This has been fixed and code now recognizes connections
marked for deletion and no longer uses them for incoming requests.
Service Request
16548888
Fixed in version
5.5.27.0
Version 5.5.27.5
Impact Level
Severity 1
Symptom
Description
The Data Mover panicked with the following DART panic/fault message: *** Divide
exception. Virt ADDRESS: c4265e ***. The stack trace showed:
0x1297c3c: 0x13b7ef waitForReboot+0x8b
0x1297c5c: 0x13b9ee fault_dump+0x5e
0x1297c7c: 0x135bb8 handle_fatal_fault+0xa
0x1297cd0: 0xc42599 __udivdi3+0x29
0xd979d8b0: 0xa90798
xxx_ZN15MGFS_FileSystem13dumpConnStatsEii+0xba8
0xd979dbd0: 0xaa1582 _Z4mgfsR6Option+0x65e
0xd979dc60: 0x15b330 _ZN11cfgcmd_desc6invokeEPKcii+0x124
0xd979dcd0: 0x1eef0e _Z10xml_cfgcmdP7macData+0x282
Fix
Code was incorrectly dividing by zero when measuring data sectors. This problem has now been
corrected.
Service Request
16720670
Fixed in version
5.5.27.0
Impact Level
Severity 1
Symptom
Description
A data recall operation can result in a 0 byte file that no longer has the stub file attributes if the HTTP
secondary server crashes.
Fix
The Celerra now verifies that the secondary server is returning the correct number of bytes before
converting this into an online file. Therefore, even if the secondary server crashes, the stub file will
retain its offline attributes so that it can retrieve the data when the secondary server is back in
operation.
Service Request
17165892
Fixed in version
5.5.27.0
305
Version 5.5.27.5
Impact Level
Severity 1
Symptom
Description
Access to offline files was lost when the DHSM (FileMover) connection to the remote host was
changed to use a different NFS version. Changing the connection involves deleting the existing
connection and recreating it specifying a different choice for NFS version.
Fix
The connection can now be deleted and recreated using a different NFS version (nfsv2 or nfsv3)
without affecting access to the offline files.
Fixed in version
5.5.27.0
Impact Level
Severity 2
Symptom
Description
If BAR (Bulk Attribute Retrieval) could not convert a file name from UTF-8 to the locale for NFS style
scans, the scan operation terminated prematurely. This issue was related to the DHSM API and was
only related to NFS style scan operations.
Fix
If BAR cannot convert file name from UTF-8 to the locale, it will output the name in UTF-8 and
continue, without aborting the job.
Fixed in version
5.5.27.0
Impact Level
Severity 2
Symptom
Description
If multiple CIFS DHSM connections were configured with the same user account and secondary
server, user was unable to modify the password after updating the password on the PDC. The
operation failed with a logon failure 0xc000006d visible in the server log.
Example problem case: account=administrator
1. \\<secondary DHSM system>\test_share1
2. \\<secondary DHSM system>\test_share2
If there was only one DHSM CIFS connection, this operation worked fine.
306
Fix
The session management logic was fixed to handle this situation.
Service Request
15771917
Fixed in version
5.5.27.0
Version 5.5.27.5
Impact Level
Severity 2
Symptom
Description
The Data Mover panicked with the following DART panic/fault message: XLT Unicode:
writes over malloc TrailerData. The stack trace showed:
0x3b8af800: 0x145ac0 waitForReboot+0x90
0x3b8af820: 0x145cd3 fault_dump+0x67
0x3b8af840: 0x145bbd PANIC+0x29
0x3b8af850: 0xf5a826 _Z14panicIfEnabledPKcPKv+0x82
0x3b8af880: 0xf5a92c _Z18checkEndOfBuffer16PKt+0x7a
0x3b8af890: 0xf5af40 _Z12XltToUnicodePKcPtRt+0x96
0x3b8af8c0: 0x10db940
xxx_ZN20Remote_Protocol_CIFS15remoteBlockReadEjR
0x3b8af970: 0x10cfa9b
xxx_ZN10MGFS_Vnode15remoteBlockReadEjR11File_Sta
Fix
In some conditions during CDMS migrations with multibyte international characters, the length passed
into Unicode translation function was truncated. The correct character length is now being passed to
the function.
Fixed in version
5.5.27.0
HighRoad (MPFS)

File System (MPFS):
Impact Level
Severity 2
Symptom
Description
User had an MPFS file systems mounted and rebooted the Data Mover. After the reboot, the file
system need to be FSCK'd. During this process the Data Mover panicked with the following DART
panic/fault message: *** Page Fault Interrupt. Virt ADDRESS: 5e534c
Err code: 0 Target addr:
fcfcfd68 **. The stack trace showed:
0x1698064: 0x13d1e0 waitForReboot+0x90
0x16980a4: 0x137582 handle_fatal_fault+0xa
0x16980e0: 0x585225 _ZN15File_NamingNode9referenceEv+0xf
0xebc95930: 0x6f16b8
xxx_Z11fmp_mount_3P9mountArgsP7svc_reqPP11FreeRes
0xebc959d0: 0x6ef0fb
xxx_ZN26fmp_IterSvc_Parent_InitialC1EP16Rpc_Svc_A
0xebc95a40: 0x3abf4e
xxx_ZN21Rpc_Svc_Child_Initial13svc_getreqsetEP7rp
0xebc95c30: 0x3ae8a0
xxx_ZN25Rpc_IterSvc_Child_Initial7svc_runEv+0x102
Fix
The code now checks for this condition and handles it correctly. FSCK will run to completion.
Fixed in version
5.5.27.0
307
Version 5.5.27.5
Impact Level
Severity 1
Symptom
Description
During times of very heavy file sharing, such as a multi-writer application, file corruption occurred.
Fix
Code now checks for and prevents the condition that caused this corruption.
Fixed in version
5.5.27.0
Install/Config/
Upgrades
308

Impact Level
Severity 2
Symptom
Description
During Pre Upgrade Health Checks (PUHC) it was found that the system had been installed using the
wrong control LUN sizes. This required a reinstall to correct.
Fix
During a fresh install, the sizes of the control LUNs are now verified.
Service Request
14911411
Fixed in version
5.5.27.0
Impact Level
Severity 2
Symptom
Description
A NAS code upgrade failed part way through the process with the error: CRITICAL FAULT:
Unable to quiesce dart logging.
Fix
Two extra checks have been added to the Pre-Upgrade Health Check (PUHC) to detect this problem
before upgrade actually starts. The first one performs additional checks on the Dart log volume, the
second one adds additional checks to make sure that the Control Station System LUNs are correctly
configured.
Service Request
14766060
Fixed in version
5.5.27.0
Version 5.5.27.5
:
Impact Level
Severity 1
Symptom
Description
There was a race condition in the handling of MPFS protocol request. If a request was being processed
and the client retransmitted the request before the Data Mover replied, the following error was seen in
the server log:
<filesystem> failed: mount_share failed: MPFS
mount: Cannot allocate memory Error mounting /mnt/<filesystem>
via MPFS.
Fix
Code now handles processing of this condition correctly.
Service Request
16725580
Fixed in version
5.5.27.0
Impact Level
Severity 3
Symptom
Description
Data Movers did not failover (in case of a panic or other issue on a Data Mover) because no standby
was configured. This is normal behavior, but running in this state could result in extended outages.
Fix
Added reporting tool to alert users of the danger of running in an unprotected environment (with no
standbys configured).
Service Request
14541820
Fixed in version
5.5.27.0
iSCSI

Impact Level
Severity 1
Symptom
Description
When managing iSCSI LUN Replication via the cbm_replicate utility, the following response was
received: Validating destination...replication destination
invalid: NBS_DPRQST_dic_cannotAuthentify.
Fix
Code now splits http_access.conf lines into multiple ones, limited to 255 characters each, to prevent
them from being truncated. See the Installing Celerra iSCSI Host Components technical module for
more information about the cbm_replicate command.
Fixed in version
5.5.27.0
309
Version 5.5.27.5
NDMP/Backup
310

Impact Level
Severity 1
Symptom
Description
When user aborted an NDMP backup process, the Data Mover panicked with the following DART
panic/fault message: >>PANIC in file: ../addrspac.cxx at line: 431 :
Out of memory. The stack trace showed:
0xccd49cf0: 0x13b8c3 waitForReboot+0x8b
0xccd49d10: 0x13bac2 fault_dump+0x5e
0xccd49d30: 0x13b9cb PANIC+0x2b
0xccd49d40: 0x137c1e
0xccd49da0: 0x17b7d2 _Z11more_memoryiP9free_list+0x4e
0xccd49dc0: 0x17b87f malloc+0x89
0xccd49df0: 0x17bf91 _Znwm+0x15
0xccd49e00: 0x17bfbe _Znam+0xc
Fix
The panic occurred because of a memory leak condition that occurred during a backup failure (media
error, abort, etc.). This has been fixed.
Service Request
16708730
Fixed in version
5.5.27.0
Impact Level
Severity 2
Symptom
Description
When performing an NDMP restore using EDM (EMC Data Manager) the name of the directory to
which the files were restored was duplicated. For example, if the original file was named /dir1/dir2/file,
the file was restored to /dir1/dir2/dir1/dir2/file.
Fix
This problem occurred with any NDMP host using NDMP v2 protocol. This has been fixed.
Service Request
16136528
Fixed in version
5.5.27.0
Version 5.5.27.5
Impact Level
Severity 1
Symptom
Description
When using NDMP and SnapSure to backup stub files, the Data Mover panicked with the following
DART panic/fault message: *** Page Fault Interrupt. Virt ADDRESS:
197fcc Err code: 0 Target addr: feeb0b0f ** . The stack trace showed:
0x1665d40: 0x13d1d3 waitForReboot+0x8b
0x1665d60: 0x13d3db fault_dump+0x67
0x1665d80: 0x13757a handle_fatal_fault+0xa
0x1665e70: 0xb5c06a
xxx_ZN10DHSM_Vnode15remoteBlockReadEjR11File_Statu
0xa621d9d0: 0xb40f16
xxx_ZN10MGFS_Vnode14migrateToBlockEjPP13Buffer_Ca
0xa621db40: 0xb5bc85
xxx_ZN10DHSM_Vnode14migrateToBlockEjPP13Buffer_Ca
0xa621db60: 0xb5a17f
0xa621dc00: 0x589901
xxx_ZN9File_Fork17readBlockInternalEjRP13Buffer_C
Fix
Code has been fixed so that Stub File attribute backup works correctly with NDMP and SnapSure.
Service Request
16767310
Fixed in version
5.5.27.0
311
Version 5.5.27.5
312
Impact Level
Severity 1
Symptom
Description
While running an NDMP backup, the Data Mover panicked with the following DART
panic/fault message: >>PANIC in file: ../bufpool.cxx at line:
309 :
couldn't get a free page. The stack trace showed:
0xc53bdb70: 0x13b7d3 waitForReboot+0x8b
0xc53bdb90: 0x13b9d2 fault_dump+0x5e
0xc53bdbb0: 0x13b8db PANIC+0x2b
0xc53bdbc0: 0x15aa77
xxx_ZN18Buffer_Cached_Pool17makeCachedBuffersEi+0
0xc53bdbf0: 0x138799
xxx_ZN19Memory_AddressSpace16makeCachedBufferEv+0
0xc53bdc10: 0x15a625
xxx_ZN18Buffer_Cached_Pool15getAnyCachedBufEji+0x
0xc53bdc40: 0x5a6783 _ZN14UFS_FileSystem11allocBufferEv+0x3f
0xc53bdc70: 0x5a5f25
xxx_ZN14UFS_FileSystem10readBlocksEliRiPP13Buffer
Fix
The Data Mover panicked because NDMP tuning parameters were set too high which caused Data
Mover resource issues. Reduced maximum values of the following 5 PAX params to avoid panic while
backing up large file systems simultaneously.
PAX.nRestore was 64, now 16;
PAX.nPrefetch was 32, now 8;
PAX.paxWriteBuff was 1024, now 256;
PAX.paxReadBuff was 512, now 256;
PAX.paxStatBuff was 2048, now 512.
Service Request
14126046
Fixed in version
5.5.27.0
Version 5.5.27.5
Impact Level
Severity 1
Symptom
Description
When using NDMP and SnapSure to backup stub files, the Data Mover panicked with the following
197fcc Err code: 0 Target addr: feeb0b0f ** . The stack trace showed:
0x1665d40: 0x13d1d3 waitForReboot+0x8b
0x1665d60: 0x13d3db fault_dump+0x67
0x1665e70: 0xb5c06a
xxx_ZN10DHSM_Vnode15remoteBlockReadEjR11File_Statu
0xa621d9d0: 0xb40f16
xxx_ZN10MGFS_Vnode14migrateToBlockEjPP13Buffer_Ca
0xa621db40: 0xb5bc85
xxx_ZN10DHSM_Vnode14migrateToBlockEjPP13Buffer_Ca
0xa621db60: 0xb5a17f
Fix
Code has been fixed so that Stub File attribute backup works correctly with NDMP and SnapSure.
Service Request
16767310
Fixed in version
5.5.27.0
Impact Level
Severity 1
Symptom
Description
While performing an NDMP backup with NDMP file history enabled and jumbo frames enabled on the
Network interface, the Data Mover panicked with the following: DART panic/fault
message: >>PANIC in file: ../malloc.cxx at line: 327 :
Memory corrupted. The stack trace showed:
0x3147bb90: 0x13d1db waitForReboot+0x8b
0x3147bbb0: 0x13d3e3 fault_dump+0x67
0x3147bbd0: 0x13d2e3 PANIC+0x2b
0x3147bbe0: 0x187aaf free+0x69
0x3147bc10: 0xc0df39 _ZN9NdmpdData9freeNlistEv+0x39
0x3147bc20: 0xc0ef25 _ZN9NdmpdData7cleanupEv+0x8d
0x3147bc80: 0xc0c1d5
xxx_ZN21NdmpConnectionHandler17connectionHandlerE
0x3147bcc0: 0xc26ff9 _ZN11ndmpInitial4forkEv+0xd3
Fix
NDMP with history and GUI now properly handle jumbo frame sizes and the Data Mover no longer
panics.
Service Request
16787182
Fixed in version
5.5.27.0
313
Version 5.5.27.5
Networking
Impact Level
Severity 2
Symptom
Description
If the WINS server was unavailable, and a NetBIOS name registration request was sent, numerous
events were logged in the server_log and the sys_log.
Fix
Code now logs an event only if the server is down after two retries, therefore reducing the number of
events in the sys_log and server_log.
Fixed in version
5.5.27.0.
NFS
314
Impact Level
Severity 1
Symptom
Description
After a Data Mover reboot, NFS services were unavailable for several hours.
Fix
This problem occurred because a host on the user's network had created thousands of NFS lock
entries with unreachable client IP addresses on the Data Mover. NFS services were unavailable as
the Data Mover attempted to notify each of the unreachable clients that the Data Mover had rebooted
and they should reclaim their locks. The following enhancements were made to avoid this problem:
Increased the number of lock notifications that can be sent simultaneously from 4 to 32.
Code limits the number of IP addresses that can be used to create locks for a single host to 100.
Service Request
16576408
Fixed in version
5.5.27.0
Impact Level
Severity 1
Symptom
Description
While running NFSv4 tests, the server_umount command failed. The following error was displayed:
"Error 16: server_2 : Device or resource busy".
Fix
When umount was invoked with only expired states, they weren't flushed and were keeping a
reference on the file system, thus preventing umount from completing. This has been fixed.
Fixed in version
5.5.27.0
Version 5.5.27.5
Impact Level
Severity 1
Symptom
Description
An NFSv4 client retained file delegations (in this case, more than 500k delegated files) for a long
period of time. Eventually, all the vnodes were used and the Data Mover panicked with the following
DART panic/fault message:
Assertion failure: 'n != NULL' on line 290
in file '../vfsnode.cxx' >>PANIC in file: ../assert.c at
line: 23 : assert failed.
0x93f09af0: 0x144caf waitForReboot+0x8b
0x93f09b10: 0x144eb7 fault_dump+0x67
0x93f09b30: 0x144db7 PANIC+0x2b
0x93f09b40: 0x1b0243 _assert+0x53
0x93f09b60: 0x73a970 _ZN9VFS_VnodenwEmP14VFS_FileSystem+0x52
0x93f09b80: 0xe9a923
xxx_ZN15DHSM_FileSystem8newVnodeEm8dirTypes+0x3b
0x93f09bb0: 0x77e05f
xxx_ZN14UFS_FileSystem16findExistingNodeEmR11File
0x93f09c00: 0x77debb
xxx_ZN14UFS_FileSystem8findNodeEmR11File_Statust8
Fix
Now aging delegations are automatically purged and NFSv4 can no longer use up all the nodes. This
can be managed using the new param nfsv4.vnodePercent.
Values: vnodepercent 0 0xffffffff, Default: 80
Limits the number of files and directories that CIFS can open to a percentage of the maximum number
of vnodes in the system.
Example:80 = CIFS can open up to 80% of the total number of vnodes.
Note: changes do not take effect until after the Data Mover reboots.
Fixed in version
5.5.27.0
Replication

using Celerra Replicator, SnapSure or Checkpoints
315
Version 5.5.27.5
316
Impact Level
Severity 1
Symptom
Description
The resync of a replication session resulted in a full fs_copy instead of a differential one. This was
because replication had just been established right before the failover was issued. This meant failover
was run before any delta sets had time to playback.
Fix
The special case of failover before playback has run is now properly handled. A flag is used to signify
this scenario so it is recognized when resync is run.
Service Request
16509740
Fixed in version
5.5.27.0
Impact Level
Severity 1
Symptom
Description
The Data Mover panicked when accessing the target file system of Celerra Replicator relationship.
The following error messages were logged in the server log:
readBlock() : Volume 226 Bad Block Index 3fffffff0
1145886817: UFS: 3: readBlock failed, err 16
2006-04-24 15:53:37: UFS: 3: Read-only file system 0x67
corrupted!!.
Fix
The target file system was corrupted. An initial fs_copy was run between the source and target file
systems but was aborted. The user then performed a differential copy from an earlier checkpoint to a
newer checkpoint, and replicator allowed this file system to be converted to an active file system.
However, since the original checkpoint hadn't been completely copied, the resulting file system was
inconsistent. Users are now prevented from running a differential copy if the original copy is
incomplete.
Service Request
15096447
Fixed in version
5.5.27.0
Version 5.5.27.5
Impact Level
Severity 1
Symptom
Description
When an fs_copy session was started and the target file system was immediately unmounted, the
Data Mover panicked with the following DART panic/fault message: *** Invalid Opcode
exception. Virt ADDRESS: c ***. The stack trace showed:
0x1666198: 0x13d1d3 waitForReboot+0x8b
0x16661b8: 0x13d3db fault_dump+0x67
0x1666220: 0xc8077e _ZN8VolMCast15checkAndReserveEv+0x2c
0xc71c7d70: 0xc8099d _ZN8VolMCast4openEv+0x17
0xc71c7d90: 0xc8b5f3
xxx_ZN18SparceFSCopySender9retryOpenEv+0x1a7
0xc71c7ee0: 0xc8d9eb
xxx_ZN18SparceFSCopySender16DoCopyFromBitMapEy+0x
0xc71c7f30: 0xc8c964
xxx_ZN18SparceFSCopySender13StartDiffCopyEv+0x3c6 xxxxxx
Fix
When fs_copy starts, the reference to the target volume is now taken on the volume first and the
pointer stored in the object for further access. This prevents the panic when the target is unmounted.
Note: fs_copy cannot complete if the target file system is unmounted.
Service Request
17303198
Fixed in version
5.5.27.0
SRDF

Symmetrix Remote Data Facility::
Impact Level
Severity 2
Symptom
Description
When using TimeFinder with RAID5 protection for BCV devices, if the new TimeFinder-Clone license
was not present when the fs_timefinder command was issued, the following error was
generated:
Error 5013 : SYMAPI_C_LICENSE_CHECK The feature being
requested is not successfully licensed.
Fix
The SYMAPI package now includes the TimeFinder-Clone license, which enables TimeFinder/FS
functionality when Celerra volumes are created on RAID5 devices.
Service Request
16084771
Fixed in version
5.5.27.0
317
Version 5.5.27.5
318
Impact Level
Severity 2
Symptom
Description
User was doing a DR test on their RDF setup. The nas_rdf -activate worked fine. The
nas_rdf -restore completed without reporting an error, but the Data Movers were failed over
to their local standbys. When an attempt to failback the Data Mover was made, the following was
returned:
2006-10-15 17:44:30.579 server_2:0:10660:E: server_standby
server_2 -r mover: replace_storage: replace_volume: failed
to complete command.
Fix
The nas_rdf -restore did not complete because the Data Movers on the source side had not
completed rebooting. The fix checks that if any servers are in reboot pending stage, nas_rdf
-restore waits for the reboot to complete.
Service Request
16084307
Fixed in version
5.5.27.0
Impact Level
Severity 2
Symptom
Description
While attempting to run a nas_rdf -restore command, failback of one Data Mover failed with
the following error:
Error 2201: server_3 : unable to acquire lock(s), try later
Fix
This problem occurred when another process held a lock on certain lines of the NAS database. Code
will now retry the operation until the lock from the other process is cleared.
Service Request
16759312
Fixed in version
5.5.27.0
Version 5.5.27.5
Impact Level
Severity 2
Symptom
Description
During the server_standby phase of the /nas/sbin/nas_rdf -active command, the

SRDF standby Data Mover's network interfaces were not properly configured. This condition existed
because this Data Mover was a production Data Mover prior to being assigned as a SRDF standby, and
the previous network configuration was not removed. Messages seen in the RDF logs were similar to:
server_2.faulted.rdf :
server_2.faulted.rdf : going standby
rdf : going active
replace in progress ...failed
failover activity complete
replace_net: fsn0:No such device or addressvdm01:No such
xxxdevice or addressvdm02:
Fix
Code now ensures that SRDF standby Data Movers do not have network interfaces configured.
Service Request
14267361
Fixed in version
5.5.27.0
Impact Level
Severity 2
Symptom
Description
If the disktype option was set for STD, when using the fs_timefinder command, it uses the first
available disks. In this case, the remote disk was the first available, so it used R1STD instead of a local
disk. When user attempted to do a snap, these "available" R1STD disks were still in a Synchronized
state and the snap failed with the following:
: Executing command -> (/nas/bin/fs_timefinder FARSN_1_R2P008
-S -n NEARSN_1_R2P008)
: Command output -> ...operation in progress (not
interruptible)...Error 5013: SYMAPI_C_INVALID_RDF_STATE
The device is not in a valid RDF state for this operation
failed to complete command Error 2208: : exec failed)
Fix
If disk_type is STD/BCV, code now only selects the local STD/BCV disks.
Fixed in version
5.5.27.0
319
Version 5.5.27.5
Impact Level
Severity 1
Symptom
Description
If a BCV file system was previously corrupted and user tried to run fs_timefinder mirror on/off/refresh
or snapshot on the corrupted BCV file system, the Data Mover on which the BCV file system was
mounted panicked with the following DART panic/fault message: >>PANIC in file:
../ufsdirhash.hxx at line: 638 :
UFS_DirHashIter_dir3::getEntry: mangled entry. The stack trace showed:
0xedc39bf0: 0x13d10f waitForReboot+0x8b
0xedc39c10: 0x13d317 fault_dump+0x67
0xedc39c30: 0x13d217 PANIC+0x2b
0xedc39c40: 0x647c5f
xxx_ZN20UFS_DirHashIter_dir38getEntryEm14BufInten
0xedc39c80: 0x656ef3
xxx_ZN20UFS_DirHashIter_dir34findEmPKcm13File_Nam
0xedc39cb0: 0x665b3f
xxx_ZN20UFS_DirHashIter_dir34findEmPK18File_NameC
0xedc39cf0: 0x657bf1
xxx_ZN15UFS_DirOps_dir37scandirEP9UFS_VnodePK18Fi
0xedc39d80: 0x66155c
xxx_ZN15UFS_DirOps_dir311getSynonymsEP9UFS_VnodeP
Fix
Code now prevents users from running fs_timefinder mirror on/off/refresh or snapshot a group file
system containing corrupted BCV member file system. attempt to run timefinder operation on a bcv fs
that is mounted. The following message is displayed: attempt to run timefinder
operation on a bcv fs that is mounted. This operation is not
permitted. You need to umount the bcv fs first.
Service Request
17291330
Fixed in version
5.5.27.0
Storage
320
Impact Level
Severity 1
Symptom
Description
After upgrading firmware on a Cisco MDS switch from version 2.03 to 3.02, Data Movers could not
reboot.
Fix
Bug in emulex bios/firmware prevented the Data Mover HBAs from working with certain FC switches
with certain code revs. HBA firmware has been upgraded to fix this problem.
Service Request
16138632
Fixed in version
5.5.27.0
Version 5.5.27.5
Impact Level
Severity 2
Symptom
Description
The Data Mover panicked with the following even though an available path to the backend was
available. DART panic/fault message: >>PANIC in file: ../BVolumeIrp.cxx at
line: 293 :
IO failure despite all retries/failovers. The stack
trace showed:
0x3613ea0: 0x13d05b waitForReboot+0x8b
0x3613ec0: 0x13d263 fault_dump+0x67
0x3613ee0: 0x13d163 PANIC+0x2b
0x3613ef0: 0x743d7b
xxx_ZN8SgIO_IRP9irpFailedE14BlockIO_Statusy+0x13f
0x3613f20: 0x7d127e
xxx_ZN14CamDisk_Device8completeEP10ccb_header+0xec
0x3613f70: 0x751f0d _ZN8Scsi_Xpt16processCallbacksEv+0x3b9
0x3613fd0: 0x751f46 _ZN18Scsi_serverInitial5startEv+0x6
Fix
Modified the logic used to select paths on Symmetrix systems to ensure that an available path is
chosen.
Service Request
16025389
Fixed in version
5.5.27.0
Impact Level
Severity 1
Symptom
Description
Due to corruption in a file system's ACL database, new files and directories could not be created and
errors were reported indicating that the database was full. UFS: 3: Filesystem x ACL
slotIndex goes beyond 1M. The ACL database for a file system can support up to 1
million unique ACLs. Depending on the corruption, the database appeared to be full when it was not.
Fix
The fix is to distinguish real ACL database full and the seemingly full caused by corruption. When
corruption in the ACL database is detected, the file system is umounted for fsck/aclchk.
2006-05-31 12:10:54: UFS: 3: Filesystem x ACL firstfree
xxxcorrupted beyond 1M
2006-05-31 12:10:54: UFS: 3: Unmounting fs x: Reason:
xxxallocAcl: Corrupt firstfree - go over upper boundary
2006-05-31 12:10:54: FSTOOLS: 2: 0: FsId: x Fsck Started
xxx(Auto - Corruption (allocAcl: Corrupt firstfree - go over
xxxupper boundary)
Service Request
14551178
Fixed in version
5.5.27.0
321
Version 5.5.27.5
UFS
322
Impact Level
Severity 1
Symptom
Description
When a file system is being mounted, the super block is checked to determine if the file system has
been marked corrupt. If it has, FSCK will be run before completing the mount. If the file system was
being mounted read-only, FSCK erroneously modified the file system. The Data Mover panicked with
the following DART panic/fault message: >>PANIC in file: ../sched.cxx at
line: 721 :
SYSTEM WATCHDOG The stack trace showed:
0x7ff54: 0x14b36d _waitForReboot+0x7d
0x7ff64: 0x14b56e _fault_dump+0x56
0x7ff78: 0x14b460 _PANIC+0x24
0x7ff84: 0x14d67c _system_watchdog+0x88
0x7ff98: 0x138e62 _t_nmi+0x62
0x7ffe4: 0x14f0ed _Sthread_startThread_internal+0x11
Fix
Instead of attempting fsck, an attempt to mount a corrupt read-only file system will now fail and an
error is logged in the server_log. The user will then need to run FSCK manually against the file
system.
Service Request
16616362
Fixed in version
5.5.27.0
Impact Level
Severity 1
Symptom
Description
While deleting a significant number of large files in a directory of a write intensive environment where
there were multiple quota trees and the file system was nearly full, the Data Mover entered a deadlock
condition or panicked with the following DART panic/fault message: >>PANIC in file:
../ufsvfsop.cxx at line: 374 : sync: no progress reducing
dirty list. The stack trace showed:
0xe5573f40: 0x13d1db waitForReboot+0x8b
0xe5573f60: 0x13d3e3 fault_dump+0x67
0xe5573f80: 0x13d2e3 PANIC+0x2b
0xe5573f90: 0x6549f3 _ZN21UFS_SyncThreadInitial5startEv+0x13d
0xe5573fe0: 0x1406f4 Sthread_startThread_internal+0xe
Fix
Corrected a locking and initialization issue where a Data Mover panicked when deleting a lot of large
files in a nearly full file system.
Service Request
16771086
Fixed in version
5.5.27.0
Version 5.5.26.4
Impact Level
Severity 1
Symptom
Description
From a CIFS client, user changed a directory (folder) to a checkpoint directory, and listed the contents
of the directory with a wildcard (dir foo*). The Data Mover did not return the list of files. The process
hung in an infinite loop.
Fix
Code has been fixed to ensure command now returns the matching names list which keeps the Data
Mover from getting into this hung state.
Service Request
17460578
Fixed in version
5.5.27.2
Version 5.5.26.4
Celerra
Management
Applications

