Sei sulla pagina 1di 30

Downtime-Optimized DMO: Introduction

Using SUM 2.0 SP 04 and higher

Cloud and Lifecycle Management, SAP SE


Boris Rubarth, Product Manager, CLM

INTERNAL
Downtime-optimized DMO (1-pager)
Reducing the technical downtime: migrate big application tables in uptime (target Suite-on-HANA)

Technical downtime of a database migration (DMO run of SUM) scales with database size

➢ Downtime-optimized DMO approach moves migration of selected big application tables to uptime

➢ A delta replay mechanism ensures that any uptime changes (by end users) are considered

SAP Note 2547309 for Prerequisites and Restrictions


© 2019 SAP SE or an SAP affiliate company. All rights reserved. ǀ PUBLIC | downtime-optimized DMO with SUM 2.0 SP 04 2
Agenda

Introduction Motivation, procedure overview

Project Considerations Table selection, impact analysis

How to Enable the feature, monitor the replication

Architecture Illustration of relevant steps

© 2019 SAP SE or an SAP affiliate company. All rights reserved. ǀ PUBLIC | downtime-optimized DMO with SUM 2.0 SP 04 3
Downtime-Optimized DMO: Motivation

▪ Technical downtime of Database Migration Option (DMO) of SUM scales with database size

▪ Optimization of DMO is possible and recommended

▪ For specific projects, further downtime reduction may be required

Downtime-Optimized DMO migrates selected big application tables already in uptime

➢ Technical downtime is reduced significantly

➢ Delta on table content (end user activity) is considered by SUM

➢ Approach is not yet general available, only used for accepted pilot projects
targeting Suite-on-HANA (not SAP S/4HANA)

© 2019 SAP SE or an SAP affiliate company. All rights reserved. ǀ PUBLIC | downtime-optimized DMO with SUM 2.0 SP 04 4
Downtime-optimized DMO of SUM – procedure overview

Standard DMO of SUM


Uptime Business Downtime

Shadow Repository Migration of


application tables

Migration part Update part


Downtime
reduced
Downtime-optimized DMO of SUM
Uptime Business Downtime

Shadow Repository Migration of big


application tables

© 2019 SAP SE or an SAP affiliate company. All rights reserved. ǀ PUBLIC | downtime-optimized DMO with SUM 2.0 SP 04 5
Downtime-optimized DMO of SUM – sequence

EU_CLONE_MIG_OPTDMO_RUN
Delta depends on
Initial migration
user change rate
RUNASYN_RRC_REPLICATE

Uptime
delta migration
RUN_RRC_REPLICATE_FINAL

Downtime
delta migration
EU_CLONE_MIG_DT_RUN

Remaining
tables migration
Downtime-optimized DMO of SUM
Uptime Business Downtime

Shadow Repository Migration of big


application tables

© 2019 SAP SE or an SAP affiliate company. All rights reserved. ǀ PUBLIC | downtime-optimized DMO with SUM 2.0 SP 04 6
Trigger are set in MAIN SHD RUN
Agenda

Introduction Motivation, procedure overview

Project Considerations Table selection, impact analysis

How to Enable the feature, monitor the replication

Architecture Illustration of relevant steps

© 2019 SAP SE or an SAP affiliate company. All rights reserved. ǀ PUBLIC | downtime-optimized DMO with SUM 2.0 SP 04 7
Downtime-Optimized DMO: Project Considerations (1/2)

➢ Feature is only available for accepted service-based projects with SAP colleagues onsite
See SAP Note 2442926 for details

➢ Recommended Sequence is
➢ Run standard DMO with optimizations
➢ Determine big tables for uptime migration
➢ Run downtime-optimized DMO on copy of PRD
➢ (optional) Run downtime-optimized DMO on PRD until end of uptime, determine change rate on selected tables, reset
➢ Start downtime-optimized DMO on PRD, using one weekend for initial data transfer, following weekend for downtime

➢ Note
➢ SUM uses its own delta record and replay technology (CRR), no DMIS AddOn is required
➢ Hardware requirements: an additional instance is used on the PAS host (size like shadow instance, but additionally)

© 2019 SAP SE or an SAP affiliate company. All rights reserved. ǀ PUBLIC | downtime-optimized DMO with SUM 2.0 SP 04 8
Downtime-Optimized DMO: Project Considerations (2/2)

Additional activities compared to standard DMO project:


