Sei sulla pagina 1di 2

10/23/2020 Document 2334851.

1
Copyright (c) 2020, Oracle. All rights reserved. Oracle Confidential.

Solaris Volume Manager (SVM) metasync Service Fails On Boot When Used in Solaris Cluster
(Doc ID 2334851.1)

APPLIES TO:

Sun Solaris Volume Manager (SVM) - Version 11.9.0 to 11.11 [Release 11.0]
Solaris Cluster - Version 3.2 to 4.3 [Release 3.2 to 4.3]
Oracle Solaris on SPARC (64-bit)
Oracle Solaris on x86-64 (64-bit)

SYMPTOMS

During a new install of Solaris Cluster using Solaris Volume Manager for the shared storage, the 'metasync' service fails to
start following each cluster node reboot.

A message such as the following is observed:

Nov 30 15:09:53 server01 svc.startd[14]: [ID 652011 daemon.warning] svc:/system/metasync:default: Method


"/lib/svc/method/svc-metasync" failed with exit status 1.

CHANGES

This was observed on a new install.

CAUSE

The installation and configuration of the Solaris Volume Manger metadbs was done to the Solaris Cluster 'did' device. This
is confirmed using 'metadb'.

# metadb

flags first blk block count


a m pc luo 16 8192 /dev/did/dsk/d15s0
a pc luo 8208 8192 /dev/did/dsk/d15s0
a pc luo 16400 8192 /dev/did/dsk/d15s0

The service fails to start because the device containing the metadbs is not online when the start method runs. This is
confirmed by checking the failed service time with the associated Solaris Cluster 'did' SMF service logs. The metasync
service failed first:

Nov 30 15:09:53 server01 svc.startd[14]: [ID 652011 daemon.warning]


svc:/system/metasync:default: Method "/lib/svc/method/svc-metasync" failed with exit status 1.

Checking the logs for the SMF services, loaddid and initdid, shows they are started after the metasync service fails:

[ Nov 30 15:10:00 Executing start method ("/usr/cluster/lib/svc/method/loaddid start"). ]


[ Nov 30 15:10:19 Executing start method ("/usr/cluster/lib/svc/method/initdid start"). ]

SOLUTION

https://support.oracle.com/epmos/faces/DocumentDisplay?_adf.ctrl-state=82skjboz1_653&id=2334851.1 1/2
10/23/2020 Document 2334851.1

Change the configuration so the metadbs are located on Solaris devices and not a Solaris Cluster 'did' device.

Locate the Solaris device:

# cldev list d15

d15 server01:/dev/rdsk/c0t3d0

Change the metadb device to use c0t3d0s0.

This will require adding a temporary copy of the metadb, deleting the copy on d15, adding a copy to c0t3d0s0, then finally
deleting the temporary copy.

If there are no metadevices or metasets defined, the metadb can simply be deleted from d15 and added to c0t3d0s0.

# metadb -f -d /dev/did/dsk/d15s0

# metadb -a -c3 c0t3d0s0

This will allow the metasync service to start during a boot as it no longer needs the did driver and the did devices to be
present during the boot sequence.

Didn't find what you are looking for?

https://support.oracle.com/epmos/faces/DocumentDisplay?_adf.ctrl-state=82skjboz1_653&id=2334851.1 2/2

Potrebbero piacerti anche