Impact Level
Severity 2
Symptom
Description
Internet Explorer 7.0 browser crashed on Celerra Manager login page.
Fix
Fixed the login page HTML to avoid the IE7 crash. Added a warning message that IE7 is not a fully
supported browser, but allow use of the browser. Celerra Manager now sets a cookie when user
chooses to continue with IE7 and no longer displays the popup when the cookie is set.
Fixed in version
5.5.26.0
Impact Level
Severity 2
Symptom
Description
User was unable to delete a checkpoint schedule. The schedule was deleted but then was
immediately re-generated by the Control Station.
Fix
Schedule names containing a "." experienced this problem. Code now supports "." in schedule name.
Service Request
16109685
Fixed in version
5.5.26.0
323
Version 5.5.26.4
324
Impact Level
Severity 2
Symptom
Description
While using Celerra Manager, an administrative user that was not the nasadmin or root account
attempting to load the schedules page and pause the schedules saw the following error:
Pause schedule <schedule name>. Unable to send message to the
scheduler.
The following error was displayed in the server log:
2004-11-23 11:44:06.216 Pause schedule <schedule name>
(complete=true, failed=true) connect: Permission denied.
Fix
All users in the nasadmin group are now able to administer (pause/resume) schedules. Previously,
only the nasadmin or root user could perform these actions.
Fixed in version
5.5.26.0
Impact Level
Severity 3
Symptom
Description
When a UPS battery fully charges, Celerra Manager should display an informational message
indicating that the battery is 100% charged. This message was displayed as an error instead of an
informational message.
Fix
Code has been modified to return informational message instead of an error if UPS is 100% charged.
Service Request
15680802
Fixed in version
5.5.26.0
Impact Level
Severity 3
Symptom
Description
While using GUI to change the severity of an Event Notification, if "Apply" button is pressed, the
refreshed properties page does not display the new severity value.
Fix
The Event Notification -> Properties page now selects the proper severity value in the severity
selector.
Fixed in version
5.5.26.0
Version 5.5.26.4
CIFS

problems:
Impact Level
Severity 1
Symptom
Description
User migrated data from Windows box to NAS box using Robocopy. After migration, each directory
had correct permissions (including inherit properties). When clients created new directory and
modified permissions, the changes were not propagated to its subfolders correctly.
Fix
The bit AUTO_INHERITED_SACL is now removed from the default Security Descriptor which is
created on the root of a new share.
Service Request
16017292
Fixed in version
5.5.26.0
Impact Level
Severity 1
Symptom
Description
../addrspac.cxx at line: 431 : Out of memory. The stack trace showed:
0xcec0ba10: 0x13b8c3 waitForReboot+0x8b
0xcec0ba30: 0x13bac2 fault_dump+0x5e
0xcec0ba50: 0x13b9cb PANIC+0x2b
0xcec0ba60: 0x137c1e
0xcec0bac0: 0x17b7d2 _Z11more_memoryiP9free_list+0x4e
0xcec0bae0: 0x17b87f malloc+0x89
0xcec0bb10: 0x17bf91 _Znwm+0x15
0xcec0bb20: 0x17bfbe _Znam+0xc
Fix
This panic occurred if requests contained invalid formatted ACLs. Added defensive code to reject
these malformed requests.
Service Request
16463752
Fixed in version
5.5.26.0
325
Version 5.5.26.4
Impact Level
Severity 1
Symptom
Description
After upgrading from NAS 5.3, CIFS users could not access files pointed to by symbolic links. The
symbolic links appeared as CIFS directories and could not be opened.This only occurred with
symbolic links that were not owned by the root user. User requested that the 5.3 behavior be restored.
Fix
As a security enhancement, NAS versions 5.4 and above require symbolic links to be created by root.
UNIX symbolic link files can point to files on file systems that a user may not have been given visibility
to. Users cannot access these files and directories without proper access rights, but could allow CIFS
users to see files that a Celerra Administrator did not expect them to be able to see.
To allow the option of user-created symbolic links, however, the meaning of the
shadow.followabsolutpath parameter was modified as follows: When set, this parameter allows
users to follow the absolute pathname in symbolic link. There are no restrictions on access, in other
words, users can easily go outside the share and also in another file system. When users go outside of
the share, the security set on the original share to which the user is connected applies. Refer to the
Configuring CIFS on Celerra for a Multi-Protocol Environment technical module or the System
Parameters Guide for more information about this parameter.
Service Request
16048692
Fixed in version
5.5.26.0
Control Station
326

Control Station:
Impact Level
Severity 1
Symptom
Description
setup_clariion wrongly identified Low Cost Fibre Channel (LCFC) disks as Fibre Channel. This
allowed storage templates to be used with these disks that are not supported on LCFC.
Fix
Code now correctly identifies all LCFC part numbers as LCFC disks.
Service Request
16751218
Fixed in version
5.5.26.2
Version 5.5.26.4
Impact Level
Severity 1
Symptom
Description
After failback of a Data Mover on a CFS-14, the file system was left mounted on the original standby
Data Mover. This happened when Data Mover failover occurred and LUN 0 was trespassed. The NAS
database was not updated to reflect a change in file system ownership from the standby to the primary
Data Mover. After the standby was rebooted, it remounted the file system.
Fix
A check has been added in the Data Mover failover code to ensure that the NAS database can be
successfully written to before failover is started. If it cannot be updated, the failover process will abort
and return the following error:
Error 10256: server_2 : Unable to build boot configuration
file at NAS_DB/dos/slot_2 for server_2
Fixed in version
5.5.26.3
CORE

Data Movers:
Impact Level
Severity 1
Symptom
Description
When file system was corrupted, listing snap of promoted iSCSI LUNs resulted in a Data Mover panic
with the following DART panic/fault message: >>PANIC in file: ../malloc.cxx at line: 327 : Memory
corrupted. The stack trace showed: 0xf53cfde0: 0x13d05b waitForReboot+0x8b 0xf53cfe00:
0x13d263 fault_dump+0x67 0xf53cfe20: 0x13d163 PANIC+0x2b 0xf53cfe30: 0x17f0a7 free+0x69
0xf53cfe60: 0xccf27c _ZN14XmlElementNodeD1Ev+0x4e 0xf53cfe80: 0xccf24b
_ZN14XmlElementNodeD1Ev+0x1d 0xf53cfea0: 0xccf6ce _ZN9XmlParserD0Ev+0x58
0xf53cfec0: 0xc9f2b0 _ZN16BlkSrvReqHandler18_handleSnapRequestEP16B
Fix
Fixed a double free condition which caused memory corruption.
Service Request
16070227
Fixed in version
5.5.26.1
327
Version 5.5.26.4
Impact Level
Severity 1
Symptom
Description
At boot time, when statd was started, the Data Mover panicked with the following DART panic/fault message:
>>PANIC in file: ../sched.cxx at line: 822 : SYSTEM WATCHDOG .
0xf0b19724: 0x13d1db waitForReboot+0x8b
0xf0b19744: 0x13d3e3 fault_dump+0x67
0xf0b19764: 0x13d2e3 PANIC+0x2b
0xf0b19774: 0x13ee08 system_watchdog+0x7c
0xf0b19784: 0x137279 _t_nmi+0x63
0xf0b197c0: 0x184c15 _ZN12Stdio_Screen6vprintEPKcPPc+0x9
0xf0b197e0: 0x18429a _ZN18Stdio_AbstractFile5printEPKcz+0x22
0xf0b19800: 0x1b7c8d
xx_ZN14logObject_File7logvmsgE11logFacility11log
Fix
Corrected a problem where error code lookup procedure accessed un-initialized memory.
Fixed in version
5.5.26.0
FileMover
328

FileMover:
Impact Level
Severity 1
Symptom
Description
While using FileMover, recalled files were up to 32k larger than they should have been. This only
occurred if the secondary server returned extra blocks to the Celerra during a read request.
Fix
This problem occurred if the secondary server returned more data than it should have. Preventative
code was added to recall the file data and only use up to the actual size of the offline file, not the size
returned by the secondary server. Any extra bytes received from the secondary server will be
discarded.
Service Request
16421328
Fixed in version
5.5.26.0
Version 5.5.26.4
Impact Level
Severity 2
Symptom
Description
User was unable to archive certain files while running policy engine servers to set files offline. In this
instance, the policy engine server logged a large number of 'stale handle' errors or negative numbers
in DHSM messages in the server log. For example,
DHSM: 6: Received set_offline_attrs for <4294967329 - 1743 - -1825141156>
Fix
Inodes that have a 32 bit generation count or inode number are now read correctly.
Service Request
16644162
Fixed in version
5.5.26.0
Impact Level
Severity 1
Symptom
Description
Processing of error codes from NFS secondary servers resulted in a memory leak for each error. This
memory leak on error code paths eventually led to a Data Mover panic with the following DART
>PANIC in file: ../addrspac.cxx at line: 431 : Out of memory.
0xdad35610: 0x13b7d3 waitForReboot+0x8b
0xdad35630: 0x13b9d2 fault_dump+0x5e
0xdad35650: 0x13b8db PANIC+0x2b
0xdad35660: 0x137b4e
0xdad356c0: 0x174bd6 _Z11more_memoryiP9free_list+0x4e
0xdad356e0: 0x174c83 malloc+0x89
0xdad35710: 0x35bff8
xxxxdr_bytes+0x62
0xdad35730: 0xac598a
xxx_Z14xdr_READ3resokP4XDR_P10READ3resok+0x70
Fix
Code was fixed to free the allocated memory on these error paths to prevent the Data Mover panics.
Service Request
16046059
Fixed in version
5.5.26.0
329
Version 5.5.26.4
Impact Level
Severity 2
Symptom
Description
FileMover offline file recalls hung on write operations if quota limits were reached/exceeded in a quota
enabled file-system. This could cause CIFS or NFS threads to hang and eventually cause a system
outage.
Fix
Fixed file recall logic to properly handle quota.
Fixed in version
5.5.26.0
HighRoad (MPFS)

File System (MPFS):
Impact Level
Severity 2
Symptom
Description
After running FSCK on an MPFS enabled file system, a file was not accessible through any protocol -MPFS, NFS, or CIFS.
Fix
Code now handles extensive MPFS sharing conflicts without any hangs.
Fixed in version
5.5.26.0
Install/Config/
Upgrades
330

Impact Level
Severity 1
Symptom
Description
While upgrading to a dual Control Station, VLAN tags & MAC address information was deleted from
the Data Mover interface configuration. This caused the Data Mover to be inaccessible. This bug
ONLY affected users who were using 802.1Q VLAN tagging.
Fix
Code now deletes only the secondary internal MAC address from the ifconfig entry to ensure the Data
Mover remains accessible.
Service Request
16096396
Fixed in version
5.5.26.0
Version 5.5.26.4
iSCSI

Impact Level
Severity 1
Symptom
Description
If a new replication session was configured while a Replicator reverse or failover operation was
running, the Data Mover panicked with the following DART panic/fault message: >>PANIC in
file: ../assert.c at line: 23 :
assert failed. The stack trace
showed:
0xe5b6fce0: 0x144b37 waitForReboot+0x8b
0xe5b6fd00: 0x144d3f fault_dump+0x67
0xe5b6fd20: 0x144c3f PANIC+0x2b
0xe5b6fd30: 0x1ae893 _assert+0x53
0xe5b6fd50: 0xf29a85
xxx_ZN8DpPolicy14registerPolicyEPc12PolicyType_eS
0xe5b6fd90: 0xefc3df
xxx_ZN22DpService_ReplicaStart11completeCmdEP9DpR
0xe5b6fec0: 0xeee108 _ZN9DpRequest4doneEv+0x84
0xe5b6fef0: 0xf3eabf
xxx_ZN13DP_RepPrimary16setDpRequestDoneE17DpReque
Fix
Code now rejects the start of replication while copy reverse is running.
Fixed in version
5.5.26.0
NDMP/Backup

Impact Level
Severity 2
Symptom
Description
All checkpoint refresh tasks failed on the Data Mover with the following error in the server_log:
"SVFS: 3: max number (5) of ckpt refresh/create exceeded".
The checkpoint refresh command returned
"Error: 5005 failed to complete command".
Fix
This occurred because of an NDMP backup thread conflict with the checkpoint process. Fixed the
backup thread hang issue so that it releases the references to the checkpoint.
Service Request
16133820
Fixed in version
5.5.26.0
331
Version 5.5.26.4
Impact Level
Severity 1
Symptom
Description
While performing NDMP backup, the Data Mover panicked with the following DART panic/fault
message:
>>PANIC in file: ../addrspac.cxx at line: 431 :
Out of memory. The stack trace showed:
0x24281bc0: 0x13b7ef waitForReboot+0x8b
0x24281be0: 0x13b9ee fault_dump+0x5e
0x24281c00: 0x13b8f7 PANIC+0x2b
0x24281c10: 0x137b6a
0x24281c70: 0x174c0e _Z11more_memoryiP9free_list+0x4e
0x24281c90: 0x174cbb malloc+0x89
0x24281cc0: 0x1753cd _Znwm+0x15
0x24281cd0: 0x1753fa _Znam+0xc
Fix
Code now uses safe memory allocation for named file stream to prevent this panic.
Service Request
16159186
Fixed in version
5.5.26.0
Impact Level
Severity 2
Symptom
Description
Pre- version 5.5 backup will restore to NAS version 5.5 as non-DAR causing slow performance if DAR
is wanted for file level restore. The following warning message shows up in the server log:
DDAR is not available for archive backed up before Celerra
5.5, Use non-DAR to continue the restore.
Fix
Prior to 5.5, RECURSIVE DAR restores were not supported. In 5.5, only RECURSIVE DAR restores
were supported. Since pre 5.5 backups don't support RECURSIVE DAR restore, DAR restores of pre
5.5 backup sets failed and a non DAR restore was performed.
This was fixed by re-implementing the non RECURSIVE DAR restore and make it the default. A
RECURSIVE DAR restore will be done only if the backup software sets the RECURSIVE variable to
yes.
NOTE: If you set the RECURSIVE variable to "yes" for a pre 5.5 restore, the DAR
will fail and a non-DAR restore will be performed.
332
Service Request
16008198
Fixed in version
5.5.26.0
Version 5.5.26.4
Impact Level
Severity 2
Symptom
Description
The following integrated checkpoint backup normal message displays as critical error on GUI:
2006-09-04 18:06:56: NDMP: 0: 2: fsid:382
mount_point:/root_vdm_2/automaticNDMPCkpts/automaticTempNDMPCkpt11-52-1157385928. This
issue did not impact any backup function.
Fix
Changed the log level to warning from emergency. It will no longer display as critical error.
Service Request
15882935
Fixed in version
5.5.26.0
NFS
Impact Level
Severity 1
Symptom
Description
When using NFSv4 under stress conditions, the Data Mover panicked with the following DART
panic/fault message: >>PANIC in file: ../sched.cxx at line: 822 :
SYSTEM WATCHDOG.
0xcc6d3594: 0x144bb7 waitForReboot+0x8b
0xcc6d35b4: 0x144dbf fault_dump+0x67
0xcc6d35d4: 0x144cbf PANIC+0x2b
0xcc6d35e4: 0x1473e3 system_watchdog+0x81
0xcc6d35f4: 0x13a799 _t_nmi+0x63
0xcc6d3650: 0x66a4d2
xxx_ZN10Delegation5matchEP15NFSv4_ClientEntP8stat
0xcc6d36a0: 0x660c6d
xxx_ZN11NFSv4_State19addShareReservationEmmP8nfs4
0xcc6d3750: 0x655eec
xxx_ZN11NFSv4_Owner4openEmmP8nfs4_ctxP8stateid4RP
Fix
Modified the referencing of delegation objects to prevent the reuse of a freed object.
Fixed in version
5.5.26.0
333
Version 5.5.26.4
Replication

Impact Level
Severity 1
Symptom
Description
When RW mounted UxFS file-system were converted to RawFS, the file system went offline. Also, RO
mounted Destination FS in a Replicator pair can be converted to RawFS which stops Playback which
can lead to replication becoming inactive. Neither of these operations should be run normally, and
would only be performed in error.
Fix
Code now prevents users from converting actively-replicated file systems to RAW type and now
reports the error in the command output: Error 10249: Conversion from Read-Write mounted UxFS to
RawFS or destination file-system in replication to RawFS is not permitted.
Service Request
14378836
Fixed in version
5.5.26.0
Impact Level
Severity 1
Symptom
Description
During local replication start operation, if playback initialization failed, changes made to the NAS
database were not reverted correctly This caused file system related commands and operations to fail.
Fix
Code now removes related references and does the cleanup correctly.
Fixed in version
5.5.26.1
UFS
334
Impact Level
Severity 1
Symptom
Description
When a large file was deleted on a busy system, a deadlock condition occurred which led to a Data
Mover panic with the following DART panic/fault message: watchdog: no progress
reducing dirty list
Fix
Code now prevents a thread requesting another update descriptor when calling asyncdelete.
Fixed in version
5.5.26.0
Version 5.5.25.2
Impact Level
Severity 1
Symptom
Description
../sched.cxx at line: 721 : SYSTEM WATCHDOG.
0xb2cf1708: 0x14b36d _waitForReboot+0x7d
0xb2cf1718: 0x14b56e _fault_dump+0x56
0xb2cf172c: 0x14b460 _PANIC+0x24
0xb2cf1738: 0x14d67c _system_watchdog+0x88
0xb2cf174c: 0x138e62 _t_nmi+0x62
0xb2cf1784: 0x6f3a89
xxx_lookupNameLocked__26passthroughShadowDirector
0xb2cf17e0: 0x6f421e
xxx_lookupDotLockedForAttr__26passthroughShadowDi
0xb2cf18d4: 0x6f4eea
xxx_lookupNameLockedForAttr__26passthroughShadowD
Fix
This problem happened in rare cases when files were added to directories immediately after the
directory was created. Code now adjusts reference count for these directories appropriately.
Service Request
16141233
Fixed in version
5.5.26.0
Version 5.5.25.2
Celerra
Management
Applications

Impact Level
Severity 2
Symptom
Description
If the Celerra Manager was not responding to client requests, and a user tried to display the file
system properties page, the browser hung.
Fix
Under this condition, the Celerra Manager now displays the page properly and the browser no longer
hangs.
Fixed in version
5.5.25.0
335
Version 5.5.25.2
CIFS
336

problems:
Impact Level
Severity 2
Symptom
Description
While using the CLI, if quota was set to OFF, file system quota configuration settings were not getting
reflected in the tree quota config.
Fix
File system quota configuration settings are now reflected in tree quota config irrespective of quota
ON/OFF.
Fixed in version
5.5.25.0
Impact Level
Severity 1
Symptom
Description
After a Data Mover reboot, widelinks configured on a VDM did not work until the CIFS service was
restarted. This happened because when the CIFS service was started during Data Mover bootup, the
VDM used for the widelink share had not been configured yet.
Fix
During bootup the widelink registry setting is now checked for each CIFS server instead of only when
starting the CIFS service. Widelinks are now started once the server hosting the share has been
configured.
Service Request #
15925355
Fixed in version
5.5.25.0
Impact Level
Severity 1
Symptom
Description
After a Data Mover reboot, widelinks configured on a VDM did not work until the CIFS service was
restarted. This happened because when the CIFS service was started during Data Mover bootup, the
VDM used for the widelink share had not been configured yet.
Fix
During bootup the widelink registry setting is now checked for each CIFS server instead of only when
starting the CIFS service. Widelinks are now started once the server hosting the share has been
configured.
Service Request
15925355
Fixed in version
5.5.25.0
Version 5.5.25.2
Impact Level
Severity 1
Symptom
Description
After upgrading to version 5.5.23 or later, NT4 and Win9x clients intermittently saw CIFS-related
errors. This occurred most frequently during write operations, and clients logged a "Lost
Delayed-Write Data" failure when this occurred.
The server logs showed errors similar to:
SMB: 3: checkSignatureS stopped by WinNT Client=DC-TSF1-S10
Cmd=0x2f Mid=a000 ctxCount=2 ActiveTime=2852
checkSignature(S): comparison failed CMD=73 MID=1000 Len=0xd4
ExpectedSeq=0x0 KeyTime=0x453e1e47 #fail=0x26/0x0
(0x114f26b7)
Fix
In versions prior to 5.5.23, SMB signing was disabled by default unless enabled by an administrator via
a registry entry or GPO setting. In versions 5.5.23-5.5.24, SMB signing was enabled by default, and
systems that normally wouldn't have been using signing ended up using it.
The root problem of SMB signing failures with these clients is not known, but signing is off by default
again. Additional information is collected to help determine why these failures are occurring.
Service Request
16284402
Fixed in version
5.5.25.0
Impact Level
Severity 2
Symptom
Description
The first attempt to get a user's mapping from the LDAP server (iPlanet) failed, however, the next
attempt succeeded. This issue happened when it was the first time the user connected (meaning the
mapping hadn't been cached) and the LDAP server had disconnected the connection to the Celerra
due to a period of inactivity.
Fix
Depending on configuration, the iPlanet server may disconnect the LDAP connection after a specified
idle time. Now, Celerra will attempt to reconnect to the LDAP server and then request the mapping.
Service Request
15802192
Fixed in version
5.5.25.0
337
Version 5.5.25.2
338
Impact Level
Severity 1
Symptom
Description
User had a very large number of Group Policies associated with a single Organizational Unit in a
Windows 2000/2003 domain, one of the LDAP search requests (to the domain controller) returned a
very large response, which required a large amount of memory to process. If the Data Mover didn't
have enough memory available to handle the response, it panicked with the following DART panic/fault
message: >>PANIC in file: ../addrspac.cxx at line: 431 : Out of
memory.
0x21f15a44: 0x14acad _waitForReboot+0x7d
0x21f15a54: 0x14aeae _fault_dump+0x56
0x21f15a68: 0x14ada0 _PANIC+0x24
0x21f15a74: 0x145ffd
xxxx_allocPages__19Memory_AddressSpaceiG12Memory_O
0x21f15ac4: 0x1983f9 _more_memory__FiP9free_list+0x51
0x21f15ae8: 0x19855b _malloc+0x73
0x21f15b04: 0x19902e ___builtin_new+0x12
0x21f15b14: 0x19905b ___builtin_vec_new+0x7
Fix
Code was changed to allocate memory by pages (if more than 1 page of memory is needed). The
operation (e.g., GPO update) will fail and log a message in the server log if there isn't enough memory
available, but the Data Mover will no longer panic.
Service Request
15244634
Fixed in version
5.5.25.0
Impact Level
Severity 1
Symptom
Description
While writing to a file system under heavy load, the server experienced high CPU utilization while
flushing all open files. Another instance was a deadlock condition when using CIFS autodisconnect
feature/parameter and users accessed a directory with a delayed SID.
Fix
Changed CIFS TCP connection list management to allow adding/removing items while browse
operations are in-progress.
Service Request
15043165
Fixed in version
5.5.25.0
Version 5.5.25.2
Impact Level
Severity 2
Symptom
Description
If client opened a new CIFS connection while setting VCNUM=0 in the request, Celerra did not close
files for the existing open connections for that client.
Fix
Files are now closed on existing connections in this case. This now mirrors the Windows reference
behavior.
Fixed in version
5.5.25.0
Impact Level
Severity 1
Symptom
Description
While using GASP management software with repository on Celerra, thousands of CIFS clients
simultaneously tried to do a wildcard search on the same large directory on a Data Mover. This caused
CPU to rise up to 100% and slow performance when accessing data using CIFS backing up using
NDMP.
Fix
The filtering mechanism was implemented to improve the search performance when doing wildcard
searches on large directories from CIFS applications.
Service Request
16078910
Fixed in version
5.5.25.0
Control Station