➢ Determination of big application tables to be migrated

➢ Estimation of change rate and possible replication rate

➢ Impact analysis

➢ SLT Trigger considerations: trigger to be re-established on target; (MaxDB as source: trigger can’t remain)

➢ Monitoring of table replication

➢ If tables are specified that can’t be replicated, an error is displayed, and the table name has to be removed from the list

© 2019 SAP SE or an SAP affiliate company. All rights reserved. ǀ PUBLIC | downtime-optimized DMO with SUM 2.0 SP 04 9
Downtime Optimized DMO: Criteria for big tables to be selected

▪ Determination of big application tables:

▪ via Report RSDMODBSIZE (see SAP Note 2153242)

▪ via SL UI after plain DMO run (menu More / Utilities -> Migration Preparation / Table Sequence)

▪ An additional, more sophisticated approach is under investigation

© 2019 SAP SE or an SAP affiliate company. All rights reserved. ǀ PUBLIC | downtime-optimized DMO with SUM 2.0 SP 04 10
Preparation: Report RSDMODBSIZE

▪ Use SAP Note 2153242 to implement the report

▪ Report shows suitable big tables

▪ Table selection can be stored in text file


-> to be provided in SUM dialog later

▪ Rough runtime estimation possible


based on migration rate from standard DMO

▪ If tables are specified that can’t be replicated,


during the SUM run an error is displayed, and the
table name has to be removed from the list

© 2019 SAP SE or an SAP affiliate company. All rights reserved. ǀ PUBLIC | downtime-optimized DMO with SUM 2.0 SP 04 11
RSDMODBSIZE with change rates

© 2019 SAP SE or an SAP affiliate company. All rights reserved. ǀ PUBLIC | downtime-optimized DMO with SUM 2.0 SP 04 12
Impact Analysis (1/2)
Why Impact Analysis is relevant

Downtime-optimized DMO includes additional activities:


➢ Changes on these table migration must be recorded via triggers

This can result in the following consequences:


➢ Change recording increases the daily database growth
➢ (MaxDB only): pre-existing triggers not allowed -> table to be removed from uptime migration

To avoid business impact on your production system, it is necessary to estimate in advance how
much additional database space will be required. The Impact Analysis approach provides an
estimation for this.

SAP Note 2402270 - Export of Table Statistics for SUM Impact Analysis
SAP Note 2481983 - SUM Impact Analysis for downtime-optimized DMO

© 2019 SAP SE or an SAP affiliate company. All rights reserved. ǀ PUBLIC | downtime-optimized DMO with SUM 2.0 SP 04 13
Impact Analysis (2/2)
HowTo execute the Impact Analysis

Procedure:
Report
➢ Create report and transport report to PRD ZRSUPG_IMPACT_ANALYSIS_EXPORT
(SAP Note 2402270)

➢ Run report and export table statistics from PRD PRD

➢ Use table statistics for DMO run on test system SUM/abap/save/ZDIMPANA.ZIP


(SAP Note 2481983)

➢ Analyze log file provided by SUM


(examples provided in SAP Note 2481983) Software Update Manager SBX
Phase
RUN_IMPACT_ANALYSIS_DODMO

SUM/abap/log/IMPANAUPG.<SID>

© 2019 SAP SE or an SAP affiliate company. All rights reserved. ǀ PUBLIC | downtime-optimized DMO with SUM 2.0 SP 04 14
Agenda

Introduction Motivation, procedure overview

Project Considerations Table selection, impact analysis

How to Enable the feature, monitor the replication

Architecture Illustration of relevant steps

© 2019 SAP SE or an SAP affiliate company. All rights reserved. ǀ PUBLIC | downtime-optimized DMO with SUM 2.0 SP 04 15
Downtime Optimized DMO: enable the option
(depending on SAPup_add.par)

Text file with table list


(one table per line)
has to be provided

© 2019 SAP SE or an SAP affiliate company. All rights reserved. ǀ PUBLIC | downtime-optimized DMO with SUM 2.0 SP 04 16
Check status on end of uptime

Link to new CRR section in SUM Utilities


(see next page)

© 2019 SAP SE or an SAP affiliate company. All rights reserved. ǀ PUBLIC | downtime-optimized DMO with SUM 2.0 SP 04 17
Replication monitoring and configuration
Part of SUM Utilities section