Control Station:
Impact Level
Severity 1
Symptom
Description
After setting up WINS information per interface and rebooting or upgrading a Data Mover, WINS
information was lost. When WINS was configured, this information wasn't saved in the NAS DB. This
caused the CIFS user to lose access to the CIFS server.
Fix
Existing code was failing when a new WINS was added for the first time on an existing interface. Code
now updates the netd file appropriately when adding WINS information to existing interfaces.
Fixed in version
5.5.25.0
339
Version 5.5.25.2
340
Impact Level
Severity 2
Symptom
Description
In CFS/CNS 14 cabinets with dual Control Stations, BoxMonitor errors (battery errors, sense cable
errors, terminator errors, power errors, etc) were continuously logged when both Control Stations were
inserted. Each error generated a callhome. These messages do not impact the functionality of either
Control Station or impact failover between the Control Stations.
BoxMonitor in NAS 5.5 was modified to run a getreason command from both Control Stations. This is
not an issue for NS based systems, but does result in these errors being erroneously logged on a
CNS/CFS system.
Fix
Code has been fixed to ensure these errors are no longer erroneously logged and a callhome is no
longer generated.
Service Request
15615902
Fixed in version
5.5.25.0
Impact Level
Severity 2
Symptom
Description
During a file system extension of a replicated file system, if the process cannot acquire the DB locks
while attempting to extend the local file system user saw a stack trace similar to: 2006-03-03
02:27:10.533 db:201:13296:E: nas_fs -x server2_fs1 size=1 -o
slice=yes: Execution failed: Segmentation violation:
Operating system signal. [NAS_SIGHN DLR.eiffel_handlr]
Fix
The error scenario is now handled properly and the user will see the proper error message Unable
to acquire locks instead of a stack trace.
Fixed in version
5.5.25.0
Impact Level
Severity 2
Symptom
Description
During nasdb_backup, the standby Data Mover server log repeated the following messages:
SMB: 3: Could not access share db
Fix
Now the nasdb_backup script will attempt to retrieve share DB information only for NAS type servers
and not for standby servers.
Service Request
15701599
Fixed in version
5.5.25.0
Version 5.5.25.2
Impact Level
Severity 1
Symptom
Description
Control Station intermittently failed over due to an out of memory condition. The logs showed lots of
'out of memory: Killed process xxxx'. Examining the processes running on the
Control Station showed a number of instances of the server_cifs -o audit commands that
did not exit.
Fix
Code now detects if a server_cifs -o audit command has been running for more than 2
hours. If it finds one that has, it issues a kill -9 on all occurrences of this command.
Fixed in version
5.5.25.0
Impact Level
Severity 2
Symptom
Description
When the Quota Block grace period is set to -1 (meaning unlimited) using nas_quotas command,
both the "Grace Period for Storage Soft Limits" and "Grace Period for File Count Soft Limits" boxes in
the WEBUI were blank. Only the WEBUI display was incorrect, the actual Inode grace period had not
changed and appeared correctly when viewed from the command line.
Fix
When the Block grace period is set to -1 using the nas_quotas command, the WEBUI now shows
the correct setting for Inode grace period.
Fixed in version
5.5.25.0
Impact Level
Severity 2
Symptom
Description
Celerra Manager allowed users to rename or delete a network interface that was being used by the
CIFS server on the VDM. The renamed interface didn't show in the Celerra Manager properties page.
Fix
Before an interface can be renamed, VDMs are now scanned to ensure the interface is not in use. If
the interface is in use, an error message is displayed:
CLI:
2216: <if_name>: is currently in use by CIFS server:
<CIFS_server_name>
GUI:
This interface is already in use by another CIFS server. Click
OK to use it for the new server instead (the existing server
will no longer be accessible on this interface.
Service Request
14167645
Fixed in version
5.5.25.0
341
Version 5.5.25.2
342
Impact Level
Severity 1
Symptom
Description
IP configurations were not transferred to the standby Data Mover during a failover. This was the result
of a setup_slot being manually run against the primary Data Mover during an attempt to correct a
previous issue with file systems mounting on the Data Mover.
Fix
The fault detection and exit code paths in the setup_slot command have been fixed to prevent leaving
the ifconfig files in this state.
Service Request
15946718
Fixed in version
5.5.25.0
Impact Level
Severity 2
Symptom
Description
If a file system was being replicated to two or more remote systems, commands specific to file
systems or replications hung.
Fix
To confirm that a remote file system is a backup of a local file system, code was building a list of all
backups. This created an infinite loop when the local file system had two remote backups on different
Celerras. This was changed to send a request to the remote Control Station to return the remote index
of the local Control Station as seen remotely. With this index, code now builds a search string
"FsId@CsId" to look for in the list of backups of the local file system. This fix eliminates the infinite loop
and the commands no longer hang.
Fixed in version
5.5.25.0
Impact Level
Severity 2
Symptom
Description
User reported problems deleting unused interface with the error "item is currently in
use by Replication". This problem occurred if multiple interfaces were configured with IP
addresses that shared the first 3 octets and first digit of the 4th octet.
Fix
Code now checks for the full IP address instead of a subset to prevent this condition.
Service Request
15342903
Fixed in version
5.5.25.0
Version 5.5.25.2
CORE

Data Movers:
Impact Level
Severity 2
Symptom
Description
The message /.etc/group does not exists and NIS not started was
logged multiple times in the server_log when NIS was not started and a local group file was not
present.
Fix
This informational message is only logged in the server_log when it is first discovered that neither NIS
or the local group file is available.
Service Request
15937867
Fixed in version
5.5.25.0
Impact Level
Severity 1
Symptom
Description
When doing a checkpoint restore, if the savVol was full, the Data Mover panicked with the following
ae6b65 Err code: 0 Target addr: eeeeeef6 ** .
0x1d003c4: 0x145867 waitForReboot+0x8b
0x1d003e4: 0x145a6f fault_dump+0x67
0x1d00404: 0x13b74a handle_fatal_fault+0xa
0x1d00440: 0xae720a _ZN6VJTreeD0Ev+0x64
0x590bf00: 0x14b426 _ZN7Sthread6detachEv+0x104
0x590bf30: 0x14b7bc _Sthread_suicide_internal+0xaa
0x590bf50: 0x14be7c _ZN7Sthread7suicideEv+0x6b4
0x562ddfe0: 0x14ae28 Sthread_startThread_internal+0x1a
Fix
The panic occurred because a linked list head was not being initialized properly. This value is now
initialized to NULL.
Fixed in version
5.5.25.0
343
Version 5.5.25.2
344
Impact Level
Severity 1
Symptom
Description
While creating a new tree quota, the Data Mover panicked with the following DART panic/fault
message: *** Page Fault Interrupt. Virt ADDRESS: a285ea Err
code: 0 Target addr: fe9884b8 ** . The stack trace showed:
0x1651ff8: 0x13d10f waitForReboot+0x8b
0x1652018: 0x13d317 fault_dump+0x67
0x1652080: 0x68b210
xxx_ZN19ShadowDirectoryBase13parsePathnameEPtiP17A
0xd66ddb90: 0x9f3278
xxx_ZN8AppLibNT10getDosAttrEPtR7dosattrR11File_St
0xd66ddbf0: 0x95db05
xxx_ZN13smb_threadCtx23callTrans2QFilePathInfoEP4
0xd66dde70: 0x95643b
xxx_ZN13smb_threadCtx10execTrans2EP4msgbPcS2_+0x1
0xd66ddeb0: 0x957058
xxx_ZN13smb_threadCtx15replyTrans2PrimEP4msgb+0x4
Fix
This issue was caused by code attempting to use a memory address which was already freed. This
has been fixed.
Service Request
16096196
Fixed in version
5.5.25.1
Impact Level
Severity 1
Symptom
Description
During the boot process, the Data Mover panicked with the following DART panic/fault message:
>>PANIC in file: ../sched.cxx at line: 822 : SYSTEM WATCHDOG
0x1474758: 0x144bb7 waitForReboot+0x8b
0x1474778: 0x144dbf fault_dump+0x67
0x1474798: 0x144cbf PANIC+0x2b
0x14747a8: 0x1473e3 system_watchdog+0x81
0x14747b8: 0x13a799 _t_nmi+0x63
0x14747f0: 0x15aec8 shared_isr+0x1a
0x1474800: 0x13adce _irq47+0x20
0x598dce0: 0x99e287 _ZN14ScsiBus_Device4scanEjj+0x13f
Fix
Changed network interface initialization code to avoid triggering multiple interrupts which could
intermittently trigger this panic.
Fixed in version
5.5.25.1
Version 5.5.25.2
Impact Level
Severity 1
Symptom
Description
While unmounting a file system with a DFS root share with DFS link on it, the Data Mover panicked with
the following DART panic/fault message: >>PANIC in file: ../assert.c at line:
23 : assert failed.
0x95de1a70: 0x144caf waitForReboot+0x8b
0x95de1a90: 0x144eb7 fault_dump+0x67
0x95de1ab0: 0x144db7 PANIC+0x2b
0x95de1ac0: 0x1b0243 _assert+0x53
0x95de1ae0: 0x7153ff
xxx_ZN15File_TriggerSet13unlockParentsEi+0x45
0x95de1b10: 0x7142c0
xxx_ZN15File_TriggerSet11unBlockPathE13File_NodeK
0x95de1b20: 0x714417
xxx_ZN15File_TriggerSet12closeDfsLinkEP14VFS_File
0x95de1b50: 0xafa2c9 _ZN10smb_server15unmountInternalEy+0x1e5
Fix
Modified code to avoid overlapping locks which caused this panic.
Fixed in version
5.5.25.0
FileMover

FileMover:
Impact Level
Severity 2
Symptom
Description
After rebooting the DHSM secondary server, offline files were not accessible and DHSM commands
failed with the following error: Cannot fsstat remote server, status 9 StaleHandle. To reestablish the connection, user needed to modify the connection through the
Celerra Control Station, for example, change the protocol from UDP to TCP or TCP to UDP and then
back. This was specific to NFS connections.
Fix
This problem occurred if NFS file handles changed. Code now detects such stale NFS handles and
disconnects and reconnects to update this information.
Fixed in version
5.5.25.0
345
Version 5.5.25.2
346
Impact Level
Severity 1
Symptom
Description
Users lost access to CIFS shares due to blocked SMB threads which can be seen in following
command: $ server_cifs server_2 .
server_2 : 256 Cifs threads started
0 messages waiting in VC collector
14473 messages waiting in SMB collector
List of blocked threads
SMB003 blocked for 4359 seconds in CreateNTX
SMB006 blocked for 4648 seconds in Close
SMB007 blocked for 4681 seconds in LockingX
SMB008 blocked for 4506 seconds in CreateNTX.
The following server log was observed:
2006-09-11 11:20:12: MGFS: 3:[VDM1] error direct read block
xxx2273 of entry 457996, status 20 - EndOfFile
2006-09-11 11:20:12: DHSM: 4:[VDM1] HTTP: Received a 0 byte
xxxresponse for a read request for 1 or 4 blocks
2006-09-11 11:20:12: MGFS: 3:[VDM1] error direct read block
xxx2273 of entry 457996, status 20 - EndOfFile
Fix
The problem was triggered only when a response that was not compliant with the HTTP protocol was
sent. When such an error was generated the HTTP response object was not freed which resulted in all
the other threads getting blocked. Code now frees response object when the HTTP client returns an
error.
Service Request
15922973
Fixed in version
5.5.25.0
Version 5.5.25.2
Impact Level
Severity 1
Symptom
Description
When using FileMover with release 5.5.22.0 or higher, a race condition caused a Data Mover to panic
with the following DART panic/fault message: >>PANIC in file: ../sched.cxx at
line: 822 : SYSTEM WATCHDOG.
0xeac99be4: 0x13d10f waitForReboot+0x8b
0xeac99c04: 0x13d317 fault_dump+0x67
0xeac99c24: 0x13d217 PANIC+0x2b
0xeac99c34: 0x13ed3c system_watchdog+0x7c
0xeac99c44: 0x1371ad _t_nmi+0x63
0xeac99ca0: 0xa716da _ZN6SmbMsg4sendEi+0x10e
0xeac99cd0: 0xa715a5 _ZN6SmbMsg7processEv+0x17
0xeac99cf0: 0xa83a46 _ZN13SmbRemoteNode5closeEv+0x8a.
User experienced a delay while the Data Mover failed over to the stand by.
Fix
The regression introduced in version 5.5.22.0 has been corrected. Code now opens/closes a CIFS
secondary file for every block I/O in order to avoid panic in UFS sync thread.
Service Request
16103198
Fixed in version
5.5.25.0
Install/Config/
Upgrades

Impact Level
Severity 2
Symptom
Description
On an integrated system, all non-critical CLARiiON events were not generating dialhomes.
Fix
The system now successfully generates all CLARiiON dialhome events.
Service Request
15206909
Fixed in version
5.5.25.0
347
Version 5.5.25.2
Impact Level
Severity 1
Symptom
Description
Trespassed LUNs caused the upgrade script to fail. The script completed with errors, but the system
was up and running, (it took a very long time for the script to run the setup_slot).
Fix
Recovery mechanism was changed to recover much faster now and upgrades will no longer fail due to
this condition.
Service Request
14767556
Fixed in version
5.5.25.0
iSCSI
Impact Level
Severity 2
Symptom
Description
While using iSCSI with a Solaris initiator, the server log was filled with the following error:
2006-08-03 01:22:41: VCS: 3:
fs70_T1_LUN0_APM00060805258_0000: uncachedWrite failed:
LastStatus 2006-08-03 01:22:41: VLU: 3:
VirtualLU_Disk::write(0x0): failed st=11 wrote 0x0
Fix
The protocol allows a transfer length of 0, which we were not handling correctly. We fixed the handling
of read/write with transfer length of 0.
Service Request
15691351
Fixed in version
5.5.25.0
NDMP/Backup
348


Version 5.5.25.2
Impact Level
Severity 1
Symptom
Description
If DMA sends NDMP environment variable RECURSIVE=n at restore time for a pre-Celerra 5.5
backup, any directory specified in the recover name list will only have itself restored. This is the
designed behavior in Celerra 5.5, but RECURSIVE=n may not be the desired behavior.
Fix
If DMA sends RECURSIVE=n and DIRECT=n (or not set, but DIRECT is default to no) at restore time,
any directory will only get itself restored. If for some reason, it's not the desired behavior, set param
NDMP.forceRecursiveForNonDAR on the Control Station to force a recursive non-DAR restore.
Service Request
15979230
Fixed in version
5.5.25.1
Impact Level
Severity 1
Symptom
Description
While doing NDMP backup/restore in a Windows 2003 environment, the Data Mover panicked with the
following DART panic/fault message: >>PANIC in file: ../sched.cxx at line:
722 :
0xa87c98c4: 0x13b7ef waitForReboot+0x8b
0xa87c98e4: 0x13b9ee fault_dump+0x5e
0xa87c9904: 0x13b8f7 PANIC+0x2b
0xa87c9914: 0x13d2de system_watchdog+0x7c
0xa87c9924: 0x1358d1 _t_nmi+0x63
0xa87c9970: 0x174c73 malloc+0x89
0xa87c99a0: 0x360eab xdr_string+0x85
0xa87c99d0: 0xb22c46 xdr_ndmp_name_v3v4+0x65
Fix
This panic occurred because of a problem handling large XDR (External Data Representation) strings.
This problem has been fixed.
Service Request
15794421
Fixed in version
5.5.25.0
349
Version 5.5.25.2
350
Impact Level
Severity 1
Symptom
Description
The creation of NDMP automatic checkpoints intermittently failed. The server log would display errors
similar to: NDMP:3:10 fsid 2783 not found. Checkpoints not created automatically
were not affected.
Fix
This problem occurred if the creation of an automatic checkpoint took longer than 5 minutes (the
default timeout). If the checkpoint creation took too long, the creation process was aborted and
cleaned up. Made changes the automatic checkpoint creation process to speed up checkpoint
creation. Also introduced a new parameter, NDMP.snapTimeout, which can be used to adjust the
timeout from 2 to 60 minutes (default of 5 minutes) if needed.
Service Request
16002836
Fixed in version
5.5.25.1
Impact Level
Severity 1
Symptom
Description
During an NDMP backup, a memory request resulted in a Data Mover panic with the following DART
>>PANIC in file: ../addrspac.cxx at line: 275 :
Out of
memory
0x2ef1da00: 0x13d05b waitForReboot+0x8b
0x2ef1da20: 0x13d263 fault_dump+0x67
0x2ef1da40: 0x13d163 PANIC+0x2b
0x2ef1da50: 0x13900c
0x2ef1da70: 0x17ef0a _Z11more_memoryiP9free_list+0x4e
0x2ef1da90: 0x17efb7 malloc+0x89
0x2ef1dac0: 0x39cb44 xdr_bytes+0x62
0x2ef1dae0: 0xbe17d5
xxx_ZN21NdmpdTapeExecuteCdbMH14xdrRequestFuncEP4X
Service Request
15867492
Fixed in version
5.5.25.0
Version 5.5.25.2
Networking
Impact Level
Severity 1
Symptom
Description
While using Data Movers with the TOE (TCP offload engine) feature enabled, if a negotiated MSS
(maximum segment size) value was too small, the Data Mover panicked with the following DART
*** Page Fault Interrupt. Virt ADDRESS: 32c6f2 Err code: 0
Target addr: fa909098 **
0x1292b74: 0x13b8c3 waitForReboot+0x8b
0x1292b94: 0x13bac2 fault_dump+0x5e
0x1292bb4: 0x135c6c handle_fatal_fault+0xa
0x1292c00: 0x32bb73
xxx_Z19atk_receive_flushedP10atk_socketP4msgbmm+0x
0x69e80: 0x31a89d _Z7atk_fsmP10atk_socket9ATK_EVENTmmm+0x10fd
0x69ed0: 0x329aaf
xxx_Z21atk_fast_receive_doneP10atk_socketP4msgbmmm+0
0x69f10: 0x328916 _Z17atk_send_completeP4msgb+0x562
0x69f60: 0x2c67ce _Z18slic_xmit_completeP10_adapter_t+0x34a
Fix
We no longer use TOE for connections with a small MSS (<=192 bytes). There would be no
performance advantages to using TOE with connections using MSS this small.
Service Request
15888950
Fixed in version
5.5.25.0
Impact Level
Severity 2
Symptom
Description
After Solaris clients rebooted (manually or due to a crash), they were unable to re-mount Celerra file
systems for up to 15 minutes. This problem only occurred if using TCP mounts, UDP mounts were not
affected.
Fix
Corrected the issue that affected TCP reconnections if clients did not terminate their existing TCP
connections during shutdown and tried to re-use the same port numbers.
Service Request
15611094
Fixed in version
5.5.25.0
351
Version 5.5.25.2
NFS
Impact Level
Severity 1
Symptom
Description
System hung when NFSv4 was used (NFSv3 and SMB threads can be blocked as well as NFSv4
threads). server_cifs showed "messages waiting in SMB collector".
This problem occurred because of a deadlock between NFSv4 file delegation and file open processes.
Fix
Code has been fixed to ensure that the lock intended to protect a list of NFSv4 states isn't held any
longer than required. This prevents the deadlock condition.
Fixed in version
5.5.25.0
Impact Level
Severity 2
Symptom
Description
A temporary communication failure caused NFS locking to some clients not to work for extended
periods of time.
Fix
Code now attempts to retry lockd and statd callbacks instead of aborting after only one request.
Fixed in version
5.5.25.0
Replication
352

Impact Level
Severity 2
Symptom
Description
When a checkpoint is refreshed a large number of messages are logged in the server log. This can
make the server log fill up quickly when there are many checkpoints in the system.
Fix
The unnecessary or redundant messages involved in a checkpoint refresh are no longer logged.
Service Request
15073167
Fixed in version
5.5.25.0
Version 5.5.25.2
Impact Level
Severity 1
Symptom
Description
On-disk corruption of a checkpoint resulted in a Data Mover panic with the following DART panic/fault
message: >>PANIC in file: ../DeltaVol.cxx at line: 8654 :
updateDouble failure. The stack trace showed:
0xd5fd9dd0: 0x13d05b waitForReboot+0x8b
0xd5fd9df0: 0x13d263 fault_dump+0x67
0xd5fd9e10: 0x13d163 PANIC+0x2b
0xd5fd9e20: 0xaf8daa
xxx_ZN12pmdvMicroLog12updateDoubleEjjjjP13SnapTra
0xd5fd9e90: 0xae98e2
xxx_ZN19PoolManagedDeltaVol17relocateIndexSlotEyy
0xd5fd9f00: 0xaf2354
xxx_ZN19PoolManagedDeltaVol5mergeERP5Event+0x6b2
0xd5fd9f80: 0xb010b2 _ZN11mergeThread5startEv+0x190
0xd5fd9fe0: 0x14038c Sthread_startThread_internal+0xe
Fix
Fixed a bug that caused some checkpoint information not to be flushed to disk immediately, and allow it
to be lost if the system panicked for an unrelated reason. This is a very unlikely bug to run into, and is
somewhat more likely to happen if a system panics shortly after creating a checkpoint.
Service Request
16142059
Fixed in version
5.5.25.0
Impact Level
Severity 1
Symptom
Description
Previously, users could copy or replicate sessions even if the source and destination versions were
different, for example, source on V5.4, destination on V5.5. The database format changes included in
this release caused Control Station communication problems when using different source and
destination versions.
Fix
Code now prevents users from creating a copy or replication session between source and destinations
running different NAS code families (5.4, 5.5). Existing copies or replications will not be affected.
Fixed in version
5.5.25.0
353
Version 5.5.25.2
Impact Level
Severity 1
Symptom
Description
During an fs_copy, if the remote side failed to respond to a network KEEPALIVE packet, the Data
Mover panicked with the following DART panic/fault message:
*** Page Fault Interrupt. Virt ADDRESS: 173016 Err code: 0
Target addr: ff000000 ** .
0x129e134: 0x13b7ef waitForReboot+0x8b
0x129e174: 0x135bb8 handle_fatal_fault+0xa
0x129e1f0: 0x1715b4 _ZN18Stdio_AbstractFile6vprintEPKcPPc+0x1a
0xbd3e9f50: 0x1a07ff
xxx_ZN9logObject11suppressMsgE11logFacility11logS
0xbd3e9f80: 0x1a11fd
xxx_ZN5logIO6logmsgE11logFacility11logSeverityPKc
0xbd3e9fb0: 0xaf7def _ZN17buildBitMapSender5startEv+0x51
0xbd3e9fe0: 0x13eb30 Sthread_startThread_internal+0xe
Fix
An error in the messaging module caused the panic, and has been fixed.
Service Request
15963338
Fixed in version
5.5.25.0
SRDF
354

Impact Level
Severity 1
Symptom
Description
When a Data Mover was failed over locally on the source side of an SRDF link, nas_rdf activate or
restore (failover or failback) commands failed with the following: "Error 4003: standby is
not configured. This affected only NS-based systems. CNS and CFS-based Celerras were
not affected.
Fix
This occurred because NBS (network block services used by the Control Station to access backend
LUNs) settings from the remote system were being used too early in the failover process. Code now
uses local NBS information to retain access to NBS until the remote configuration can be used.
Fixed in version
5.5.25.0
Version 5.5.25.2
UFS
Impact Level
Severity 1
Symptom
Description
User lost access to a file system when the CIFS Security Event Log was on a user file system and this
file system was replicated or used for checkpoint.
Fix
Corrected a deadlock that occurred between the event log writing process and file system pause
process (used with Replicator or Snapsure) which blocked access to the file system.
Service Request
15693612
Fixed in version
5.5.25.0
Impact Level
Severity 2
Symptom
Description
User created a file in the test directory, set the timestamps on the file to an earlier date, and then wrote
into the file. After closing the file the "last write time" was set to the current time rather than the earlier
time that was set before the write. This is different from Windows Server, which performs the same
except that the "last write" timestamp on the file is set to the earlier time that was set before the write.
Fix
Appropriate CIFS algorithm was modified to mirror the Windows platform behavior. If last modification
date was set on the file, Celerra wont update the last modification date in the inode of this operation.
Service Request
15262426
Fixed in version
5.5.25.0
355
Version 5.5.24.2
Version 5.5.24.2
Celerra
Management
Applications
356

Impact Level
Severity 2
Symptom
Description
GPO settings became inaccessible during user management while using the Celerra Manager snap-in
due to a loss of connection between the CIFS server and the domain. Possible causes of this problem
are the CIFS server is no longer joined to the domain as required to communicate to the snap-in, the
snap-in was not connected to the CIFS server via the console, the computer acting as the domain
controller was down, or the network was not operational.
Fix
To help rectify this condition the error message that is displayed has been updated to include areas the
users can check to resolve the issue.The new message reads: Internal Error: Unable
to retrieve GPO settings from server. Please ensure that the
CIFS server is still connected to the domain. Possible
solution would be to rejoin the domain.
Fixed in version
5.5.24.0
Impact Level
Severity 2
Symptom
Description
In Celerra Manager, the Celerra --> Storage --> Systems --> <Clariion>
page Disk Device summary displayed No Items to display on systems with a single
enclosure CX300, CX400, CX500 CLARiiON.
Fix
Corrected regular expression in get backend status polling script.
Service Request
15768384
Fixed in version
5.5.24.0
Version 5.5.24.2
Impact Level
Severity 1
Symptom
Description
After the Tree Quotas list was shown, and the sort column was clicked, Celerra Manager hung and
nothing was displayed in the sort order.
Fix
Code was modified to use the correct return value under this condition and the Data Mover no longer
hangs.
Service Request
15680487
Fixed in version
5.5.24.0
CFS
The following entries describe fixed Celerra File System problems:
Impact Level
Severity 1
Symptom
Description
If a file system name greater than 1024 was used, for example, "nas_quotas -report -fs
<fsname> 1 2 3 .... 1025", the Data Mover panicked with the following DART panic/fault
message: PANIC in file:../malloc.cxx at line:327 : Memory
corrupted.
0xd85a7320: 0x13d047 waitForReboot+0x8b
0xd85a7340: 0x13d24f fault_dump+0x67
0xd85a7360: 0x13d14f PANIC+0x2b
0xd85a7370: 0x17c33b free+0x69
0xd85a73a0: 0x17c954 _ZdlPv+0x10
0xd85a73b0: 0x17c96a _ZdaPv+0xc
0xd85a73c0: 0x161f3f _ZN6OptionD1Ev+0x15
0xd85a73d0: 0x5bc7a6
xxx_Z17File_QuotaControlR6OptionP18VFS_FileSystem
Fix
Code now restricts the number of IDs to 1024 and prevents the panic.
Service Request
15588711
Fixed in version
5.5.24.0
357
Version 5.5.24.2
Impact Level
Severity 1
Symptom
Description
An attempt to failover a VDM didn't succeed because the VDM root file system could not be
unmounted. This happened because extra references being taken on shares were not released. The
server log showed the repeated sequence of the following messages:
2006-08-07 14:37:05: VRPL: 3: Error while getting file system
ptr
2006-08-07 14:37:06: VRPL: 4: fs:327 is in unmounting state,
xxxwill retry in a few seconds (../vrplTools.cxx:93).
2006-08-07 14:37:10: VRPL: 4: last message repeated 1 times
2006-08-07 14:37:10: CFS: 4: fs 0x214 type = dhfs being
xxxunmounted. xWaiting for quiesce ...
2006-08-07 14:37:10: CFS: 4: fs 0x214 xtype = dhfs unmounted
2006-08-07 14:37:10: VRPL: 4: Playback: v:916, Delta:986,
xxxg:986, xxad:115344384, nDelta:1.
Fix
When shares are created, the extra references are no longer taken.
Service Request
15645839
Fixed in version
5.5.24.0
Impact Level
Severity 2
Symptom
Description
User was unable to mount BCV file system after a fs_timefinder -M refresh was done. In this
instance, diskmark was using a path that was unavailable.
Fix
Code now properly handles errors if diskmark command fails.
Service Request
13947461
Fixed in version
5.5.24.0
CIFS
358

problems:
Version 5.5.24.2
Impact Level
Severity 1
Symptom
Description
In a CIFS environment, the Data Mover panicked with the following DART panic/fault message:
>>PANIC in file: ../sched.cxx at line: 722 : SYSTEM WATCHDOG.
0x65f34: 0x13b747 waitForReboot+0x8b
0x65f54: 0x13b946 fault_dump+0x5e
0x65f74: 0x13b84f PANIC+0x2b
0x65f84: 0x13d236 system_watchdog+0x7c
0x65f94: 0x135829 _t_nmi+0x63
0x65fe0: 0x13ea88 Sthread_startThread_internal+0xe
Fix
The Data Mover GPO functionality has been optimized to avoid costly and frequent thread lookups
during normal operation which could lead to this panic.
Service Request
15575623
Fixed in version
5.5.24.0
Impact Level
Severity 2
Symptom
Description
When using Microsoft VSS (Volume ShadowCopy Service) and DFS (distributed file systems) to
access file systems, users could not see previous versions of file systems. If DFS was not used,
previous versions could be seen and accessed.
Fix
When using DFS, CIFS clients were sending requests with pathnames ending in a trailing \, which
the code considered part of the pathname. The code now correctly strips the trailing \ from target
pathnames.
Service Request
15357258
Fixed in version
5.5.24.0
359
Version 5.5.24.2
360
Impact Level
Severity 1
Symptom
Description
When using an LDAP server for user/group mappings (iPlanet/SunOne), new CIFS connections were
intermittently denied when many users were trying to connect at the same time. The problem
happened because querying the LDAP server for the groups each connecting user belonged to took a
long time.
Fix
A more efficient method of identifying the groups a user is a member of is now employed.
Service Request
15828272
Fixed in version
5.5.24.0
Impact Level
Severity 1
Symptom
Description
When multiple CIFS sessions were open and simultaneously writing to the event log, the Data Mover
../addrspac.cxx at line: 275 : Out of memory. The stack trace
showed:
0xe6c97b50: 0x13d05b waitForReboot+0x8b
0xe6c97b70: 0x13d263 fault_dump+0x67
0xe6c97b90: 0x13d163 PANIC+0x2b
0xe6c97ba0: 0x13900c
0xe6c97bc0: 0x17ef0a _Z11more_memoryiP9free_list+0x4e
0xe6c97be0: 0x17efb7 malloc+0x89
0xe6c97c10: 0x456269
xxx_Z26asn1buf_remove_octetstringP15code_buffer_r
0xe6c97c30: 0x8d9836
xxx_ZN13smb_threadCtx16SSXAuth_KERBEROSER16cSessi
Fix
Fixed memory leak condition that occurred when multiple CIFS sessions were open and logging
messages to event log. This panic will no longer occur.
Service Request
15773186
Fixed in version
5.5.24.0
Version 5.5.24.2
Impact Level
Severity 2
Symptom
Description
During a CDMS migration, if a file on the source file server had permissions for a local user with the
same name as a DOMAIN user, and the client attempted to do a SID lookup for the domain user,
Celerra incorrectly returned it. Executing this same call against a Windows server as the target
returned SID not found.
Fix
Local name or SID is now kept local to the Data Mover to prevent this condition.
Fixed in version
5.5.24.0
Impact Level
Severity 1
Symptom
Description
When allocating domain record description in Usermapper, the Data Mover panicked with the following
DART panic/fault message: >>PANIC in file: ../addrspac.cxx at line: 431
: Out of memory.
0xd37ff9a0: 0x13b7d3 waitForReboot+0x8b
0xd37ff9c0: 0x13b9d2 fault_dump+0x5e
0xd37ff9e0: 0x13b8db PANIC+0x2b
0xd37ff9f0: 0x137b4e
0xd37ffa50: 0x174bd6 _Z11more_memoryiP9free_list+0x4e
0xd37ffa70: 0x174c83 malloc+0x89
0xd37ffaa0: 0x175395 _Znwm+0x15
0xd37ffab0: 0x1753c2 _Znam+0xc
Fix
Code now requests only the memory needed to perform this operation and the Data Mover no longer
panics.
Service Request
15315965
Fixed in version
5.5.24.0
361
Version 5.5.24.2
Impact Level
Severity 2
Symptom
Description
A CIFS user that was a member of the Administrators group was unable to mange the server using
Windows tools. This happened if the user had connected to the system initially through a non-CIFS
protocol (for example, FTP).
Fix
Fixed a problem where the Administrative privilege were not granted in this situation.
Fixed in version
5.5.24.0
Impact Level
Severity 2
Symptom
Description
When the Security Descriptor (SD) on a subdirectory was changed, the parent directory's mtime was
updated. This is because the archive bit was set when the SD of the subdirectory was changed.
Fix
A new param was added (cifs.acl.archive) which controls whether the archive bit is effected by a
Security Descriptor change.
usage
Bit 0 - controls the behavior for files
Bit 1 - controls the behavior for directories
Setting a bit to 1 specifies that the archive bit will be updated if the SD is changed. The default setting
of 3 has Celerra continue its existing behavior of updating the archive bit when the SD is changed for a
file or a directory.
Note: Changing the archive bit does change the last modification time of the parent directory
since dos attributes are stored in the parent directory.
362
Service Request
15269475
Fixed in version
5.5.24.0
Impact Level
Severity 1
Symptom
Description
Due to a regression introduced in version 5.5.21.0, Win 3.11 and Win9x clients could no longer
connect to the Data Mover using the net use command. The following error message was returned in
DOS : Error 58: The network has responded incorrectly.
Fix
Code now accepts W3.11 and Win9x logins.
Fixed in version
5.5.24.0
Version 5.5.24.2
Impact Level
Severity 1
Symptom
Description
When user invoked the server_cdms server_2 -Convert <filesystem>

command, ALL CIFS users lost access to shares and the Data Mover panicked with the following
DART panic/fault message: >>PANIC in file: ../dskdump.cxx at line:
1450 : SMBBlocked.
0xf52fda60: 0x13d10f waitForReboot+0x8b
0xf52fda80: 0x13d317 fault_dump+0x67
0xf52fdaa0: 0x13d217 PANIC+0x2b
0xf52fdab0: 0x853b64 _Z11paniccfgcmdR6Option+0x178
0xf52fdbf0: 0x167f62
xxx_ZN11cfgcmd_desc6invokeEPKcPP11CCMDMessageii+0
0xf52fdc60: 0x2070be _Z10xml_cfgcmdP7macDataRP8CCMD_Msg+0x312
0xf52fdcb0: 0x20342a
xxx_ZN17macCCMDCommand_DB6invokeEPKcP7macDataRP8C
0xf52fdce0: 0x203958
xxx_Z17mac_CCMDDocParserP7macDataRP8CCMD_Msg+0x20
Fix
This problem was caused by a deadlock in CIFS thread that has been resolved.
Service Request
15832013
Fixed in version
5.5.24.1
Impact Level
Severity 1
Symptom
Description
Starting from version 5.5.22.0, some CDMS migrations failed to complete and the following errors
were displayed in the server_log:
2006-09-11 14:39:43: MGFS: 6: connection disconnected,
xxxattempt 1 in 1 sec
2006-09-11 14:39:44: MGFS: 6: connection attempt 1 of 2
2006-09-11 14:39:44: MGFS: 6: connection disconnected,
xxxattempt 2 in 1 sec
2006-09-11 14:39:45: MGFS: 6: connection attempt 2 of 2
2006-09-11 14:39:45: MGFS: 3: CIFS : Error: read remote dir
xxxfailed, ntstatus: c000020c
Fix
This issue occurred while reading large directories from the secondary storage, if all results werent
returned in a single reply. The code now handles this properly.
Service Request
15920270
Fixed in version
5.5.24.2
363
Version 5.5.24.2
Control Station
364

Control Station:
Impact Level
Severity 2
Symptom
Description
Following an upgrade to a nas version 5.4 or higher, user was trying to extend a 2 TB file system to 2.2
TB file system via the WebUI. The extend failed with the following error in the server log.
2005-07-10 17:28:46: CFS: 3: extend: size 4702337024 over 2TB
not enabled yet 2005-07-10 17:28:46: ADMIN: 3: Command
failed: file extend uxfs 172=59 s=4702337024
Fix
After upgrading to the newer version of code, file systems must be upgraded to support larger than 2
TB capacities. If this conversion is still in progress, extends will fail with the following: Filesystem
cannot be extended beyond 2TB, as high capacity mode is disabled.
Service Request
13384845
Fixed in version
5.5.24.0
Impact Level
Severity 2
Symptom
Description
High CPU and memory utilization occurred on systems with more than 10,000 exports when Celerra
Manager issued requests to retrieve configuration data for all Data Movers.
Fix
Code was altered so the data requested is independent per server instead of the entire system. This
reduces the amount of memory being used by Celerra Manager.
Service Request
14786517
Fixed in version
5.5.24.0
Impact Level
Severity 2
Symptom
Description
Duplicate disk names existed in nas_disk -list after adding disk volumes with
server_devconfig -create or Celerra Manager Rescan operation. This condition only
occurred when an existing disk volume was renamed to a conflicting name prior to the rescan.
Fix
Corrected unique name generation code.
Service Request
15681190
Fixed in version
5.5.24.0
Version 5.5.24.2
Impact Level
Severity 1
Symptom
Description
The Data Mover failover process aborted if the file system mount operations took longer than 60
seconds. An error was logged in the server log showing mount operation timeout and the Data Mover
failover aborted.
Fix
Extended timeout for the 'mount' operation. Timeout value now has a minimum of 120 seconds and
increases with the size of the file system.
Service Request
15685378
Fixed in version
5.5.24.0
Impact Level
Severity 2
Symptom
Description
If the Data Mover could not determine what UDP port on the client to send NFS lock responses to, lock
operations failed, and the following error message was logged in the server log:
2006-07-19 23:56:05: IP: 3: UDP error indication, errno = 99
2006-07-19 23:56:05: LOCK: 3: Bad LOCKD request dropped.
One of three conditions could cause this problem.
1.The client didn't respond to port lookup requests.
2. the client reported a port number of "0" which means lock services were not running, or
3. port lookups are blocked or dropped by the Network.
Fix
The Data Mover will now report a more useful message, and retry the port lookup operation. The new
error message is: IP: 3: UDP error indication, errno = 99
LOCK:
4: Lockd: UDP couldn't assign addr: <ip>, port 0.
Service Request
15611998
Fixed in version
5.5.24.0
365
Version 5.5.24.2
Impact Level
Severity 2
Symptom
Description
The server_export -refresh command was invoked from nasdb_backup, which itself
was scheduled to run every hour. Server_export -refresh took around 20-30 minutes.
This caused performance overhead and blocked the Celerra Manager from displaying the exports.
Fix
Code now uses cookie-based caching mechanism which significantly increases performance of the
server_export -refresh command to help prevent this condition.
Service Request
14741136
Fixed in version
5.5.24.0
FileMover
366

FileMover:
Impact Level
Severity 1
Symptom
Description
When corruption occurred on a DHSM file system, FSCK did not correct all corruption for some offline
inodes. If the system attempted to access the corrupted inodes, the Data Mover panicked with a
message similar to the following DART panic/fault message: readBlock: Bad Block
Index.
0xd4bf3440: 0x13b7ef waitForReboot+0x8b
0xd4bf3460: 0x13b9ee fault_dump+0x5e
0xd4bf3480: 0x13b8f7 PANIC+0x2b
0xd4bf3490: 0x5a8097
xxx_ZN14UFS_FileSystem28markCorruptedFsForUnmount
0xd4bf3520: 0x5b89f8
xxx_ZN14UFS_FileSystem9readBlockEliRP13Buffer_Cac
0xd4bf3590: 0x5eda3a _ZN9UFS_Vnode14readOpaqueDataEPci+0xd8
0xd4bf35d0: 0xac27f5 _ZN10DHSM_Vnode9typeMorphEv+0xe3
0xd4bf3620: 0xac2a4e _ZN10DHSM_Vnode8initNodeEi8dirTypes+0x3a
Fix
The code now more aggressively scans for corruption in offline inodes during fsck, and corrects these
problems when found.
Service Request
15646561
Fixed in version
5.5.24.0
Version 5.5.24.2
Impact Level
Severity 1
Symptom
Description
Offline files appeared to be truncated on pass-through read when reading with large (>8k) block sizes.
This symptom was visible when using NFS client (rsize/wsize > 8K), or the server_archive
-rw command. Although data appeared to be truncated to clients, the files were not truncated on the
disk.
Fix
Code now correctly handles end of file processing for offline file large reads.
Service Request
15850070
Fixed in version
5.5.24.0
Impact Level
Severity 2
Symptom
Description
The fs_dhsm -connection command failed when the primary and secondary file systems
were on the same Data Mover. This issue was the result of the internal communications code not
making a copy of the message before delivering it to the network layer.
Fix
These messages are now copied and internal communication for DHSM now works properly.
Fixed in version
5.5.24.0
Impact Level
Severity 2
Symptom
Description
NDMP based backups with automatic checkpoint creation failed to backup attributes of stub files on a
DHSM-enabled file system. Instead of backing up the attributes, it backed up the file data and
generated the following error message in the server_log:
2006-08-03 14:18:06: NDMP:4: Unable to write offline
attributes.
Fix
The stub file attributes are now properly backed up, and code no longer pulls the data from the
secondary server.
Fixed in version
5.5.24.0
367
Version 5.5.24.2
368
Impact Level
Severity 2
Symptom
Description
Users were unable to stop the CIFS service on a system with high amounts of FileMover (DHSM)
activity. The following error was logged in the server log: DHSM: httpd service
BLOCKED, unable to stop cifs - need to reboot
Fix
Code has been modified to handle this condition and CIFS service is now unaffected.
Fixed in version
5.5.24.0
Impact Level
Severity 1
Symptom
Description
When using Arkivio DHSM, user saw memory usage on the Data Mover increase over time. This
memory leak led to a Data Mover panic with the following DART panic/fault message: >>PANIC
in file: ../addrspac.cxx at line: 431 : Out of memory. The stack
trace showed:
0xa4d98aa0: 0x13b7ef waitForReboot+0x8b
0xa4d98ac0: 0x13b9ee fault_dump+0x5e
0xa4d98ae0: 0x13b8f7 PANIC+0x2b
0xa4d98af0: 0x137b6a
0xa4d98b50: 0x174bc6 _Z11more_memoryiP9free_list+0x4e
0xa4d98b70: 0x174c73 malloc+0x89
0xa4d98ba0: 0x7a9350
0xa4d98bc0: 0x7a8b6f
Fix
The code now prevents the memory leak that led to this Data Mover panic.
Service Request
15582999
Fixed in version
5.5.24.0
Version 5.5.24.2
Impact Level
Severity 3
Symptom
Description
When a file system was deleted and a share remained, the following exception was generated in the
webui.log: ERROR 2006-03-15 23:06:04,642 [COMMON]: Celerra
Manager: IndicationDistributer: Marshal Exception: objects of
this type must match the following regular expression:
[1-9]+[0-9]*: 1142492764641
Fix
Code has been fixed so this error is no longer generated and the share is displayed correctly.
Service Request
14651316
Fixed in version
5.5.24.0
Installation/
Upgrades
The following entries describe fixed issues related to installations and

upgrades:
Impact Level
Severity 1
Symptom
Description
In an NS40 environment, under certain CIFS load, the Data Mover rebooted and the following
messages were displayed in the server log:
Sep 19 15:09:21 2006 CHAMIIENCMON:3:101 Slot 2: 1158678414:
xxxOEM_ERROR: 09/19/06 15:05:17 : PCI Exp Fatal Error Message
xxxBus:00H Dev:06H Fn:00H PS:61H EIP:000000000013E671H
xxxESP:000000000364FFC4H
xxxOEM_ERROR: 09/19/06 15:05:17 : PCI Exp First Uncorrectable
xxxFatal Bus:00H Dev:06H Fn:00H PS:61H EIP:000000000013E671H
xxxESP:000000000364FFC4H
xxxOEM_ERROR: 09/19/06 15:05:17 : PCI Exp Multiple
xxxUncorrectables Bus:00H Dev:06H Fn:00H PS:61H
xxxEIP:000000000013E671H ESP:000000000364FFC4H
xxxOEM_ERROR: 09/19/06 15:05:18 : -R- PCI Exp Poisoned TLP
xxxBus:00H Dev:06H Fn:00H SID:0C00H FEP:0CH
xxxHL:0000000000CD5505FF10200D20400040H
Fix
As of this release, code updates the Data Mover BIOS to version 3.48, which improves stability and
prevents this problem.
Service Request
15968545
Fixed in version
5.5.24.0
369
Version 5.5.24.2
iSCSI
370

Impact Level
Severity 1
Symptom
Description
Under certain conditions, when multiple iSCSI logins were attempted simultaneously, the Data Mover
>>PANIC in file: ../malloc.cxx at line: 308: Memory corrupted.
0xd844de40: 0x13b7ef waitForReboot+0x8b
0xd844de60: 0x13b9ee fault_dump+0x5e
0xd844de80: 0x13b8f7 PANIC+0x2b
0xd844de90: 0x174d7f free+0x69
0xd844dec0: 0x175398 _ZdlPv+0x10
0xd844ded0: 0xbf7d46 _ZN18IscsiEventReceiverD0Ev+0x12
0xd844dee0: 0xc03ed6 _ZN12IscsiSession11decRefCountEv+0x5a
0xd844df00: 0xc04b15 _ZN9SessProxy11decRefCountEv+0x15.
Fix
Fixed a race condition in iSCSI login process where the reference count could become incorrect and
cause a Data Mover panic.
Fixed in version
5.5.24.0
Impact Level
Severity 1
Symptom
Description
During an upgrade from 5.4.16 to 5.5.21, iSCSI LUN was deleted because the UXFS file system was
full. The following error was displayed in the server_log:
2006-07-30 09:24:27: NBS: 3: invalid path or upgrade
incomplete: path=/fs66_T4_LUN1_APM00043702222_0000.
Fix
Code now checks for available space in UXFS file system before the upgrade, and preserves original
LUN when upgrade fails. iSCSI upgrade will resume after the Data Mover reboots.
Service Request
15271963
Fixed in version
5.5.24.0
Version 5.5.24.2
Impact Level
Severity 1
Symptom
Description
While running iSCSI Replication and attempting to abort a copy reverse, the Data Mover panicked with
the following DART panic/fault message: Assertion failure: 'dicSessionp !=
NULL' on line 743 in file '../DpTaskM anager.cxx' >>PANIC in
file: ../assert.c at line: 23 : assert failed.
0xc10c7e50: 0x13d10f waitForReboot+0x8b
0xc10c7e70: 0x13d317 fault_dump+0x67
0xc10c7e90: 0x13d217 PANIC+0x2b
0xc10c7ea0: 0x1873f3 _assert+0x53
0xc10c7ec0: 0xba28f5
xxx_ZN13DpTaskManager20createSyncDicSessionERP15d
0xc10c7f00: 0xba77ce
xxx_ZN33DpTaskManagerCopyReverseInitiator8abortSe
0xc10c7f40: 0xba6da9
xxx_ZN33DpTaskManagerCopyReverseInitiator8abortAl
0xc10c7f70: 0xba4034
xxx_ZN33DpTaskManagerCopyReverseInitiator8AbortAl
Fix
Code now handles this condition properly and the Data Mover no longer panics.
Fixed in version
5.5.24.0
Impact Level
Severity 2
Symptom
Description
While iSCSI replication was in progress, the user was unable to delete a LUN. The server_iscsi
<server> -d command returned the following error:
server_8 : LUN 75 (fs7192_T83_LUN75_000190100550_005C) is in
use by other applications, can't be deleted.
Error 4020: server_8 : failed to complete command
Fix
This issue occurred because the LUN delete was unable to stop replication of the LUN. A race
condition was found where this process may fail if a replication abort is already in progress. This
operation now completes successfully.
Fixed in version
5.5.24.0
371
Version 5.5.24.2
NDMP/Backup
372

Impact Level
Severity 2
Symptom
Description
NDMP backup performance was slow, for example, about 5MB/sec when running 3-way backups
inside the same Data Mover.
Fix
Code has been modified to improve performance when running 3-way backups inside the same Data
Mover.
Service Request
15610948
Fixed in version
5.5.24.0
Impact Level
Severity 2
Symptom
Description
While using Legato Networker, if a drive was in use by backup/recovery session, TAPE_OPEN took a
long time (nearly 8 minutes) and eventually timed out.
Fix
The code no longer retries tape open in raw mode so the Networker inquiry command can get the tape
device information quickly.
Fixed in version
5.5.24.0
Impact Level
Severity 2
Symptom
Description
When user issued the server_pax server_x -stat -verbose command while NDMP restore was in
progress, the Data Mover panicked with the following DART panic/fault message:
>>PANIC in
file: ../sched.cxx at line: 802 : SYSTEM WATCHDOG
0x3629f34: 0x13d05b waitForReboot+0x8b
0x3629f74: 0x13d163 PANIC+0x2b
0x3629f84: 0x13eb6e system_watchdog+0x7c
0x3629f94: 0x1370f9 _t_nmi+0x63
Fix
Fixed a race condition that occurred at the end of a restore process, when a PAX statistics command
was issued.
Fixed in version
5.5.24.1
Version 5.5.24.2
Networking
Impact Level
Severity 1
Symptom
Description
The server answers to TCP packets that are coming from a multicast address. This is known as the
'spank' denial of service attack. An attacker might take advantage of this behavior and cause
temporary data unavailability problems.
Fix
Added check for multicast source address when inspecting packets and now drop them appropriately.
Service Request
15270918
Fixed in version
5.5.24.0
Impact Level
Severity 1
Symptom
Description
While using CDMS over a large offline file, the Data Mover panicked with the following DART
panic/fault message: >>PANIC in file: ../bufpool.cxx at line: 322 :
couldn't get a free page. The stack trace showed:
0xef6650e0: 0x13d05b waitForReboot+0x8b
0xef665100: 0x13d263 fault_dump+0x67
0xef665120: 0x13d163 PANIC+0x2b
0xef665130: 0x160e93
xxx_ZN18Buffer_Cached_Pool17makeCachedBuffersEi+0
0xef665160: 0x139f1b
xxx_ZN19Memory_AddressSpace16makeCachedBufferEv+0
0xef665180: 0x160a21
xxx_ZN18Buffer_Cached_Pool15getAnyCachedBufEji+0x
0xef6651b0: 0x7075ae
xxx_ZN17UnCachedIO_Object9readBlockEym16BlockIO_P
0xef665230: 0x60944c
xxx_ZN14UFS_FileSystem9readBlockEliRP13Buffer_Cac
Fix
Fixed a memory leak condition in CDMS code to prevent this panic.
Service Request
15798941
Fixed in version
5.5.24.0
373
Version 5.5.24.2
Impact Level
Severity 1
Symptom
Description
Diskmark process run when adding new LUNS completed successfully, but new devices were not
marked.
Fix
Corrected a problem where errors were not reported by a Data Mover when I/O errors prevented
writing the diskmark. The diskmark process will now report the error to the user and the command will
fail properly.
Service Request
11776228
Fixed in version
5.5.24.0
NFS
374
Impact Level
Severity 1
Symptom
Description
While using NFSV4 to remove nested mount file systems, the Data Mover panicked with the following
DART panic/fault message: >>PANIC in file: ../vfsnode.cxx at line:738
: VFS_Vnode::release - releasing pinned node. The stack trace showed:
0xbcf19a40: 0x13d047 waitForReboot+0x8b
0xbcf19a60: 0x13d24f fault_dump+0x67
0xbcf19a80: 0x13d14f PANIC+0x2b
0xbcf19a90: 0x5c53f7
xxx_ZN9VFS_Vnode21releaseInternalLockedEPv11File_
0xbcf19aa0: 0x5c5abf
0xbcf19ad0: 0x566e33 _ZN15File_NamingNode7releaseEv+0x27
0xbcf19af0: 0x574339 _ZN12File_DirNode12removeLock
Fix
Code has been fixed to ensure the Data Mover no longer panics when removing nested mountpoint
using NFSV4.
Fixed in version
5.5.24.0
Version 5.5.24.2
Impact Level
Severity 2
Symptom
Description
Under high stress load, the Data Mover became unresponsive for up to 30 seconds. This happened
when the ufslog was filled to high threshold limit, and resulted in poor performance.
Fix
When the ufslog is full or near full, data will continue to be processed without requiring 30-second
timeouts which results in better performance.
Fixed in version
5.5.24.0
Impact Level
Severity 1
Symptom
Description
In an NFS environment, if several clients with similar names holding locks were rebooted
simultaneously, the Data Mover panic with the following DART panic/fault message:
>>PANIC in file: ../malloc.cxx at line: 308 : Memory
corrupted.
0xd1565d80: 0x13b7d3 waitForReboot+0x8b
0xd1565da0: 0x13b9d2 fault_dump+0x5e
0xd1565dc0: 0x13b8db PANIC+0x2b
0xd1565dd0: 0x174d47 free+0x69
0xd1565e00: 0x175360 _ZdlPv+0x10
0xd1565e10: 0x585768
xxx_ZN13Statd_Monitor13deleteMonitorEP17Statd_Mon
0xd1565e30: 0x58468b _Z12statd_notifyP4msgb+0x2ed
0xd1565fa0: 0x58409e _ZN13Statd_Initial5startEv+0x308
Fix
The lock monitoring system has been fixed to ensure no false hits are processed between clients with
similar hostnames, and the Data Mover no longer panics.
Service Request
15572682
Fixed in version
5.5.24.0
375
Version 5.5.24.2
Impact Level
Severity 2
Symptom
Description
The Data Mover hung when user attempted to access a certain directory via NFS. The ls commands
simply hung, as would any attempt to access the directory. Access was restored after the Data Mover
was failed over.
Fix
Fixed a deadlock that occurred when changing file or directory attributes.
Fixed in version
5.5.24.0
Impact Level
Severity 2
Symptom
Description
While using CIFS, the Data Mover panicked with the following DART panic/fault message: >>PANIC
in file: ../malloc.cxx at line: 327 :
Memory corrupted.
0xd7063290: 0x13d047 waitForReboot+0x8b
0xd70632b0: 0x13d24f fault_dump+0x67
0xd70632d0: 0x13d14f PANIC+0x2b
0xd70632e0: 0x17c33b free+0x69
0xd7063310: 0x17c954 _ZdlPv+0x10
0xd7063320: 0x17c96a _ZdaPv+0xc
0xd7063330: 0x624371
xxx_ZN12UFS_DirHashQ21deleteOffsetChunkListEv+0x5
0xd7063350: 0x6244ae _ZN12UFS_DirHashQD1Ev+0x24
Fix
Memory allocation function has been fixed to prevent the condition that led to the panic.
Service Request
15662236
Fixed in version
5.5.24.0
Replication
376

Version 5.5.24.2
Impact Level
Severity 2
Symptom
Description
After an IP replication session was failed over successfully, a resync was then attempted but failed
because of a communication problem. A firewall blocked communications between two Celerras,
however, the resync was run specifying an interface which was allowed by the firewall to communicate
to the remote side which should have succeeded.
Fix
Instead of using the specified interface, the resync was using the default interface which was blocked
by the firewall. Changes were made to ensure that the specified interface is used for all operations
needed by the resync and the operation will now succeed.
Service Request
15443985
Fixed in version
5.5.24.0
:
Impact Level
Severity 3
Symptom
Description
When checking the status of a replication session using fs_replicate -info -verbose,
the Dur (duration) field overwrote the Create Time. This caused the display to give the wrong delta set
create time and appear to be missing the duration.
Fix
The spacing on the delta set table shown when using the -verbose option was corrected so each
field is now readable.
Fixed in version
5.5.24.0
Impact Level
Severity 2
Symptom
Description
If an application other than replication or fs_copy (such as a port scanner) established a TCP
connection on Replicator's port (TCP port 8888), a network interface was marked as in-use by
Replicator. When user attempted to delete that interface, the operation failed with the following error:
currently in-use by replication.
Fix
System now verifies that the interface is truly in-use by Replicator when a user attempts to delete it.
Fixed in version
5.5.24.0
377
Version 5.5.24.2
378
Impact Level
Severity 1
Symptom
Description
When using Shadow Copy (VSS) from a Windows client, one the following problems could be seen:
1. If using Win/XP mounted to an NMFS share, user received an error when attempting to view copies
or restore an entry from a checkpoint.
2. If using Windows 2000 clients, shadow copy functionality did not work regardless of how the file
system was mounted.
Fix
Code now allows access to checkpoints of file systems mounted on NMFS through Shadow Copy on
both Win/Xp and W2K clients. The checkpoint name passed from the client may occur at any point in
the path - not just the first component. This will allow the checkpoint name to be looked up after the
path parsing moves from the NMFS to the checkpoint file system. This fix resolves the incompatibility
of the W2K client paths that prevented this client from using shadow copy.
Fixed in version
5.5.24.0
Impact Level
Severity 2
Symptom
Description
In some circumstances where replication databases between Celerras were out of sync, replication
commands failed with an error similar the following:
Eiffel dump: 2006-02-09 18:03:43.913 db:201:3673:E:
fs_replicate -a <file system>: Execution failed: Segmentation
violation: Operating system signal.
[NAS_SIGHNDLR.eiffel_handlr].
Fix
Code now returns the appropriate error, depending on the original failure condition.
Fixed in version
5.5.24.0
Impact Level
Severity 2
Symptom
Description
When FSCK was run on a replicated file system that was virtually mounted on the standby Data
Mover, replication became inactive.
Fix
Code no longer allows the server_mount -volume command on a replicated file system to
mount on any Data Mover other than the mount server.
Service Request
15301382
Fixed in version
5.5.24.0
Version 5.5.24.2
Impact Level
Severity 2
Symptom
Description
When inactivating the oldest checkpoint with a large amount of snapped data (in this case more than
256 gigabytes of data), it took a long time to free up the chunks, and other processes were affected,
causing CIFS outages and I/O errors.
Fix
The algorithm to release chunks has been optimized to help prevent this condition. This condition
occurred under very rare circumstances where extremely large checkpoints were inactivated.
Service Request
15013478
Fixed in version
5.5.24.0
Impact Level
Severity 1
Symptom
Description
After a failover, during the resync operation of replicated file system, a full copy was run instead of an
incremental copy.
Fix
When checkpoint is replicated on secondary file system, code now runs an incremental copy.
Service Request
15362023
Fixed in version
5.5.24.0
Storage
Impact Level
Severity 2
Symptom
Description
MAC OSX clients running in an active directory were required to supply a username, domain name
and password every time they tried to map to a CIFS share on a Data Mover. This occurred because
Kerberos authentication failed and clients fell back to NTLM authentication.
Fix
Added support for OSX Kerberos checksum type (MD5/MD4).
Fixed in version
5.5.24.0
379
Version 5.5.23.2
UFS
Impact Level
Severity 2
Symptom
Description
When a file system was built, the creation time for lost+found and the root directory of a file system
was Jan 1970.
Fix
The creation time for these directories is set to the current Data Mover time when the file system is
built.
Fixed in version
5.5.24.0
Version 5.5.23.2
Celerra
Management
Applications
Impact Level
Severity 2
Symptom
Description
When using the Celerra setup wizard, if Unicode was enabled, the first part of the wizard setup
process showed that all servers were Unicode enabled. The CIFS server section of the setup process,
however, incorrectly showed that Unicode was not enabled.
Fix
After enabling Unicode with the setup wizard, code now correctly detects that the Unicode is enabled
later in the CIFS server section. A reboot is no longer required.
Fixed in version
5.5.23.0
CIFS
380


problems:
Version 5.5.23.2
Impact Level
Severity 2
Symptom
Description
When a Data Mover detected an internal volume error, the sys_log filled up with Incorrect Reference
messages, and callhomes were being generated almost every minute.
Fix
The code now generates only one callhome per occurrence of this problem. After correcting the
problem, if overlap occurs again on the same volume, a new call-home event will be generated.
Fixed in version
5.5.23.0
Impact Level
Severity 2
Symptom
Description
When ABE (access-based enumeration) was enabled and a search mask was specified in a directory
listing, the list of files displayed on a Windows NT client did not display the file/directory attributes
properly. For example, file sizes were reported by the Data Mover to be over 24 petabytes. Running
dir *.* on both NT4 and XP clients displayed fine, but using a masking by running dir 0*.*
in the NT4 client exposed this problem.
Fix
When using a wildcard other than * or *.* and ABE is enabled on the share, the entry attributes
returned by the Data Mover are now filled in.
Service Request
15478699
Fixed in version
5.5.23.0
Impact Level
Severity 2
Symptom
Description
When SMB signing support was left as "undefined" in GPOs, the Data Mover assumed that SMB
signing was disabled. This prevented clients from domains where signing is required from logging into
the Data Mover.
Fix
Signing is now enabled by default. If GPO entry is explicitly disabled, the Data Mover will use that
setting.
Fixed in version
5.5.23.0
381
Version 5.5.23.2
Impact Level
Severity 2
Symptom
Description
MAC OS clients using CIFS and running MAC OS version 10.4 saw the Finder hang while attempting
to browse a list of shares on Celerra.This problem only happened with Power PC clients. Intel based
MACs were not affected.
Fix
Made changes to force clients to retrieve share lists the way they would have in OSX 10.3.
Service Request
15352887
Fixed in version
5.5.23.0
Impact Level
Severity 2
Symptom
Description
ACL check reported that it duplicated inode ACL IDs to primary ACL IDs, but did not.
Fix
Fixed the bug in duplication check where certain inode's ACL ID was not duplicated to primary ACL ID
correctly, even though the server log said the duplicated ACL ID was changed to primary ACL ID.
Fixed in version
5.5.23.0
Control Station
382

Control Station:
Impact Level
Severity 1
Symptom
Description
Control Station intermittently failed over due to an out of memory condition. The logs showed lots of
'out of memory: Killed process xxxx'. Examining the processes running on the
Control Station showed a number of instances of the server_cifs -o audit commands that
did not exit.
Fix
Code now detects if a server_cifs -o audit command has been running for more than 2
hours. If it finds one that has, it issues a kill -9 on all occurrences of this command.
Fixed in version
5.5.23.0
Version 5.5.23.2
Impact Level
Severity 2
Symptom
Description
System dialed home with a Jserver core dump error when a navicli -h <ipaddress>
command failed. However, if the command was run manually, it completed successfully.
Fix
Navicli memory management issue was resolved in navicli 6.22.21.0.0 which is included in this
release.
Service Request
15386674
Fixed in version
5.5.23.1
Impact Level
Severity 2
Symptom
Description
Renaming reserved disk volumes caused the rename operation to fail and left the database in an
improper state.
Fix
These disk volume names are now changed in the disks and volumes table correctly.
Fixed in version
5.5.23.0
Impact Level
Severity 1
Symptom
Description
In rare circumstances, a Data Mover can power-off suddenly (for example, due to hardware issues).
When this occurred, the Control Station monitoring functionality did not failover the Data Mover.
Fix
If a Data Mover loses power on its own and does not recover within 6 minutes, it will be failed over to
any available standby Data Mover. If no standby Data Mover is available, code will generate a call
home event.
Service Request
15143175
Fixed in version
5.5.23.0
383
Version 5.5.23.2
Impact Level
Severity1
Symptom
Description
While using the nas_fs -x command to extend a replicated file system, only the source file
system was extended, leaving the destination file system at a smaller size than the source. This
caused corruption on the destination file system and panicked the destination Data Mover, causing
replication to go inactive.
Fix
Code now prevents production file system extension when destination file system information can't be
retrieved. This ensures that both the destination and source file systems are extended when the
nas_fs -x command is invoked.
Fixed in version
5.5.23.2
CORE
384

Data Movers:
Impact Level
Severity 1
Symptom
Description
Users were able to create CIFS alternate data streams on files that had been archived with File Level
Retention.
Fix
Added access checks on stream creation to ensure streams cannot be created or removed on files
that have been archived with File Level Retention.
Fixed in version
5.5.23.0
Impact Level
Severity 1
Symptom
Description
When server handles many connection or open files, a client with limited memory resources can
cause an internal memory messages leak by not reading all the replies to its requests. This led to a
panic when all the messages resources were consumed.
Fix
Code now frees all MSRPC reply messages which have not been read by the client on pipe close and
on TCP disconnection.
Service Request
15386975
Fixed in version
5.5.23.0
Version 5.5.23.2
Impact Level
Severity3
Symptom
Description
When user changed the state of a VDM (virtual Data Mover) to mounted, running any
server_export command on that VDM returned the following Eiffel trace:
Execution failed: from_in_range: Precondition violated.
Fix
As per the expected behavior, in mounted state of VDM, server_export <vdm_name>

related operations like list/create/delete should not be carried out. Code will now report
error - VDM is not loaded.
Fixed in version
5.5.23.0
Impact Level
Severity 2
Symptom
Description
While doing a unicode (I18N) conversion, the Data Mover panicked with the following DART panic/fault
message:
*** Page Fault Interrupt. Virt ADDRESS: 5f2dd7 Err code: 0
Target addr: f714d5d4 **.
_waitForReboot+0x7d;
_fault_dump+0x56;
_err_exit+0xa;
_inactive__9UFS_VnodePvi11File_Statusi+0x380;
_releaseInactive__9VFS_VnodePvi11File_Statusi+0;
_releaseInternal__9VFS_VnodePvi11File_Status+0x;
_discardNewNode__9UFS_VnodeP20UFS_UpdateDescrip;
_dirConvert__17UFS_DirOps_compatP9UFS_VnodePv+0
Fix
The check for a valid name length field has been changed to prevent an I18N conversion to fail on any
directory that had an entry whose name length equaled 255 characters.
Service Request
12196039
Fixed in version
5.5.23.0
385
Version 5.5.23.2
Impact Level
Severity 1
Symptom
Description
The Data Mover panicked with the following DART panic/fault message:
checkMemory:CPU=2 Thread=0xcb3b5528 Initial=0xcb790858 Memory
corrupted at address=0xdee0 Value=0x8
waitForReboot+0x8b
fault_dump+0x67
PANIC+0x2b
_Z11checkMemoryPKvti+0x96
_ZN31Sthread_Condition_unconstructed4waitEP34S
_ZN14RawIO_CallBack4waitEv+0x2b
_ZN17UnCachedIO_Object10lio_listioE7LioModePP5
_ZN9UFS_Vnode15blockChainWriteEiP10BlockChaini
Fix
Updated BIOS for NSX and NS-40 to fix a problem that caused memory corruption.
Fixed in version
5.5.23.0
FileMover
386

FileMover:
Impact Level
Severity 1
Symptom
Description
While creating a CIFS connection using DHSM and CDMS, a null pointer condition resulted in a Data
Mover panic with the following DART panic/fault message: *** Page Fault Interrupt.
Virt ADDRESS: ea89aa Err code: 2 Target addr: 118 **.
0x1bd6604: 0x1451f7 waitForReboot+0x8b
0x1bd6624: 0x1453ff fault_dump+0x67
0x1bd6644: 0x13b0ee handle_fatal_fault+0xa
0x1bd6790: 0xe8112f
---_ZN15MGFS_FileSystem11mountRemoteEP15MGFS_Conne
0xa21ab660: 0xe9bb24 _Z18connectionInternalR6Option+0xc52
0xa21ab760: 0xe87110 _ZN15MGFS_FileSystem9reconnectEPci+0x2e
0xa21ab7d0: 0xe80ee4
---_ZN15MGFS_FileSystem18rebuildCidDBLockedEv+0x1
0xa21ab820: 0xe80d62
---_ZN15DHSM_FileSystem18rebuildCidDBLockedEv+0x1
Fix
Code now performs a status check that avoids using this null pointer.
Fixed in version
5.5.23.0
Version 5.5.23.2
Impact Level
Severity 2
Symptom
Description
Using the CIFS protocol to establish a FileMover connection to the secondary store, if the secondary
store became disconnected, two attempts were necessary to reconnect.
Fix
Code has been modified to ensure that the reconnect is successful on the first attempt.
Fixed in version
5.5.23.0
Installation/
Upgrades

upgrades:
Impact Level
Severity 1
Symptom
Description
While installing a gateway NS or NSX Celerra, NAS software installation would fail with the error
message:
Error returned from Agent Client reporting network transfer
error. Err: RWInetHostNotFound: No such host <hostname>.
Fix
This problem happened if the IP address of a CLARiiON SP (service processor) attached to the
Celerra was a subset of the IP address of the Control Station. For example, if the Control Station IP
address was 10.200.1.200, and an SP had an IP address of 10.200.1.20 (or 10.200.1.2), then the
installation failed. This has been corrected.
Fixed in version
5.5.23.0
NDMP/Backup

Impact Level
Severity 1
Symptom
Description
If the last backup data write to tape encountered "End of Media", the backup process did not
continue and caused subsequent backup jobs to hang.
Fix
The hang problem has been fixed.
Service Request
12313252
Fixed in version
5.5.23.0
387
Version 5.5.23.2
388
Impact Level
Severity 2
Symptom
Description
While using NDMP, a restore job to a file system where there was no inode space available was
reported as successful from DMA (data management administrator).
Fix
When no inode space is available, code now reports the following error to the client:
No space left on device .
Fixed in version
5.5.23.0
Impact Level
Severity 3
Symptom
Description
When trying to open a tape device with no tape loaded, the server log was flooded with FCDMTL
Initiator Check condition events. Sample of messages logged repeatedly:
: DRIVERS: 4: FCDMTL 2 [1.4.1] Initiator check condition:
loopID = 1., SK = 02, ASC/Q = 3a00
: DRIVERS: 4: last message repeated 1 times
: DRIVERS: 4: FCDMTL 2 [1.4.1]
Suppressed all but 2 out of 5 identical log entries.
Fix
Eliminated these informational server log messages.
Service Request
14200891
Fixed in version
5.5.23.0
Impact Level
Severity 1
Symptom
Description
When user attempted to do single file NDMP restores using Veritas, the restore failed if the backup
was created with a Celerra Network Server version earlier than 5.5.
Fix
Veritas requires the use of non-DAR restore to do an NDMP backup/restore from a Celerra version
earlier than 5.5. If the environment variable "DIRECT=y" flag was set, it was not cleared while NAS
switched to non-DAR operation. Code has been modified to clear the flag and ensure a successful
restore.
Service Request
15521559
Fixed in version
5.5.23.0
Version 5.5.23.2
Impact Level
Severity 1
Symptom
Description
Port 4658 on the Celerra Data Mover is designed to receive server_archive requests from the Celerra
Control Station. These commands should only be accepted from the Control Station over the private
network to the Data Mover. However, the code honored requests from the external networks to which
the Data Mover was connected.
Fix
Requests are only accepted from the Control Station over the internal/private network. Requests from
the external network are not accepted.
Fixed in version
5.5.23.0
Networking
Impact Level
Severity 2
Symptom
Description
In some cases on single CS/NS Gateway systems, when the internal network switch was suddenly
powered-off, the Control Station monitoring module rebooted the Data Movers due to loss of network
connectivity to the Data Movers even though the Data Movers were functioning properly.
Fix
When there is a physical problem in the internal network switch or the cables that connect the Control
Station or the Data Movers to it, Data Movers will no longer be rebooted as long as the Data Movers
are still responding over their serial ports. Instead, a callhome message will be logged indicating the
problem.
Service Request
15360738
Fixed in version
5.5.23.0
389
Version 5.5.23.2
Impact Level
Severity 1
Symptom
Description
After LUNs were removed, the Data Mover left an object in memory that could be reused while
creating or deleting file systems or volumes and cause the Data Mover to panic with the following
DART panic/fault message: >>PANIC in file: ../volume.cxx at line: 6718
: Device Serial Number mismatched.
0xd814d9f0: 0x13d047 waitForReboot+0x8b
0xd814da10: 0x13d24f fault_dump+0x67
0xd814da30: 0x13d14f PANIC+0x2b
0xd814da40: 0x6f1f50
---_ZN11BasicVolume17checkSerialNumberEv+0xf8
0xd814da70: 0x6ee776 _Z13volumeCommandR6Option+0xfdc
0xd814dc10: 0x161765
---_ZN11cfgcmd_desc6invokeEPKcPP11CCMDMessageii+0
0xd814dc80: 0x1fb25e _Z10xml_cfgcmdP7macDataRP8CCMD_Msg+0x30e
0xd814dcd0: 0x1f7652
---_ZN17macCCMDCommand_DB6invokeEPKcP7macDataRP8C
Fix
The object is now removed from memory to prevent this condition.
Service Request
15111530
Fixed in version
5.5.23.0
NFS
390
Impact Level
Severity 2
Symptom
Description
File Level Retention state was compromised after openattr/read operation was done over NFSv4.
Fix
Code has been modified to ensure streams cannot be created or removed on files that are in File Level
Retention state.
Fixed in version
5.5.23.0
Version 5.5.23.2
Impact Level
Severity 1
Symptom
Description
When updating an ACL using NFSV4, if the ACL was long or the message was fragmented (network
busy), the Data Mover panicked with the following DART panic/fault message:
*** Watch Point exception. Virt ADDRESS: 6624f5 ***
0x952f1958: 0x144a9b waitForReboot+0x8b
0x952f1978: 0x144ca3 fault_dump+0x67
0x952f1998: 0x13a986 handle_fatal_fault+0xa
0x952f19c0: 0x6647f1
____Z8nfs4_acliiP14nfs4_attr_infoP8nfs4_ctxP13nfs
0x952f1ac0: 0x6600f6
____ZN16nfsv4_attributes8VerfAttrEP8nfs4_ctxP15Fi
0x952f1b00: 0x63745e
____Z12nfs4_setAttrRP4msgbS1_RPhR8nfs4_ctx+0x31a
0x952f1d30: 0x62f2e4 _Z23nfs4_convertNfsv3Statusm+0x92c
0x952f1ea0: 0x60b4c4 _ZN17nfs_serverInitial5startEv+0x123c
Fix
The Data Mover will no longer panic under this condition. Code has been modified to ensure the ACL is
always updated, unless there is a problem in the request processing in which case an error is returned
to the user.
Fixed in version
5.5.23.0
Impact Level
Severity 1
Symptom
Description
If more than 800 files were opened with NFSv4 when user attempted to unmount the file system on the
Data Mover, the Data Mover hung and messages similar to the following were logged in the server log:
STREAM WATCHDOG: handle/modid: 0x594c2a8/0x6, client's
IP/Port: 0xac1846f6/0x3e8, server's IP/Port: 0xac184718/0x801.
Fix
Solved locking issue to ensure the Data Mover no longer hangs under this condition and the unmount
operation will succeed.
Fixed in version
5.5.23.0
391
Version 5.5.23.2
Impact Level
Severity 2
Symptom
Description
If mounting of an NFS file system could not complete because of CLARiiON sector errors, the Data
Mover returned a Stale filehandle error to the NFS client, causing the operation to immediately fail.
Fix
The Data Mover no longer responds in this situation, allowing the client to retry. Once the sector error
is repaired, the Data Mover will continue to respond successfully.
Fixed in version
5.5.23.0
Replication
392

Impact Level
Severity 2
Symptom
Description
When running replication-related commands such as fs_replicate -info and the remote
Control Station was heavily loaded, commands failed with an eiffel dump similar to
Error 2237: Execution failed: Segmentation violation:
Operating system signal. [NAS_SIGHNDLR.eiffel_handlr].
Fix
Instead of the command failing on the first try, it now retries 3 times before failing. The fix improves
remote queries reliability.
Fixed in version
5.5.23.0
Impact Level
Severity 2
Symptom
Description
After encountering bad sectors in a Celerra Replicator SavVol, playback state showed active even
though it was inactive.
Fix
The fs_replicate command now reports playback state properly.
Fixed in version
5.5.23.0
Version 5.5.23.2
Impact Level
Severity 1
Symptom
Description
When a differential fs_copy was started while one of the checkpoints between the two specified was
unmounted, a full fs_copy was run. This means the copy took longer than the administrator expected.
Fix
A message is now logged in the server log to indicate that a full fs_copy has been started instead of a
differential.
Service Request
15375168
Fixed in version
5.5.23.0
SRDF

Impact Level
Severity 2
Symptom
Description
After running the /nas/sbin/rootfs_rdf <fs> -Restore -Force command, user

was unable to mount the production file system. The server log displayed errors similar to:
: STORAGE: 3: IsMarkerValid(): BadDiskMark vname:341
disk_id:152 marker:96
: STORAGE: 3: 1: Error: Path:c0t1l13 for volume:341 not valid
(Path/DiskMark inconsistent)
Fix
Code now restores the diskmarks as part of the nearcopy / farcopy restore process.
Fixed in version
5.5.23.0
Storage
393
Version 5.5.22.2
Impact Level
Severity 1
Symptom
Description
The Data Mover panicked with the following DART panic message:
>>PANIC in file: ../volume.cxx at line:7010 : Physical device
has been disconnected.
0xbc2538f0: 0x13d1d3 waitForReboot+0x8b
0xbc253910: 0x13d3db fault_dump+0x67
0xbc253930: 0x13d2db PANIC+0x2b
0xbc253940: 0x717f46
____ZN11BasicVolume17checkSerialNumberEv+0x126
0xbc253a50: 0x7145cf _Z13volumeCommandR6Option+0xed3
0xbc253bf0: 0x16803e
____ZN11cfgcmd_desc6invokeEPKcPP11CCMDMessageii+0
0xbc253c60: 0x2073b6 _Z10xml_cfgcmdP7macDataRP8CCMD_Msg+0x312
0xbc253cb0: 0x203722
____ZN17macCCMDCommand_DB6invokeEPKcP7macDataRP8C
Fix
During volume serial number check, code now checks the validity of inquiry data as well as the status
return. The Data Mover will no longer panic and the operation will succeed.
Fixed in version
5.5.23.2
Version 5.5.22.2
Celerra
Management
Applications
394

Impact Level
Severity 2
Symptom
Description
Upon logging into the GUI, the following error was presented:
Query Celerras All. The process for task 48496 has terminated
abnormally. It will not be restarted.
Fix
If a user altered one of their Data Mover's ACL values to something other than 1000, that Data Mover
was considered to be hidden (or could only be accessed by logging with the correct logging level).
When hidden, the Data Mover didn't show up in the list of available Data Movers. This has been
corrected to ensure the Data Movers are visible.
Service Request
15236968
Fixed in version
5.5.22.0
Version 5.5.22.2
CIFS

problems:
Impact Level
Severity 2
Symptom
Description
When using version 5.5.16.0 or higher, users could follow softlinks when using a Windows client, but
were unable to follow the same softlinks using a scanner application.
Fix
In a CIFS lookup request, if DFS flag is set, the Data Mover returned a file not found error. Clients
should not be setting this flag unless the pathname specified is a DFS pathname. However, code now
returns the name specified regardless of whether or not this bit is set.
Service Request
14897535
Fixed in version
5.5.22.0
Impact Level
Severity 2
Symptom
Description
Clearcase test jobs were periodically failing with access denied when trying to rename files.
Fix
In cases where files are created and recreated quickly and inode numbers are reused, files could get
flagged with an attribute that prevented them from being renamed. This attribute is no longer set
incorrectly.
Fixed in version
5.5.22.0
Impact Level
Severity 2
Symptom
Description
In a very large domain, where it could take over an hour for replication within the active directory to
complete, the cifs.acl.mappingErrorAction parameter was invoked to allow new users
to be added to ACLs before active directory replication completed. Since upgrading to version
5.5.21.0, permissions could not be set on a new/unresolved SID. This impacted operations when
adding new users, as they could not be added or modified until active directory replication finished.
Fix
This was caused by a regression introduced in 5.5..21.0 which has been corrected.
Service Request
14860313
Fixed in version
5.5.22.0
395
Version 5.5.22.2
396
Impact Level
Severity 1
Symptom
Description
User needed to enable Guest access to a CIFS share to provide unauthenticated access to a shared
resource. He enabled the Guest account on the CIFS Server and granted it Full Control share and file
level permissions. When performing directory listing of the share, he received the error message:
"Logon failure: unknown user name or bad password." When executing a
net view against the CIFS server, user received: "System error 5 has occurred.
Access is denied." This problem only occurred if the guest password was empty.
Fix
Code now supports unauthenticated logon when local guest password is empty.
Service Request
15190095
Fixed in version
5.5.22.0
Impact Level
Severity
Symptom
Description
When configuring iPlanet, domain names were entered in the fully-qualified-domain-name form
(abc.def.ghi). This was only interpreted as dc=abc,dc=def,dc=ghi which didn't match the user's
configuration.
Fix
Added support for x509-formed names, for example: ou=abc,o=def,c=ghi.
Service Request
14814380
Fixed in version
5.5.22.0
Impact Level
Severity 2
Symptom
Description
After changes were made to the profile on the LDAP server, the Data Mover continued to use the old
setting for the attribute.
Fix
Celerra now correctly checks the LDAP server for updates to the attributes.
Service Request
15111527
Fixed in version
5.5.22.0
Version 5.5.22.2
Impact Level
Severity 1
Symptom
Description
In a Kerberos environment, the Data Mover panics with the following DART panic/fault message:
waitForReboot+0x8b;
fault_dump+0x5e;
PANIC+0x2b;
_ZN19Memory_AddressSpace10allocPagesEi12Memory;
_Z11more_memoryiP9free_list+0x4e;
malloc+0x89;
krb5_copy_data+0x45;
_Z18krb5_mcc_next_credP13_krb5_contextP12_krb5.
or
DART panic/fault message: >>PANIC in file: ../malloc.cxx at
line: 213 : malloc() - incorrect size - not a positive value
Stack trace:
waitForReboot+0x8b;
fault_dump+0x5e;
PANIC+0x2b
malloc+0x3d
krb5_copy_principal+0x142
krb5_get_cred_via_tkt+0xac6
krb5_get_cred_from_kdc+0x30
krb5_get_credentials+0x116
Fix
A problem was fixed in Kerberos where a memory leak or dereferencing a NULL pointer caused a
panic in a multi-realm Kerberos environment.
Service Request
15193890
Fixed in version
5.5.22.0
397
Version 5.5.22.2
398
Impact Level
Severity 2
Symptom
Description
DNS updates failed for a CIFS server joined to a child domain when the DNS authority was contained
in another child domain. A CIFS server joined to the child domain containing the DNS authority had
no problems updating DNS.
Fix
Changes were made to correctly traverse from one domain child to another (sibling domain). This
corrects the issue where authentication across realms using Kerberos was failing and prevented the
secure DNS updates from succeeding.
Service Request
14741088
Fixed in version
5.5.22.0
Impact Level
Severity 1
Symptom
Description
When using Microsoft Services for NIS to map UIDs and GIDs, if a user was a member of more than
1024 groups, the Data Mover panicked if services for NIS sent more than one line from the NIS group
database for the same key (group name). DART panic/fault message: >>PANIC in file:
../addrspac.cxx at line: 431 : \n Out of memory.
ef waitForReboot+0x8b
fault_dump+0x5e
PANIC+0x2b
_ZN19Memory_AddressSpace10allocPagesEi12Memory
_Z11more_memoryiP9free_list+0x4e
malloc+0x89
_Znwm+0x15
_Znam+0xc
Fix
The Data Mover no longer panics when the NIS data base contains more than one line for the same
group, but subsequent lines are discarded.
Service Request
15220335
Fixed in version
5.5.22.0
Version 5.5.22.2
Impact Level
Severity 1
Symptom
Description
A design limitation in the original implementation of ACLs for CIFS created a window which allowed
duplicate ACLs to be created in the Data Mover. This did not cause any problems for security, however
it allowed a situation where some files created their own copy of the ACL rather than link to the existing
copy of the ACL. When the ACL database was nearly full, or actually filled, CIFS users were unable to
access some shares, files or directories. They were unable to create new files or directories, or modify
permissions on existing files and directories. There were also some signs of this condition adding to
the load on the CPU, thus leading to a performance impact.
Fix
The mechanism for managing ACLs on the file system has been redesigned. Rather than keeping an
LRU (least recently used) cache of the most recently used ACLs in memory, the Data Mover now
keeps a hash list of all ACLs on disk in memory. This means that the opportunity for a cache miss has
been eliminated which was the reason for duplicate ACLs being created.
Service Request
13830658
Fixed in version
5.5.22.0
Impact Level
Severity 1
Symptom
Description
When the CIFS server name was NULL or invalid, the Data Mover panicked with the following DART
>>PANIC in file: ../malloc.cxx at line: 297 : free() called
with NULL argument.
waitForReboot+0x8b;
fault_dump+0x5e;
PANIC+0x2b free+0x1d;
krb5_free_principal+0x28;
krb5_free_ticket+0x26;
krb5_free_ap_req+0x23;
krb5_rd_req+0x1d7;
Fix
Prior to calling a function which uses the length of the CIFS server name to control a loop, code now
checks the data to ensure it's not NULL before freeing invalid memory. This fix prevents the Data
Mover panic.
Service Request
15031775
Fixed in version
5.5.22.0
399
Version 5.5.22.2
Control Station
400

Control Station:
Impact Level
Severity 1
Symptom
Description
Eiffel error followed by No disks attached to server error during initial install.
Fix
Corrected parsing of the hostinq output to properly handle the 5e encoded device ids. Installation is
successful.
Fixed in version
5.5.22.0
Impact Level
Severity 1
Symptom
Description
The nas_fs -xtend size= command allows a user to specify the size of the extension to the
file system in megabytes (M), gigabytes (G) or terabytes (T). When running the extend from the
command line, if a user specified the space for the extension as MB, GB or TB, the file system could
be extended to a huge size, consuming all available space in a pool, or else the command failed
reporting not enough space available.
Fix
The presence of the "B" caused the command to interpret the size as a hex number. This has been
fixed and only decimal values are allowed for file system size.
Service Request
14512251
Fixed in version
5.5.22.0
Impact Level
Severity 1
Symptom
Description
If the serial number in /nas/sys/callhome.config contained a space, the transfer_support_materials

script went into an infinite loop when it was invoked.
Fix
Fixed the script to handle the space so it does not go into the infinite loop.
Fixed in version
5.5.22.0
Version 5.5.22.2
Impact Level
Severity 2
Symptom
Description
The following false power supply failure messages were seen in the sys_log:
JServer:4:0 "Polling of data mover in slot 3 has been shut
down"
CHAMIIENCMON:1:6 Slot 3: 1130027853: encmon: Power Supply A
failed. PS faulted.
CHAMIIENCMON:1:16 Slot 3: 1130027853: encmon: Power Supply B
failed. PS faulted.
Fix
Code now inhibits certain power supply signals for 20 seconds to prevent erroneous power supply
failure reporting.
Service Request
14024758
Fixed in version
5.5.22.0
Impact Level
Severity 1
Symptom
Description
Some messages were too long to fit into message logged in /nas/log/sys_log, causing the message to
be truncated. For example: May 17 09:35:42 2006 OTHERS:6:1106
transfer_support_materials[3331]: The transfer script has
finished successfully. Your authorized service provider
should check host 123.112.4.56 for a dump saved as
dump_nsx.060517_0933.dump and a dump header file
header_nsx.060517_0933.txt in /incoming/n
the ending "sx" is truncated from the message.
Fix
Changed and split up texts from a successful ftp upload so it fit into 256 characters limit.
Fixed in version
5.5.22.0
401
Version 5.5.22.2
Impact Level
Severity 2
Symptom
Description
While upgrading from version 5.5.16.2 to 5.5.17.2, user saw the following message. Creating
IDE cache of NAS file system ########## done
Starting
services... Restarted crond Restarted NIS
Performing
Inventory....Error: T2NETCMD (T2netCMD_I2C_Device) - operand
38 out of range Done. System Inventory is placed in
/nas/log/inventory. Fri Jan 13 14:04:46 EST 2006
Root Disk
Upgrade Succeeded
Stopping artificial heartbeat. Fri Jan 13
14:04:50 EST 2006 Performing final upgrade tasks ... done
Restarting the other Control Station ... done
Model: NSX
Fix
This problem occurred because the cabinet type was detected incorrectly. To prevent this condition, a
loop now waits until valid cabinet information is returned.
Fixed in version
5.5.22.0
Impact Level
Severity 2
Symptom
Description
Celerra Monitor didn't respond when viewing CPU/Memory Usage. Also, when Tools ->
Celerra was clicked, the following error was displayed: This HTTP Status 503 Connection refused. This usually occurred after an upgrade.
Fix
When a system was upgraded, the pci_param file was placed on the system without the signature file.
The pci_param file needs the signature file in order to work. When upgrading, the signature file is now
included within the upgrade package.
Fixed in version
5.5.22.0
CORE
402

Data Movers:
Version 5.5.22.2
Impact Level
Severity 1
Symptom
Description
The Data Mover went into a rolling panic if a fs_copy command was issued on a read-only file system.
>>PANIC in file: ../sched.cxx at line: 802 :
SYSTEM WATCHDOG.
waitForReboot+0x8b;
fault_dump+0x67;
PANIC+0x2b;
system_watchdog+0x7c;
_t_nmi+0x63;
_ZN9RawIO_IRP7irpDoneEv+0x53;
_ZN8SgIO_IRP7irpDoneEv+0x57;
_ZN14CamDisk_Device8completeEP10ccb_header+0x16.
Fix
Code attempted to make changes to the source file system which was read-only. Needed changes
are now only kept in memory to prevent the Data Mover panic.
Service Request
15135621
Fixed in version
5.5.22.0
Impact Level
Severity 1
Symptom
Description
On a file system with checkpoints, if the SavVol becomes full, code tries to inactivate the oldest
checkpoint to reclaim space. If the oldest checkpoint is mounted, it needs to be unmounted before the
inactivation can occur. This umount occurs synchronously, and causes a write to the PFS. In this case,
the write couldn't progress because a process was being waited upon. One process was waiting for
the umount, the umount was waiting for the write to pass, and the write was waiting for the process to
unmount. This resulted in a deadlock condition.
Fix
The oldest checkpoint is now unmounted ahead of time 'asynchronously' when space is on the
SavVol. As of this release, If the SavVol fills up and the oldest checkpoint needs to be inactivated, the
oldest checkpoint would already have been umounted, so the inactivation can proceed.
Service Request
14561267
Fixed in version
5.5.22.0
403
Version 5.5.22.2
Impact Level
Severity 2
Symptom
Description
VDM with thousands of CIFS shares were being listed slowly in the CLI and in the GUI. When using
the GUI, if users were making changes to the exports, the changes were not reflected immediately.
Users needed to refresh the GUI in order to see the changes.
Fix
The performance of the listing of the CIFS shares on VDM has been improved by Implementing a
cookie based caching mechanism.
Fixed in version
5.5.22.0
FileMover
404

FileMover:
Impact Level
Severity 1
Symptom
Description
In DHSM configurations using CIFS connections to a secondary server under heavy load, when the
secondary server was dropping and re-establishing connections, the Data Mover panicked with the
following DART panic/fault message: >>PANIC in file: ../dskdump.cxx at
line: 1428 : cifs hang. The stack trace showed:
_fault_dump+0x56;
_PANIC+0x24;
_paniccfgcmd__FR6Option+0x13f ;
_invoke__11cfgcmd_descPCcii+0xf5;
_macthr_v_cfgcmd__FP4msgb+0x681;
_start__11Mac_Initial+0x5fa;
_Sthread_startThread_internal+0x11.
Fix
New code now protects session and connection objects with a new locking scheme to prevent this
condition.
Service Request
15031612
Fixed in version
5.5.22.0
Version 5.5.22.2
Impact Level
Severity 2
Symptom
Description
When running a policy engine server to archive files with FileMover, it was possible that the put offline
operation failed for some files. This showed up as an error in the policy engine even though it was
caused by the Celerra.
Fix
This was caused by a race condition on the Celerra that truncated the date string when replying to the
Policy Engine server. This has been fixed.
Fixed in version
5.5.22.0
Impact Level
Severity 2
Symptom
Description
When reading stub files from a CIFS server, the maximum data recall rate was 10MB/sec even when
using a GigE connection between the PC client / FileMover and secondary CIFS server and reading
multiple stub files at the same time using 32k sized reads.
Fix
Celerra can now sustain a data rate of at least 25MB/sec when performing DHSM recalls.
Service Request
14840741
Fixed in version
5.5.22.0
HighRoad (MPFS)

File System (MPFS):
Impact Level
Severity 2
Symptom
Description
When an NFS client truncated a file, setting EOF resulted in the MPFS server replying incorrectly with
a RevokeList notification of length zero.
Fix
Moved call to break MPFS locks to the correct place in the commit sequence to avoid this problem.
Fixed in version
5.5.22.0
405
Version 5.5.22.2
Impact Level
Severity 2
Symptom
Description
Using MPFSi, when doing large sequential write tests followed by large sequential reads, it was
possible for the client application to hang during the read cycle.
Fix
Code now skips notifications when the associated file has FMP (file mapping protocol) files open and
the associated block is locked for FMP.
Fixed in version
5.5.22.0
Impact Level
Severity 2
Symptom
Description
When writing files smaller than 96 KB, it could take up to twice as long using MPFS than using NFS.
Fix
Optimized space allocation routines to align with internal file system block boundaries and improve
MPFS performance.
Fixed in version
5.5.22.0
iSCSI
Impact Level
Severity 1
Symptom
Description
The nas_create_lun function failed to create LUN with the following error "failed to create
file versionset 421:/fs421_T3_LUN1_HK192200821
RA_0000/fs421_T3_LUN1_HK192200821 RA_0000 err
VersionSet_VersionNotFound!" Therefore, iSCSI sessions did not work.
Fix
The correct information in the inquiry data is now returned to iSCSI so it can create the iSCSI LUNs.
Fixed in version
5.5.22.0
NDMP/Backup
406


Version 5.5.22.2
Impact Level
Severity 2
Symptom
Description
NDMP backup/restore tape drive's unit-attention forced a Data Mover to inquiry scan for new LUN
configurations. Some tape drives could not handle command overlap and caused the library inventory
or backup operation to fail when swapping tapes.
Fix
Code no longer forces an inquiry scan on unit-attention to generic devices.
Service Request
14509950
Fixed in version
5.5.22.0
Impact Level
Severity 1
Symptom
Description
While using NDMP, user was unable to perform multiple simultaneous backups or any backups at all.
Fix
Corrected a race condition that caused NDMP threads to hang.
Service Request
15176844
Fixed in version
5.5.22.0
Impact Level
Severity 1
Symptom
Description
Celerra NDMP files were not restored if the file's M83 (DOS 8.3) name already existed in the
destination. This happened if
1. the backed up file name was longer than 8 characters.
2. after the file was backed, it was renamed or deleted, then another file was created having the
same first 8 character name as the original file.
Fix
Code now changes the M83 name to avoid the collision.
Service Request
13684065
Fixed in version
5.5.22.0
407
Version 5.5.22.2
Impact Level
Severity 1
Symptom
Description
The maximum allowed NDMP sessions was 16. Subsequent NDMP connection requests were
rejected by DART.
Fix
The limit on Celerra NDMP sessions has been increased to 2048.
Service Request
13912310
Fixed in version
5.5.22.0
Networking
408
Impact Level
Severity 2
Symptom
Description
User saw very slow performance when using NFS over TCP. Using NFS over UDP showed
significantly better performance This was only seen on clients talking across network links that were
dropping packets. The network packet drops will always cause a performance hit, but the Data Mover
was taking an excessively long time to retransmit dropped data.
Fix
Fixed a bug that caused slow TCP recovery if packet drops occurred and client had a reduced TCP
window size.
Service Request
14910654
Fixed in version
5.5.22.0
Impact Level
Severity 1
Symptom
Description
On a network with many packet drops, TFTP transfers would abort. In this particular case, user could
not boot from a Data Mover via TFTP.
Fix
The retransmit transfers now times out if the same three frames are retransmitted sequentially.
Previously, the code aborted if three frames were retransmitted during the entire TFTP session.
Service Request
14288992
Fixed in version
5.5.22.0
Version 5.5.22.2
Impact Level
Severity 1
Symptom
Description
After experiencing a TOE NIC panic or reset, the Data Mover panicked with the following DART
*** Page Fault Interrupt. Virt ADDRESS: 3147aa Err code: 0
Target addr: fd000058 **.
waitForReboot+0x8b
fault_dump+0x5e
handle_fatal_fault+0xa
_Z7atk_fsmP10atk_socket9ATK_EVENTmmm+0x5cc
_Z14atkfsm_timeoutPv+0x2d1
_Z13RTtimer_schediPv+0x77
_ZN15SchedRT_Initial5startEv+0xf6
Sthread_startThread_internal+0xe
Fix
Fixed a problem with handling of outstanding timers and events that could fire while recovering the
TOE NIC. If these timers fired during this time, memory could be corrupted. This has been resolved.
Service Request
14982663
Fixed in version
5.5.22.0
Impact Level
Severity 1
Symptom
Description
When using Data Movers with TCP-offload-engine (TOE) interfaces, during the TCP recovery that
occurs after packet drops, the following device message was logged to syslog: slicNicDevice: TOE
(aceX) panic detected. Saving TOE core image to '/slicdump'. Any TCP connections which were
off-loaded to the TOE device at the time of the panic were lost and had to be re-established after the
device was auto-restarted. Any connections/protocols ineligible for offload, or offload-eligible
connections not currently off-loaded to the TOE device, typically survived any TOE panic/restart
unimpacted. The TOE device dump-collection/restart time is typically around 30 seconds.
Fix
The bug resulting in the mis-management of the TCP state field 'snd_cwnd' has been identified/fixed.
Service Request
14891425
Fixed in version
5.5.22.0
409
Version 5.5.22.2
Impact Level
Severity 1
Symptom
Description
The FGE (fibre gigabit ethernet) device reset during a period of high traffic resulting from an fs_copy.
The following messages were logged in the server log when this occurred:
DRIVERS: 4: fge0 : Mbuf Low - Resetting Device
DRIVERS: 4: Restarting the port : fge0
DRIVERS: 4: fge0: Shutting down Tigon3
DRIVERS: 4: 2: fge0: link down
DRIVERS: 4: 3: fge0: link up
DRIVERS: 4: fge0: Link status -1000 Mbps full duplex, receive
& transmit flow control ON.
Fix
Instead of resetting the device when this condition occurs, the threads are now scheduled to allow the
host to process packets. This restores the device to normal operation.
Service Request
14729083
Fixed in version
5.5.22.0
Impact Level
Severity 1
Symptom
Description
In some cases where one or more paths failed due to HBA, switch or FA problems which cause I/O
timeout errors, the Data Mover panicked even though there were good alternate paths.
Fix
Failed I/Os are now retried on every alternate path until success.
Service Request
15190095
Fixed in version
5.5.21.405
NFS
410
Version 5.5.22.2
Impact Level
Severity
Symptom
Description
When an NFS client references a stale file handle and the treeid within the handle is also different, a
message similar to the following was logged in the server log:
UFS: 3: foundNode(valid), with treeId 1 different than passed
treeId 3, please unmount & re-mount fs 61 from nfs clients.
Fix
This message was a false warning and is no longer logged. The code now properly identifies stale
NFS file handles, it no longer identifies them as file handles with incorrect tree IDs.
Service Request
14971492
Fixed in version
5.5.22.0
Impact Level
Severity 1
Symptom
Description
The Data Mover panicked when a showmount -e command from a client was run at exactly the same
time a server_export -u command was run on the very first export entry. DART panic/fault message:
*** GP exception. Virt ADDRESS: 50036b. Err code: 0 ***.
waitForReboot+0x8b;
fault_dump+0x67;
handle_fatal_fault+0xa;
_ZN10Nfs_Export14processOptionsEv+0x80d ;
_ZN10Nfs_Export12mount_exportEPvP7svc_reqPP11F;
_Z18mountproc_export_1PvP7svc_reqPP11FreeResul;
_ZN28mount_IterSvc_Parent_InitialC1EP16Rpc_Svc
_ZN21Rpc_Svc_Child_Initial13svc_getreqsetEP7rp
Fix
In the routine which handles showmount -e command, we now acquire the global mutex on the export
table before we start parsing it.
Service Request
15183174
Fixed in version
5.5.22.0
411
Version 5.5.22.2
412
Impact Level
Severity 1
Symptom
Description
When Clients were performing large ACL updates with many ACEs in a network with a lot of
fragmented traffic, the Data Mover panicked with the following DART/panic fault message:
>>PANIC in file: ../assert.c at line: 23 : assert failed.
waitForReboot+0x8b;
fault_dump+0x6;
PANIC+0x2b;
_assert+0x53;
free+0xd6;
_ZN13nfs4_attrhold17nfs4SetattrVerifyEmRPm+0xc;
_Z15do_unix_setAttriP14nfs4_attr_infoP13nfs4_a;
_Z9nfs4_unixiiP14nfs4_attr_infoP8nfs4_ctxP13nf
Fix
Memory corruption condition has been fixed to prevent this panic.
Fixed in version
5.5.22.0
Impact Level
Severity 1
Symptom
Description
When running a tree quota report, the Data Mover panicked with the following DART panic/fault
message:
waitForReboot+0x8b;
fault_dump+0x5e;
PANIC+0x2b;
malloc+0x89;
_Znwm+0x15;
_Znam+0xc
Fix
We allocate 1284 bytes of memory for writing tree list record. We were not freeing that memory so
there was 1284 bytes of memory leak per tree list command. Code now deletes allocated memory and
eliminates the memory leak.
Service Request
15074297
Fixed in version
5.5.22.0
Version 5.5.22.2
Impact Level
Severity 1
Symptom
Description
When working on a file system with a large number of files in the root directory (in this case 217,000
files), access to the file system was very slow, and at times appeared to have hung.
Fix
Changes were made to improve the speed of scanning large directories.
Service Request
14717349
Fixed in version
5.5.22.0
Impact Level
Severity 2
Symptom
Description
When Solaris client attempted to run the -ls command in the root of an NMFS-mounted file system,
the command hung.
Fix
When an NFS client attempted to look up a file handle for a file or directory that didn't exist in the root of
an NMFS file system, the Data Mover returned a file handle, incorrectly. This no longer occurs.
Service Request
14760079
Fixed in version
5.5.22.0
Replication

Impact Level
Severity 1
Symptom
Description
User created a PFS, and generated continuous I/Os from the client side script. User then created 64
checkpoints and randomly deleted the checkpoints continuously. The Data Mover panicked with the
following assertion failure: 'trailer->FsVolBlockNumber ==
mrgSt->fromTOC[mrgSt->fromIndex]' on line 6259 in file
'../DeltaVol.cxx' >>PANIC in file: ../assert.c at line: 23 :
assert failed
Fix
Fixed issues found internally involving out-of-order delete of checkpoints.
Fixed in version
5.5.22.0
413
Version 5.5.22.2
Impact Level
Severity 1
Symptom
Description
During fs_copy, the Data Mover panicked with the following DART panic/fault message:
*** Page Fault Interrupt. Virt ADDRESS: 6e98a3 Err code: 2
Target addr: ffe64cba ** .
waitForReboot+0x8b;
fault_dump+0x5e;
handle_fatal_fault+0xa; _
ZN9BtreeHash9findEntryEP9BtreeSnap8NodeTypejP1;
_ZN13Btree_Read_CB4doneEv+0x5ba;
_ZN9BtreeSnap8lookupCBEP13Btree_Read_CB+0x45;
_ZN13PagedBlockMap10GetMappingEjRj+0x91;
_ZN19PoolManagedDeltaVol10GetMappingEyRy+0x51
Fix
The panic happened when a differential fs_copy and a checkpoint out-of-order delete both attempted
to reference the same object. The out-of-order checkpoint delete released an in-memory object at the
same time as the differential fs_copy attempted to access the same object. The object is now locked
until the fs_copy is finished.
Service Request
14962420
Fixed in version
5.5.22.0
SRDF
414

Impact Level
Severity 2
Symptom
Description
During a TimeFinder restore and farcopy backup create operation, the symapi.log file showed a
SYMAPI_C_CANT_RESTORE_PROTECT error related to the backed operation.
Fix
When SYMAPI_C_CANT_RESTORE_PROTECT error is returned, the device pairs are in a

transient state and this is considered a warning. The application now retries the same symapi control
command after a small sleep for max of 12 retries.
Fixed in version
5.5.22.0
Version 5.5.22.2
Storage
The following entries describe fixed issues related to Storage:
Impact Level
Severity 1
Symptom
Description
When writing on secondary file system volume with PlaybackFilter inserted, the I/O timed out in the
volume layer which resulted in a Data Mover panic with the following
>>PANIC in file: ../sched.cxx at line: 802 : SYSTEM WATCHDOG".
waitForReboot+0x8b
fault_dump+0x67
PANIC+0x2b
system_watchdog+0x7c
_t_nmi+0x63
_ZN9RawIO_IRP7irpDoneEv+0x53
_ZN8SgIO_IRP7irpDoneEv+0x57
_ZN14CamDisk_Device8completeEP10ccb_header+0x16
Fix
Removed deadlock condition that occurred when irp timeout has expired. If I/O Lost occurs in volume
layer, instead of panicking due to SYSTEM WATCHDOG, the server would panic with the string
reporting I/O Lost and report last volume reached.
Fixed in version
5.5.22.0
UFS
Impact Level
Severity 1
Symptom
Description
When a checkpoint of a Celerra Replicator target file system was being accessed through .ckpt, the
Data Mover panicked with the following DART panic/fault message:
readBlock: Bad Block Index. The target file system was not
corrupt and one of the last messages logged to the server log
indicated that it was being unmounted for playback:
CFS: 4:
fs 0x27 being unmounted. Waiting for quiesce ...
Fix
Protection was added against race conditions during an unmount or freeze of a file system while it is
being accessed through .ckpt.
Service Request
14802352
Fixed in version
5.5.22.0
415
Version 5.5.22.2
416
Impact Level
Severity 1
Symptom
Description
When an administrator disabled user or group quotas on a file system that had a paused MPD
conversion, the Data Mover panicked with the following DART panic/fault message:
waitForReboot+0x8b;
fault_dump+0x5e;
PANIC+0x2b;
_t_nmi+0x63;
_ZN9UFS_Vnode10blocksReadEjjRiPP13Buffer_Cache;
_ZN14pax_cmd_thread10readBlocksEiRP4msgbmx+0x1;
_ZN14pax_cmd_thread9wr_rdfileEP5ARCHDiP11MsgLi.
Fix
When the MPD translation thread needs to check if quota space is available to complete a translation,
it now verifies that the quota descriptor pointers are valid before using them.
Service Request
14443983
Fixed in version
5.5.22.0
Impact Level
Severity 1
Symptom
Description
User was unable to export a new tree quota if another tree on the same file system had been deleted
but its associated export was not removed. During the check for conflicts in the export table, the stale
handle for the missing path caused the export operation to fail.
Fix
Changes were made so the export operation completes even when a stale handle for the missing path
is encountered.
Service Request
15125123
Fixed in version
5.5.22.0
Version 5.5.21.4
Impact Level
Severity 2
Symptom
Description
While accessing CDMS or DHSM file systems, user saw messages such as "Offline
attributes are badly formatted for handle " in the server log. Running
FSCK did not correct this problem.
Fix
FSCK now corrects this condition.
Fixed in version
5.5.22.0
Version 5.5.21.4
Celerra
Management
Applications

Impact Level
Severity 2
Symptom
Description
Users were unable to create multi-level mountpoints (for example, /new1/new2) from the WebUI. The
validation code for this field restricted the user from entering the second "/". The following error was
generated:
Invalid mountpoint name.
Fix
WebUi code was changed and now allows the user to enter multi-level mountpoints.
Service Request
14297479
Fixed in version
5.5.21.0
Impact Level
Severity 2
Symptom
Description
If the user altered any of the Data Mover hostnames, the code that calculates the enclosure value
failed to calculate the correct value. This resulted in a failure to correctly list components, component
status and location strings.
Fix
The enclosure calculation routine has been modified so it uses the Data Mover slot numbers when
calculating the enclosure indexes. These slot numbers can not be altered.
Fixed in version
5.5.21.0
417
Version 5.5.21.4
Impact Level
Severity 2
Symptom
Description
Previously, if '/nbsnas' exceeded space limit (67% usage) but '/nas' didn't, the WebUI would keep
sending an alert every 10 minutes (as long as this condition was maintained on the Control Station).
This was inconvenient as the user should be notified of this only once. Also, these frequent alerts filled
up the 'alert_log' sooner, so user would find it difficult to locate other more important alerts.
Fix
Now, when this condition appears on the Control Station, the WebUI will log the alert only once and
not repeatedly.
Fixed in version
5.5.21.0
CIFS
418

problems:
Impact Level
Severity 1
Symptom
Description
A design limitation in the original implementation of ACLs for CIFS created a window which allowed
duplicate ACLs to be created in the Data Mover. This did not cause any problems for security, however
it allowed a situation where some files created their own copy of the ACL rather than link to the existing
copy of the ACL. When the ACL database was nearly full, or actually filled, CIFS users were unable to
access some shares, files or directories. They were unable to create new files or directories, or modify
permissions on existing files and directories. There were also some signs of this condition adding to
the load on the CPU, thus leading to a performance impact.
Fix
The mechanism for managing ACLs on the file system has been redesigned. Rather than keeping an
LRU (least recently used) cache of the most recently used ACLs in memory, the Data Mover now
keeps a hash list of all ACLs on disk in memory. This means that the opportunity for a cache miss has
been eliminated which was the reason for duplicate ACLs being created.
Service Request
13830658
Fixed in version
5.5.21.0
Version 5.5.21.4
Impact Level
Severity 1
Symptom
Description
A CIFS client was unable to reconnect to a Data Mover after several reboots of the Data Mover for an
unrelated reason. This problem occurred because the Data Mover changed its machine account
password faster than replication of passwords between domain controllers. When clients
authenticated through a domain controller that had not yet been updated, it received out-of-date
information for the Data Mover and failed.
Fix
Machine account passwords are no longer updated after each reboot by default. The Data Mover only
updates its machine account password based on the cifs.svrpwd.updtMinutes
parameter setting. The Data Mover can change passwords up to the number of times defined in the
cifs.svrpwd.maxhistory parameter (default of 2) between domain controller replications
without causing this issue.
Fixed in version
5.5.21.0
Impact Level
Severity 1
Symptom
Description
Because of a CIFS memory leak, a 0% CPU idle condition can be observed on Data Movers when an
internal thread is launched periodically to defragment memory. The Data Mover panicked with the
>>PANIC in file:../dskdump.cxx at line:1437 :Performance
issue.
waitForReboot+0x8b;
fault_dump+0x5e;
PANIC+0x2b;
_Z11paniccfgcmdR6Option+0x178;
_ZN11cfgcmd_desc6invokeEPKcii+0x124;
_Z10xml_cfgcmdP7macData+0x282;
_ZN13macCommand_DB6invokeEPKcP7macData+0x74;
_Z13mac_DocParserP7macData+0x19
Fix
This memory leak condition has been fixed.
Service Request
14526192
Fixed in version
5.5.21.0
419
Version 5.5.21.4
Impact Level
Severity 1
Symptom
Description
CIFS authentication from a Windows 2003 server failed and resulted in the following lines being
written to the server log:
2006-03-18 11:58:13: SMB: 3:[DKDENSX07_VDM] MS_PAC::build
PAC buffer[1] unknown type=11
2006-03-18 11:58:13: SMB: 3:[DKDENSX07_VDM] Cannot build
kerbPAC Client=xx.xx.xxx.xxx Asn1Error=xxxxxxxxxx origin=0x80
Fix
The code now supports additional Kerberos structures that can be used by Windows 2003 and will
prevent these connection failures. It also does better error handling when unknown Kerberos
structures are sent to the Data Mover.
Service Request
14858693
Fixed in version
5.5.21.0
Impact Level
Severity 1
Symptom
Description
>>PANIC in file:../malloc.cxx at line: 308 : Memory corrupted.
However, this was not a memory corruption but a double free issue. This occurred when trying to
reconnect to a CIFS share - the TCP connection succeeded but the negotiate failed with (negotiate
error status=c000020c). When trying to reconnect once again, the encryptionKey was deleted but
encryptionKeyLength was not 0, thus performing a double deletion that led to this panic.
Fix
Fixed code so that when the encryption key object is deleted during a disconnection, the encryption
key length is set to 0 to prevent this panic condition. Unless there are additional circumstances, our
iteration of reconnect will work with this fix.
Service Request
14920850
Fixed in version
5.5.21.0
Control Station
420

Control Station:
Version 5.5.21.4
Impact Level
Severity 1
Symptom
Description
The nas_fs -xtend size= command allows a user to specify the size of the extension to the
file system in megabytes (M), gigabytes (G) or terabytes (T). When running the extend from the
command line, if a user specified the space for the extension as MB, GB or TB, the file system could
be extended to a huge size, consuming all available space in a pool, or else the command failed
reporting not enough space available.
Fix
The presence of the "B" caused the command to interpret the size as a hex number. This has been
fixed and only decimal values are allowed for file system size.
Service Request
14512251
Fixed in version
5.5.22.0
Impact Level
Severity 2
Symptom
Description
iSCSI services would not start when the Data Mover could not contact DNS server. The following
error was reported:
Server_3 : Input/output error Time out Server_3 : Modify
iscsi configuration with iscsi service ?on?
Fix
Code has been modified by removing DNS dependency during iSCSI start.
Fixed in version
5.5.21.0
Impact Level
Severity 2
Symptom
Description
If the system has multiple backend arrays of the same type (more than 1 CLARiiONs or more than one
Symmetrix) and file system auto-extension was performed, the extension failed. This was because
the free space calculation used the available space across all arrays. File system extend is only
supported across one array.
Fix
The code now looks at the free space only on the array where the file system is located.
Fixed in version
5.5.21.0
421
Version 5.5.21.4
422
Impact Level
Severity 2
Symptom
Description
In certain circumstances, Data Movers may automatically run an FSCK against a file system.
However, the reason for this is only logged in the server log and could be easily overwritten.
Fix
The events posted to the sys_log on the Control Station now include details of the symptom detected
which triggered the Data Mover to start the FSCK. The sys_log will not contain all the details of the
FSCK, only what triggered it.
Service Request
14870949
Fixed in version
5.5.21.0
Impact Level
Severity 2
Symptom
Description
The server_standby command failed with the following when setting up an RDF standby :
not_after: Precondition violated. [LINKED_LIST.forth]
************* Thread exception ******** In thread
Child
thread
0x91ce970 (thread id) ***************
-------------------- Class / Object
Routine
Nature of exception
Effect ----------- MOUNT_TABLE
forth @2
not_after: <00000000412E7590> (From
LINKED_LIST)
Precondition violated.
Fail
------------------ MAC_STANDBY
process_args @69
<00000000412CFAC8>
Routine failure.
Fail ----------MAC_STANDBY
parse_and_exec @25 <00000000412CFAC8> (From
MAC_EXEC)
Routine failure.
Fail --------- MAC_STANDBY
nexec_process_args @16 <00000000412CFAC8> (From MAC_EXEC)
Routine failure.
Fail ------------ MAC_STANDBY
local_init @40 <00000000412CFAC8> (From NAS_EXEC)
Routine failure.
Rescue ----------------
Fix
Corrected a problem introduced in version 5.5.19.4 which triggered this condition.
Fixed in version
5.5.21.0
Version 5.5.21.4
Impact Level
Severity 2
Symptom
Description
Celerra Manager displayed outdated status for CLARiiON components like write cache, or disks.
Fix
A race condition existed between the process which polls the CLARiiON status and the Celerra
Manager caching of the status. This race condition was removed to ensure the outdated status is not
re-cached.
Fixed in version
5.5.21.0
CORE

Data Movers:
Impact Level
Severity 1
Symptom
Description
During a NAS code upgrade to version 5.3.13.1 or later, the Data Movers would not boot past reason
code 3. A 32 bit to 64 bit conversion caused the Data Movers to overwrite an area of memory in use
during the boot process. The issue was dependent on the memory usage on the Data Mover, the
number of LUNs assigned and in use, and the number of interfaces configured on the Data Mover.
Therefore, the problem may or may not have been seen.
Fix
The problem was caused by a code error that has been fixed.
Service Request
12320524
Fixed in version
5.5.21.0
Impact Level
Severity 2
Symptom
Description
During quota check, if usage of user or group crossed 4TB, an error message was not displayed.
Fix
An event is now logged in CFS log whenever user or group quota usage crosses 4TB.
Fixed in version
5.5.21.0
423
Version 5.5.21.4
424
Impact Level
Severity 1
Symptom
Description
An invalid RPC message was received by the Data Mover which caused a panic with the following
*** GP exception. Virt ADDRESS: 1fc132. Err code: 0 ***.
_fault_dump+0x56;
_t_machine_check+0x36;
_WriteSrvp+0x137;
_STRservice__FiPv+0x16b;
_start__15SchedRT_Initial+0x104;
_Sthread_startThread_internal+0x11
Fix
The RPC messages are now verified and rejected if they are invalid.
Service Request
14831213
Fixed in version
5.5.21.0
Impact Level
Severity 1
Symptom
Description
While trying to create a file system, the mount command failed and the Data Mover panicked with the
mangled directory entry.
waitForReboot+0x8b;
fault_dump+0x5e;
PANIC+0x2b;
_ZN14UFS_FileSystem28markCorruptedFsForUnmount;
_ZN9UFS_Vnode6dirbadEmPKc+0x71;
_ZN20UFS_DirHashInfo_dir3C1EP11Local_Vnode+0x5;
_ZN15UFS_DirOps_dir311initDirHashEP9UFS_Vnode+;
_ZN15UFS_DirOps_dir37scandirEP9UFS_VnodePK18Fi.
Fix
The code no longer panics if a file named /etc exists in the root of a file system. This is not supported
and when it occurs, code logs an error in the server log.
Service Request
14446064
Fixed in version
5.5.21.0
Version 5.5.21.4
Impact Level
Severity 1
Symptom
Description
The server_standby -activate command failed with the following:

Execution failed: Segmentation violation
[NAS_SIGHNDLR.eiffel_handlr].
This occurred if disabling the original primary Data Mover returned an error. The Eiffel trace happened
because the exit path of failover (server_standby activate/restore) used an uninitialized string to
display the command state.
Fix
The command will no longer generate an Eiffel dump. Code has been modified to ensure the display
string is used in its initialized boundary. Failover will now succeed, and both Data Movers will be
rebooted.
Service Request
14057715
Fixed in version
5.5.21.0
Impact Level
Severity 1
Symptom
Description
When ufslog system log device was running close to full, the Data Mover panicked with the following
>>PANIC in file: ../ufsvfsop.cxx at line:149 : watchdog: no
progress reducing dirty list.
waitForReboot+0x8b;
fault_dump+0x5e; P
ANIC+0x2b;
_Z17swatchdogCallbackP5timer5tmarg+0x102;
_Z13RTtimer_schediPv+0x77;
_ZN15SchedRT_Initial5startEv+0xf6;
Sthread_startThread_internal+0xe.
Fix
Under heavy load, the file system log device could become close to full. When this happens, any
thread that needs to write to log device, waits until there is enough space to do so. If the waiting thread
holds an ACL database lock, it can deadlock other threads trying to modify ACL database. With this
fix, all threads first acquire the logging space and then the ACL database lock. This will avoid the
deadlock panic.
Fixed in version
5.5.21.4
425
Version 5.5.21.4
FileMover

FileMover:
Impact Level
Severity 2
Symptom
Description
While running an ACL check, the Data Mover panicked with the following DART panic/fault message:
0 length offline inode.
_fault_dump+0x56;
_PANIC+0x24;
_markCorruptedFsForUnmounting__14UFS_FileSyste;
_opaqueDataLength__9UFS_Vnode+0x1d;
_typeMorph__10DHSM_Vnode+0x3c;
_initNode__10DHSM_Vnodei8dirTypes+0x3a;
_findExistingNode__14UFS_FileSystemUlR11File_S.
426
Fix
This problem occurred as a result of corruption on the file system leaving some inodes marked as
OFFLINE but still referencing data blocks on the file system. The OFFLINE bit is set in a mode 0 inode
only through corruption that fsck cannot and should not fix. This fix will prevent aclchk panicking on
this condition, and an informational message will be printed in the server_log :
"VNON offline inode %d\n". The file system can be returned to
production.
Service Request
14778968
Fixed in version
5.5.21.0
Impact Level
Severity 2
Symptom
Description
If a DHSM secondary server became unavailable, possibly because it was shut down, there were
network problems, or its IP address failed, fs_dhsm -i commands failed with the error message:
Problem communicating with DART.
Running this command several times may have caused all commands against Data Movers to fail with
a similar error message for approximately 30 minutes.
Fix
The code no longer tries to contact DHSM secondary servers for 25 minutes before responding. The
code times out and returns expected output to the user within 30 seconds. Also, the fs_dhsm -i
command no longer generates multiple calls to the Data Mover, and should run faster overall.
Service Request
13409868
Fixed in version
5.5.2.10
Version 5.5.21.4
Impact Level
Severity 2
Symptom
Description
When doing intensive open/read of files on the DHSM or CDMS secondary server, there may have
been regular interruptions of the traffic for 20 seconds decreasing performance. The CIFS client
wrongly assumed that the header of the SMB response could not be fragmented among several
network frames.
Fix
If the received SMB header is fragmented, code now compacts it before processing it.
Fixed in version
5.5.21.0
Hardware
The following entries describe fixed hardware-related problems:
Impact Level
Severity 1
Symptom
Description
Memory errors were incorrectly reported as CPU module errors.
Fix
Individual memory module failures are now reported correctly.
Fixed in version
5.5.21.0
Installation/
Upgrades

upgrades:
Impact Level
Severity 1
Symptom
Description
Setup_enclosure were unable to discover all the switches during install and failed. The following error
was shown on the serial console, and was also be logged in the install log:
Setting up the Enclosure...ERROR setting up the enclosure...
Executing -rebuildConfig option discoverOnly mode
System discovery on both subnets ... OK
Rebuilding ENCL_DB & DHCPD_CFG (via live discovery) ...
Error: T2NETUTIL - Enclosure ID range error or Bad IP address
"123.456.7.890"
Error: Found Mgmtswitch-B (ID #0) MAC address mismatch
Error: rebuildConfig FAIL (EMISMATCHMAC)
Error: REBUILDCONFIG_CMD retval = -5 (EMISMATCHMAC)
Script was stopped due to Error!
ERROR - Operation Timed Out - Expecting 'Done'- Element 41
Fix
We fixed the problem in setup_enclosure so that the Data Mover or blade enclosures are detected
correctly and the command no longer fails.
Fixed in version
5.5.21.0
427
Version 5.5.21.4
iSCSI

Impact Level
Severity 2
Symptom
Description
Aclchk during FSCK truncated the ACL DB files in order to shrink them. This truncation caused phase 3
of FSCK to report errors in the file system even though there was no actual corruption.
Fix
Added code to skip truncation of ACL DB files if there is no corruption in the ACL database and the file
sizes are below their max values of 16MB and 1GB respectively.
Fixed in version
5.5.21.1
NDMP/Backup
428

Impact Level
Severity 1
Symptom
Description
While using NDMP and doing file-by-file restore, the presence of a VBB (volume-based backup) file in
the .vbbtemp directory led to a Data Mover panic with the following DART panic/fault message:
*** GP exception. Virt ADDRESS: a789de. Err code: 0 ***.
waitForReboot+0x8b;
fault_dump+0x67; handle_fatal_fault+0xa;
_ZN14vbbFileRestore10flushFilesEv+0x1a1;
_ZN14vbbFileRestore11getMetadataEv+0x139;
_ZN14vbbFileRestore12startRestoreEv+0x170;
_ZN10vbbSession9doRestoreEv+0x9f;
_ZN10vbbSession5startEv+0x9f.
Fix
To prevent this panic, code now deletes these files from the .vbbtemp directory.
Fixed in version
5.5.21.0
Version 5.5.21.4
Impact Level
Severity 1
Symptom
Description
After many days of running NDMP backups, and there were some failed backups, the Data Mover ran
out of memory and panicked with the following DART panic/fault message:
>>PANIC in file: ../bufpool.cxx at line: 309 : couldn't get a
free page.
waitForReboot+0x8b;
fault_dump+0x5e;
PANIC+0x2b; _ZN18Buffer_Cached_Pool17makeCachedBuffersEi+0x;
_ZN19Memory_AddressSpace16makeCachedBufferEv+0x;
_ZN18Buffer_Cached_Pool15getAnyCachedBufEji+0x8;
_ZN14UFS_FileSystem11allocBufferEv+0x3f;
_ZN14UFS_FileSystem10readBlocksEliRiPP13Buffer_ .
Fix
Fixed the memory leak condition that led to this Data Mover panic.
Service Request
14695663
Fixed in version
5.5.21.0
Impact Level
Severity 1
Symptom
Description
The Data Mover panicked on allocating a large chunk of memory (45KB) when processing
NDMP_CONFIG_GET_FS_INFO message. After many days of run time, the Data Mover didn't have
enough contiguous page of memory, and panicked with the following DART panic fault message:
waitForReboot+0x8b;
fault_dump+0x5e;
PANIC+0x2b;
malloc+0x89;
_Znwm+0x15;
_Znam+0xc
Fix
Code now uses reserved memory for config_get_fs_info command to prevent this panic.
Service Request
14936598
Fixed in version
5.5.21.0
429
Version 5.5.21.4
430
Impact Level
Severity 1
Symptom
Description
NDMP backups failed. The following message was logged multiple times in the server_log:
cifs sd length for /filename_path too long 11858
Fix
Increased the CIFS SD (security descriptor) buffer length to handle 32K characters.
Service Request
15065536
Fixed in version
5.5.21.0
Impact Level
Severity 1
Symptom
Description
While running NDMP DAR restores, the restore process hung and messages similar to the following
were repeated in the server log:
: NDMP: 3: DAR (xxxx): seek failed (../ar_subs.cxx: 2541)
: NDMP: 3: Mover failed to seek to image offset (-1).
: NDMP: 3: PaxDataModule failed to seek position
Fix
This problem was caused by a non-recoverable media error that occurred while writing to tape.
Restore is now terminated immediately and messages are logged to indicate that a problem has
occurred.
Service Request
14649004
Fixed in version
5.5.21.0
Impact Level
Severity 1
Symptom
Description
During NDMP operations, a Data Mover experienced memory loss which led to a Data Mover panic
with the following DART panic/fault message:
waitForReboot+0x8b; fault_dump+0x67;
PANIC+0x2b; system_watchdog+0x7c;
_t_nmi+0x63; _Znwm+0x15;
_ZN8Scsi_Xpt13xpt_ccb_allocEv+0xcd;
Fix
This memory leak condition has been fixed.
Fixed in version
5.5.21.0
Version 5.5.21.4
Networking
Impact Level
Severity 1
Symptom
Description
When the code tried to abort an already completed fibre channel I/O, the Data Mover panicked with the
>>PANIC in file: ../dart/dart_debug_util.cxx at line: 29 :
fc_osw_dart_panic.
waitForReboot+0x8b;
fault_dump+0x5e;
PANIC+0x2b;
_osw_panic+0x32;
fc_exm_send_sequence+0x48;
fc_exm_restart_init_tcb+0x100;
fc_exm_abort_exchanges+0x576;
fc_osw_start_io+0x2d9.
Fix
Code now prevents the driver from processing I/O requests with bad context.
Service Request
14736952
Fixed in version
5.5.21.0
Impact Level
Severity 2
Symptom
Description
>>PANIC in file: ../sched.cxx at line: 722 :
SYSTEM WATCHDOG.

waitForReboot+0x8b;
fault_dump+0x5e;
PANIC+0x2b;
_t_nmi+0x63;
_ZN12SelectStream6selectEiP8LinkListS1_m+0x3a;
_ZN19NdmpEventDispatcher8dispatchEim+0x353;
_ZN21NdmpConnectionHandler17connectionHandlerE.
Fix
Decreased the size of queue for incoming packets for the network device so the system will not timeout
while trying to process incoming packets and cause a watchdog panic.
Service Request
14603745
Fixed in version
5.5.21.0
431
Version 5.5.21.4
Impact Level
Severity 2
Symptom
Description
With some network applications, write performance to Celerra was poor. This occurred with Windows
clients using iSCSI, and would be more likely if there are packet drops on the network.
Fix
The code did not generate TCP acknowledgements frequently enough if clients were not emptying their
TCP buffers immediately. Acknowledgements are now sent regularly.
Service Request
14103049
Fixed in version
5.5.21.0
Impact Level
Severity 1
Symptom
Description
_fault_dump+0x56;
_PANIC+0x24;
_system_watchdog+0x88;
_t_nmi+0x62;
_Sthread_startThread_internal+0x11
openLockCheck__8AppLibNTUsP8fid_infoUlUlUlRUsP13oplockRequest
PUsR10extparam_t, CIFS
Fix
Corrected a locking problem that could lead to corruption of UDP stream information and led to this
panic.
Service Request
14523884
Fixed in version
5.5.21.0
NFS
432
Version 5.5.21.4
Impact Level
Severity 1
Symptom
Description
On heavily loaded systems, CIFS clients experienced intermittent access when trying to access the
server and when using a large group file.
Fix
Reduced contention for a lock that was serializing access to group file for all CIFS threads.
Service Request
14834075
Fixed in version
5.5.21.0
Impact Level
Severity 1
Symptom
Description
In high-usage environments, Data Movers stopped responding to NFS lock requests. This only
occurred if file systems were mounted on the Data Mover in "uncached" mode.
Fix
A deadlock could occur in rare conditions that would prevent locks from being processed and
completing. The problems leading to the deadlock have been fixed.
Service Request
14376473
Fixed in version
5.5.21.0
Replication

Impact Level
Severity 2
Symptom
Description
A Data Mover failover timed out while trying to recover an inactive IP Replication session.
Fix
On failover, recovery will no longer be attempted on replication sessions which were already inactive
prior to the failover.
Service Request
14691368
Fixed in version
5.5.21.0
433
Version 5.5.21.4
Impact Level
Severity 1
Symptom
Description
Under heavy load conditions on a file system with checkpoints, a Data Mover panicked with the
Assertion failure: 'status==FALSE' on line 446 in file
'../RedoLog.cxx' >>PANIC in file: ../assert.c at line: 23 :
assert failed.
waitForReboot+0x8b;
fault_dump+0x5e;
PANIC+0x2b;
_assert+0x53;
_ZN7RedoLog18RecordBTransactionE18BlockTransac;
_ZN12LogFilterVol20DoWriteBlockInternalEP10Dis;
_ZN20LogFilterIO_CallBack4doneEv+0x1a9;
_ZN7RedoLog18RecordBTransactionE18BlockTransac
434
Fix
Code has been changed to avoid buffer overflows in heavy load situations.
Service Request
14760189
Fixed in version
5.5.21.0
Impact Level
Severity 2
Symptom
Description
When a Data Mover failover was not successful, IP Replication sessions could become inactive when
the original Data Mover resumes operation.
Fix
When failover fails to configure the standby Data Mover, the Control Station will send a shutdown
request before sending the reboot request to the standby Data Mover. The shutdown step was added
so that DART can flush any panic handler data on the standby Data Mover before it is rebooted.
Service Request
14691368
Fixed in version
5.5.21.0
Version 5.5.21.4
Impact Level
Severity 1
Symptom
Description
Data Mover failover did not complete when an invalid checkpoint was detected.
Fix
We now post a message to the /nas/log/sys_log if a checkpoint fails to mount, and the failover will
succeed.
Service Request
14760189
Fixed in version
5.5.21.0
Impact Level
Severity 1
Symptom
Description
*** GP exception. Virt ADDRESS: a6a3a9. Err code: 0 *** .
waitForReboot+0x8b;
fault_dump+0x5e;
_ZN11HyperVolume14notifyChildrenEN7LVolume8noti;
_ZN14PlaybackFilter14notifyChildrenEN7LVolume8;
_ZN11HyperVolume14notifyChildrenEN7LVolume8not;
_ZN7LVolume5closeEv+0x4a; _ZN9IO_Object5closeEv+0x11.
This panic didn't occur when there were two checkpoints for each PFS.
Fix
The race condition that caused the Data Mover to panic has been eliminated. This occurred when a
PFS volume had only a single checkpoint which was closed while being refreshed.
Service Request
14401422
Fixed in version
5.5.21.0
435
Version 5.5.21.4
Impact Level
Severity 1
Symptom
Description
The Data Mover panicked with the following messages in the server log:
: SVFS: 4: 107:D36107_16:Merge hdr=1560806400 prev=1560544256
id=36 chunk=145 stableEntry=15412
Fix
Increased the number of index slots used by each checkpoint to accommodate checkpoints that do not
get refreshed often. This will prevent the system from running out of index slots allocated to each
checkpoint.
Service Request
14837737
Fixed in version
5.5.21.0
Impact Level
Severity 1
Symptom
Description
*** GP exception. Virt ADDRESS: a6a3a9. Err code: 0 *** .
waitForReboot+0x8b;
fault_dump+0x5e;
_ZN11HyperVolume14notifyChildrenEN7LVolume8noti;
_ZN14PlaybackFilter14notifyChildrenEN7LVolume8;
_ZN11HyperVolume14notifyChildrenEN7LVolume8not;
_ZN7LVolume5closeEv+0x4a; _ZN9IO_Object5closeEv+0x11.
This panic didn't occur when there were two checkpoints for each PFS.
436
Fix
We eliminated the race condition that caused the Data Mover to panic. This occurred when a PFS
volume had only a single checkpoint which was closed while being refreshed.
Service Request
14401422
Fixed in version
5.5.21.0
Version 5.5.21.4
Impact Level
Severity 1
Symptom
Description
The memory allocated in checkpoint shadow's readDirectory() call was not released. This caused a
memory leak which led to a Data Mover panic with the following DART panic/fault message:
>>PANIC in file: ../addrspac.cxx at line: 431 : Out of
memory.
waitForReboot+0x8b;
fault_dump+0x5e;
PANIC+0x2b; _ZN19Memory_AddressSpace10allocPagesEi12Memory;
_Z11more_memoryiP9free_list+0x4e; malloc+0x89;
_ZN9UFS_Vnode12getAclLockedER11File_StatusRmS2;
_ZN9UFS_Vnode6getAclER11File_StatusRmS2_S2_+0x
Fix
Code now releases the allocated memory before returning from readDirectory() to prevent this
condition.
Service Request
14782945
Fixed in version
5.5.21.0
Impact Level
Severity 1
Symptom
Description
On a system configured for MPFS and Replication, if the Replication was inactive and the
/nas/sbin/uc_config -convert start jp-euc1.txt -mover server_4
command was issued, the following error was returned:
You are using the translation file (jp-euc1.txt), which will
translate. It is recommended that you back up your file system
before running this command. Have you backed up your file
system? Please make sure you are using the right translation
file, otherwise data on the disk will be corrupted and
unrecoverable! Do you wish to continue? [yes or no]:. If the
user answers yes, the following error will occur followed by a
panic: server_4 : Error 4020: server_4 : Input/output error
Timed out.
Fix
Changed MPFS code to check the status of Replication (active/inactive) and return the correct value.
Fixed in version
5.5.21.0
437
Version 5.5.21.4
438
Impact Level
Severity 1
Symptom
Description
When many (ten or more) checkpoints were scheduled to run every two or three hours, after three or
four days, they stopped running and could not be restarted.
Fix
Schedule checkpoint code was changed so that two processes wouldn't collide in the internal tables
which caused mismanagement of the table's cursor - leading to checkpoints being unable to be
scheduled and run.
Fixed in version
5.5.21.0
Impact Level
Severity 2
Symptom
Description
If a file system is exported from the Celerra without root access (neither root=IP nor anon=0 is used),
then root user from clients gets "group" permissions instead of "other" permissions. This is the
intended behavior and is the same as Linux.
Fix
Added a param nfs.rootSquashAllGids. When the param is on, NFS will convert all GIDs of
superuser to nobody if UID is also set to nobody. This implementation makes Celerra behave like
Solaris. By default, the param is off.
Service Request
12844187
Fixed in version
5.5.21.0
Impact Level
Severity 3
Symptom
Description
When multiple WINS servers were added on the CIFS configuration page of the WebUI, using comma
for separation, the entry was displayed as a single entity on the CIFS Server wizard. When the entry
was selected for editing, the two servers could be replaced with a single one only.
Fix
The code was erroneously looking for a colon to separate the WINS servers instead of a comma. Code
now looks only for a comma.
Fixed in version
5.5.21.0
Version 5.5.21.4
Impact Level
Severity 1
Symptom
Description
During checkpoint recovery, under rare circumstances, it was possible for the Data Mover to create two
checkpoints for different file systems with the same timestamp, resulting in a name collision in the
(DART) volume database. When this occurred, checkpoint recovery prematurely terminated, and
checkpoints older than the one with the duplicate timestamp were lost.
Fix
Code has been modified to ensure volume name uniqueness.
Fixed in version
5.5.21.0
SRDF

Impact Level
Severity 2
Symptom
Description
Timefinder related commands failed with

SYMAPI_C_NO_DEVICES_SELECTED or SYMAPI_C_NO_BCVS_LEFT.
Fix
This problem was corrected by applying latest symapi patch.
Fixed in version
5.5.21.0
UFS
The following entries describe fixed problems encountered with the

UNIX File System:
Impact Level
Severity 2
Symptom
Description
If a file system's ACL database was corrupted, mounting the file system hung.
Fix
Fixed the infinite loop condition that occurred while initializing ACL cache due to ACL database
corruption.
Fixed in version
5.5.21.0
439
Version 5.5.20.1
Version 5.5.20.1
Celerra
Management
Applications
440

Impact Level
Severity 2
Symptom
Description
The WebUI was not responding for some pages when many pages were clicked continuously.
Fix
Fixed a Java runtime deadlock issue encountered during concurrent applet shutdown and start up.
Fixed in version
5.5.20.0
Impact Level
Severity 2
Symptom
Description
If system came out of manufacturing with faulted-resume faulted state, the WebUI did not recognize
this state. Users saw the message "Invalid Component Information" while using Celerra Manager.
Fix
Code now recognizes "Faulted - Resume Faulted" as a potential component state and if seen it will
report the component as "Faulted" and no longer report the "Invalid component" error messages.
Fixed in version
5.5.20.0
Impact Level
Severity 2
Symptom
Description
When modifying a Data Mover's NTP server via xml request, only the first line on the server list was
taken. The rest were discarded.
Fix
Code was changed to take in all servers on the list and not just the first one.
Fixed in version
5.5.20.0
Version 5.5.20.1
Impact Level
Severity 2
Symptom
Description
Celerra Manager did not display the latest information when quotas or shares are newly created,
deleted or modified. Seven hundred user quotas were created using the command line interface, but
only 566 of them were displayed in the WebUI.
Fix
Celerra Manager has been fixed to always display the latest quotas and shares when added, deleted
or modified via the command line interface.
Fixed in version
5.5.20.0
CIFS

problems:
Impact Level
Severity 1
Symptom
Description
In a Windows 2003 environment, during a Kerberos activity, a Data Mover panicked with the following
DART panic message:
>>PANIC in file: ../malloc.cxx at line: 213 : malloc() incorrect size - not a positive value.
waitForReboot+0x8b;
fault_dump+0x5e;
PANIC+0x2b;
malloc+0x3d;
krb5_build_principal_ext+0x7c;
_Z12krb5_tgtnameP13_krb5_contextPK10_krb5_data;
krb5_get_cred_via_tkt+0xa9d;
krb5_get_cred_from_kdc+0x30.
Fix
We were using 0 as a pointer in this section of code. Code now checks the pointers for a NULL value
before using them.
Service Request
14720674
Fixed in version
5.5.20.0
441
Version 5.5.20.1
442
Impact Level
Severity 1
Symptom
Description
Following a Data Mover panic, a VDM was not mounting and none of the file systems on the VDM
could be accessed. The server_log for the Data Mover showed only the following:
CFS: 4:[whdc03vdm14]
firstOffset=0x74468
endOffset=0x74468
firstNum=0x2dab5
endNum=0x2dab5
endStamp=0x43f119d6;
Range discarded;
firstOffset=0x74468 e
ndOffset=0x74468
firstNum=0x2dab5
endNum=0x2dab5
endStamp=0x43f119d6;
Range discarded.
The nas_server -info -vdm command showed this VDM as loaded, server_cifs
showed the VDM as unloaded. server_mount showed root_fs_vdm_whdcs03vdm14 as
mounted, but a server_df did not list root_fs_vdm_whdcs03vdm14 in the output suggesting it
was really unmounted.
Fix
The security.evt file on the root file system of the VDM was corrupt. This caused an infinite loop in the
code when it was attempting to recover the file. The mechanism for recovering the .evt file has been
modified to prevent the cause of the infinite loop.C
Service Request
14674295
Fixed in version
5.5.20.0
Impact Level
Severity 1
Symptom
Description
CIFS access was lost when using CDMS or FileMover (DHSM) due to the remote server being
disconnected at the same time as a reconnection was attempted.
Fix
Code now protects disconnection processing to avoid another thread from trying to simultaneously
reconnect a share.
Service Request
14637049
Fixed in version
5.5.20.0
Version 5.5.20.1
Impact Level
Severity 1
Symptom
Description
When using a Sun One Directory Server (iPlanet) for user and group resolution, the Data Mover
>>PANIC in file: ../malloc.cxx at line: 308 : Memory
corrupted.
waitForReboot+0x8b;
fault_dump+0x5e;
PANIC+0x2b;
free+0x69; _
ZdlPv+0x10; _
ZdaPv+0xcl
_ZN23cLsaRpc_MultipleRequestD1Ev+0xa8;
_ZN20NT_Access_Credential16NLreadGroupNamesEP1.
This happened when multiple threads queried the domain for user and group objects at the same time.
Fix
The LdapClient objects are no longer shared between threads so the potential of memory corruption
has been eliminated.
Service Request
14634017
Fixed in version
5.5.20.0
Impact Level
Severity 2
Symptom
Description
Messages like the following were being logged to the server log.
SMB: 3: Cannot get attributes for directoryname
(81=FileNotCovered)
Fix
The directory name in the message is actually a DFS link. When a CIFS client requested the contents
of the directory containing this link, an attempt was made internally to get the attributes. The internal
get attributes request is no longer made on a path not covered by the server.
Service Request
14685072
Fixed in version
5.5.20.0
443
Version 5.5.20.1
Impact Level
Severity 1
Symptom
Description
The creation of DFS links failed and generated the following error in the server log:
error 1359 : ERROR_INTERNAL_ERROR.
The Data Mover panicked with a header similar to :
Assertion failure: 'linksCount == 0' on line 1878"
Fix
Code was changed to fix the panic - when an error occurs when creating DFS links, a cleanup of the
DFS erroneous link is performed.
Fixed in version
5.5.20.0
CORE
444

Data Movers:
Impact Level
Severity 1
Symptom
Description
When configuring replication, if a Virtual Data Mover (VDM) was created with the same name as an
existing file system on the destination file system, the file system was not listed on the available
destination list page of Celerra Manager. The initially created file system with the same name was not
displayed, only the newly created VDM was listed. From the CLI, both the VDM and the file systems
were listed.
Fix
Both the file system and VDM with same names are now displayed properly in the Celerra Manager.
Fixed in version
5.5.20.0
Impact Level
Severity 2
Symptom
Description
When unmounting file systems, the removal of a NFSv4 client by administrator caused server_umount
to hang.
Fix
This was fixed by suppressing an unnecessary lock when removing NFSv4 clients which caused the
deadlock.
Fixed in version
5.5.20.0
Version 5.5.20.1
FileMover

FileMover:
Impact Level
Severity 2
Symptom
Description
While using FileMover, putting a virtual tape file offline and then attempting a restore operation of a file
that was on the virtual tape caused a panic because offline virtual tape files are not readable.
Fix
Virtual tape files can now be put offline and then read by the Data Mover during a restore operation.
Fixed in version
5.5.20.0
Installation/
Upgrades

upgrades:
Impact Level
Severity 2
Symptom
Description
A NAS code 5.5.19.4 installation failed because the CLARiiON IP address setup would not accept the
IP addresses of the CLARiiON backend. It failed with the message:
You must enter a valid IP Address in dotted-decimal format.
Fix
This installation issue has been fixed. While validating the IP address, code now allows each octet to
be 0 to 255.
Fixed in version
5.5.20.1
NDMP/Backup

Impact Level
Severity 1
Symptom
Description
During NDMP backup, if there was an error opening a file and the file name contained a percent (%)
character, the Data Mover panicked with the following DART panic/fault message:
*** Divide exception. Virt ADDRESS: c13cce *** .
Fix
Code has been enhanced to allow the percent (and other characters such as *) in the file name.
Fixed in version
5.5.20.0
445
Version 5.5.20.1
Impact Level
Severity 2
Symptom
Description
During three way NDMP backups that involved multiple jobs in a schedule, the Data Mover panicked
with one of the following:
DART panic/fault message: *** GP exception. Virt ADDRESS: 1b9e86. Err
code: 0 ***. The stack trace showed: waitForReboot+0x8b;
fault_dump+0x5e; handle_fatal_fault+0xa; _end+0x1e8d147d;
CADULSELECT+0xffa0; _end+0xfe6cf45c.
1b9ed9 Err code: 0 Target addr: e9fb6c62 **. DART stack trace:
waitForReboot+0x8b; fault_dump+0x5e; handle_fatal_fault+0xa
DART panic/fault message: ***Divide exception. Virt ADDRESS: 409cd ***.
DART stack trace: waitForReboot+0x8b; fault_dump+0x5e;
handle_fatal_fault+0xa; _end+0x47341bc6; _end+0x227552fd
DART panic/fault message: *** Invalid Opcode exception. Virt ADDRESS:
477b5 ***. DART stack trace: waitForReboot+0x8b; fault_dump+0x5e;
handle_fatal_fault+0xa; _end+0x6dd29bdd
446
Fix
Previously an NDMP backup with multiple jobs in a schedule could use a stale stream handle. This led
to a condition where the stale handle was reused and that usage led to a panic. Code now removes the
stream handle from the list, which prevents improper usage.
Service Request
14164183, 14773834
Fixed in version
5.5.20.0
Impact Level
Severity 2
Symptom
Description
While using CommVault to do NDMP backups, direct access restore (DAR) failed or appeared to hang
for a backup image which was just one tape block in size.
Fix
Increased the DAR/DDAR prefetch size so that 3-way performance for one tape block works the same
as it does for large files.
Fixed in version
5.5.20.0
Version 5.5.20.1
Networking
Impact Level
Severity 2
Symptom
Description
When using TOE enabled interfaces, and VLAN tagging, network connectivity was lost. Priority bits in
the VLAN tag were being incorrectly interpreted as part of the VLAN number. Since the VLAN number
was unknown, traffic was being dropped.
Fix
The code no longer considers the priority bits to be part of the VLAN number.
Service Request
14453272
Fixed in version
5.5.20.0
Replication

Impact Level
Severity 1
Symptom
Description
Checkpoints scheduled to refresh using the WebUI checkpoint scheduler were failing with
Unable to acquire locks.
The failed checkpoint refreshes led to orphaned tasks on the Control Station which impacted
performance and memory usage, and the number of changes accumulating on the checkpoints
impacted Data Mover reboot and failover times.
Fix
Checkpoint refreshes are now serialized to prevent this condition.
Service Request
14487628
Fixed in version
5.5.20.0
Impact Level
Severity 1
Symptom
Description
Under extremely rare circumstances, if a user extended a file system which had replication or
checkpoint, and at the exact same time, the system or user issued an unmount of the file system, a
Data Mover panic occurred.
Fix
Modified code by adding a lock to protect the user list during extend for filters.
Fixed in version
5.5.20.0
447
Version 5.5.20.1
Impact Level
Severity 1
Symptom
Description
A checkpoint was refreshed shortly before upgrading from 5.3 to 5.5. After the upgrade, the Data Mover
panicked with a watchdog timeout: no progress reducing dirty list.
Fix
After a checkpoint refresh, a checkpoint merge operation is started to perform the refresh. As part of
the upgrade from 5.3 to 5.5, existing checkpoints are converted from version 2.0 to version 2.5. After
the upgrade and the checkpoint conversion, the merge is continued. There was an incompatibility
between a merge begun in 5.3 and continued in 5.5 that caused the watchdog timeout. Modified the
conversion to be aware of the merge-in-process status of the checkpoints to properly handle the
incompatibility.
Service Request
14222898
Fixed in version
5.5.20.0
Impact Level
Severity 1
Symptom
Description
When using SnapSure, a race condition led to a Data Mover panic with the following DART panic/fault
message: >>PANIC in file: ../rawirp.cxx at line: 193 : I/O not
progressing for the specified time .
waitForReboot+0x8b;
fault_dump+0x67;
PANIC+0x2b;
_ZN9RawIO_IRP13checkProgressEv+0x7c;
_ZN17IrpProgressThread5startEv+0x6;
Sthread_startThread_internal+0xe.
448
Fix
Code has been fixed to eliminate the possibility of this race condition.
Fixed in version
5.5.20.0
Version 5.5.20.1
Impact Level
Severity 2
Symptom
Description
When a differential fs_copy was started, no progress was shown until data was actually transmitted.
For a large file system the preprocessing time can take hours and lead the administrator to believe the
copy is hung.
Fix
A message will be logged in the server log for each 5% of preprocessing for a differential fs_copy. The
message will be similar to: VMCAST: 4: FSCopySender::FSid:220 Building
Differential BitMap 5% complete.
Service Request
14516362
Fixed in version
5.5.20.0
SRDF

Impact Level
Severity 2
Symptom
Description
The fs_timefinder -M off command reported an error condition, however, the process
continued and actually split the devices according to the api.log.
: Execution failed: have_symm: Precondition violated.
[SYMAPI_DG_RDF.connect]
Error 2208: :exec failed
Fix
Added the Symmetrix ID in the storage table to the name of the device group. This change corrects the
situation where multiple Symmetrix resulted in two or more device groups having the same name even
when the devices belong to different Symmetrix.
Fixed in version
5.5.20.0
Impact Level
Severity 2
Symptom
Description
After a fresh installs, the /nas/sbin/nas_rdf -init command failed with the following:
unexpected EOF while looking for matching `"'
Fix
Corrected a typo where an extra " caused the failure.
Fixed in version
5.5.20.0
449
Version 5.5.20.1
Impact Level
Severity 1
Symptom Description
The nas_rdf -init command which adds the devices to the RDF config file on the src side did
not handle return code properly.
Fix
The nas_rdf -init command tries to find the remote device for the source volumes. The command will now retry three times with 10 second interval if it cannot find the device. If after three times
the device is still not found, the script will exit with an ERROR.
Service Request
14573743
Fixed in Version
5.5.20.0
Impact Level
Severity 1
Symptom
Description
While using SRDF to do failover and failback tests, activating the target site failed because the
destination Symmetrix could not communicate with the source Symmetrix. The following error was
presented to the user:
id type acl slot groupID state name
1
1
1000 2
0
server_2
2
1
1000 3
0
server_3
server_2 :
server_2 : going offline
rdf : going active
replace in progress ...failed
failover activity complete
replace_storage:
replace_volume: volume is unreachable
root_disk,root_ldisk,d3,d4,d5,d6,d7,d8,d9,
d10,d11,d12,d13,d14,d15,d16,d17,d18,root_ldisk,root_ldisk
server_3 : Error 4003: server_3 : standby is not configured
450
Fix
Code now gets the remote mirror configuration information from the cached src side symapi DB
during activate. This eliminates the dependency on the current symapi DB for the information, as it
might not exist.
Fixed in version
5.5.20.1
Troubleshooting and getting help
Troubleshooting and getting help

EMC support, product, and licensing information can be obtained as
follows.
Product information For documentation, release notes, software
updates, or for information about EMC products, licensing, and
service, go to the EMC Powerlink website (registration required) at:
http://Powerlink.EMC.com
Technical support For technical support, go to EMC Customer

Service on Powerlink. To open a service request through Powerlink,
you must have a valid support agreement. Please contact your EMC
sales representative for details about obtaining a valid support
agreement or to answer any questions about your account.
Copyright 2007 EMC Corporation. All rights reserved.

EMC believes the information in this publication is accurate as of its publication date. The
information is subject to change without notice.
THE INFORMATION IN THIS PUBLICATION IS PROVIDED AS IS. EMC CORPORATION
MAKES NO REPRESENTATIONS OR WARRANTIES OF ANY KIND WITH RESPECT TO THE
INFORMATION IN THIS PUBLICATION, AND SPECIFICALLY DISCLAIMS IMPLIED
WARRANTIES OF MERCHANTABILITY OR FITNESS FOR A PARTICULAR PURPOSE.
Use, copying, and distribution of any EMC software described in this publication requires an
applicable software license.
For the most up-to-date listing of EMC product names, see EMC Corporation Trademarks on
EMC.com.
All other trademarks used herein are the property of their respective owners.
451
452

5 5 32 4

Caricato da

Informazioni sul documento

Descrizione originale:

Titolo originale

Copyright

Formati disponibili

Condividi questo documento

Condividi o incorpora il documento

Opzioni di condivisione

Hai trovato utile questo documento?

Questo contenuto è inappropriato?

Copyright:

Formati disponibili

5 5 32 4

Caricato da

Copyright:

Formati disponibili

Celerra

New features and changes

New features and changes

Where <movername> is the Data Mover name and

CSA version check

CSA version check prevents running the current CSA version on an

Celerra Network Server Release Notes

New features and changes

Each setting should reside on its own line. For example:

Impact of modem settings during upgrades

Upgrading from Upgrading to

Settings are not preserved,

The settings must be in the

The settings must be in the

The settings must be in the

Celerra Network Server introduced support for binding disks as

New features and changes

This section describes the features introduced in version 5.5.31.6.

Celerra Network Server Release Notes

New features and changes

If the serial number is automatically detected, the following message

If the serial number is not detected, it gets updated accordingly and

Complete details about the nas_connecthome command can be

This version of Celerra Network Server introduced support for the

New features and changes

The network topology used to access the public IP addresses of the

Celerra Network Server Release Notes

New features and changes

The Celerra Startup Assistant (CSA) tool simplifies the installation of

Verify cabling and connectivity are working properly prior to

Configure Control Station and blades

Input site-specific information and configure call-home

Create NFS Export, CIF shares, or iSCSI LUNs

Register for services

The server_iscsi command supports the following new option:

This option changes the specified LUN to read-only (-readonly yes)

You cannot use the -modify option with a promoted snapshot.

Before changing a LUN to read/write or read-only, you must

Changing a LUN to read-only while it is being accessed by a host

Celerra Network Server Release Notes

New features and changes

If the LUNs mask grants access to an iSCSI initiator and that

Do not create a replication-based clone of a LUN that is in an

To change LUN 24 (a read-only LUN) on target t1 on server_2 into a

To change LUN 11 (a read/write LUN) on target t1 on server_2 into

NS20, NS20FC and

Celerra Network Server Release Notes

New features and changes

This section describes the features introduced in version 5.5.29.1.

Balance LUNs across disk adapters/directors.

Control Volumes must be created on mirrored devices, not on

Do not build a file system on a single Symmetrix metavolume.

Do not mix multiple RAID types (mirrors/RAID 1, RAID 5, parity

AVM is not RAID1, RAID5, RAID6, RAID-S aware on

Celerra Network Server Release Notes

New features and changes

The system-defined AVM profiles for Symmetrix will attempt

If you manually create stripe volumes on Symmetrix based disk

RAID 6 performance is more sensitive to write intensive workloads

This section describes the features introduced in version 5.5.28.1.

Celerra Network Server Release Notes

New features and changes