Transaction CRR_CONTROL
is deprecated (read only)
with SUM 2.0 SP 04 and higher

© 2019 SAP SE or an SAP affiliate company. All rights reserved. ǀ PUBLIC | downtime-optimized DMO with SUM 2.0 SP 04 18
SUM 2.0 SP 04: Replication based on TMP instance and R3load

SUM uses one approach for replication for

• downtime-optimized DMO (targeting SoH)

• downtime-optimized Conversion (targeting S/4HANA)

Non-cluster tables are replicated by an additional instance (TMP instance)

Cluster tables are replicated by R3load mechanism (with SUM 2.0 SP 04 and higher)
New kernel versions 7.53 are no longer cluster-aware, can’t read clusters.
R3load is used for cluster table replication, even for lower target releases like Suite-on-HANA

© 2019 SAP SE or an SAP affiliate company. All rights reserved. ǀ PUBLIC | downtime-optimized DMO with SUM 2.0 SP 04 19
Agenda

Introduction Motivation, procedure overview

Project Considerations Table selection, impact analysis

How to Enable the feature, monitor the replication

Architecture Illustration of relevant steps

© 2019 SAP SE or an SAP affiliate company. All rights reserved. ǀ PUBLIC | downtime-optimized DMO with SUM 2.0 SP 04 20
Downtime-optimized DMO: Procedure Start

PAS Host
SUM is started
PRD Instance
PAS
Source
Kernel
SUM

SHD Kernel TGT Kernel

PRD REP Legende

Target
Release

Application Data Source


Release

Source DB Host SAP HANA DB Host


© 2019 SAP SE or an SAP affiliate company. All rights reserved. ǀ PUBLIC | downtime-optimized DMO with SUM 2.0 SP 04 21
Downtime optimized DMO: Shadow repository

PAS Host
Uptime:
PRD Instance Shadow system
PAS is created
Source
Kernel
SUM

SHD Instance
SHD Kernel

PRD REP

SHD REP
Application Data

Source DB Host SAP HANA DB Host


© 2019 SAP SE or an SAP affiliate company. All rights reserved. ǀ PUBLIC | downtime-optimized DMO with SUM 2.0 SP 04 22
Downtime optimized DMO: Migration of shadow repository

PAS Host
Uptime:
PRD Instance
Shadow repository
PAS
is copied
Source
Kernel Later, trigger are activated by SUM
SUM
R3load R3load

SHD Instance
SHD Kernel TGT Kernel

PRD REP TGT REP


SHD REP
Application Data

Source DB Host SAP HANA DB Host


© 2019 SAP SE or an SAP affiliate company. All rights reserved. ǀ PUBLIC | downtime-optimized DMO with SUM 2.0 SP 04 23
Downtime optimized DMO: Initial Transfer of Triggered Tables
EU_CLONE_MIG_OPTDMO_RUN
PAS Host
Uptime:
PRD Instance
Initial transfer of
PAS
triggered tables
Source
Kernel
SUM
R3load R3load

SHD Instance
SHD Kernel TGT Kernel

PRD REP TGT REP


SHD REP
Application Data Application Data

Source DB Host SAP HANA DB Host


© 2019 SAP SE or an SAP affiliate company. All rights reserved. ǀ PUBLIC | downtime-optimized DMO with SUM 2.0 SP 04 24
Downtime optimized DMO: Delta migration of triggered tables
RUNASYN_RRC_REPLICATE
PAS Host
Uptime:
PRD Instance
Delta transfer of
PAS
triggered tables
Source
Kernel

* SUM
R3load R3load

SHD Instance
CRR replicator Writer
SHD Kernel TMPTGT
Instance
Kernel MAIN_MIGTMPSTART/
SUBMOD_MIGTMP_REPLICATOR_START/
RUNASYN_RRC_REPLICATE

PRD REP TGT REP


SHD REP
Application Data Application Data

Source DB Host SAP HANA DB Host * Cluster table replication via R3load
© 2019 SAP SE or an SAP affiliate company. All rights reserved. ǀ PUBLIC | downtime-optimized DMO with SUM 2.0 SP 04 25
Downtime optimized DMO: remaining migration of triggered tables
RUN_RRC_REPLICATE_FINAL
PAS Host
Downtime:
PRD Instance
Remaining Delta
PAS
transfer of
Source
triggered tables
Kernel

* SUM
R3load R3load

SHD Instance
CRR replicator Writer
SHD Kernel TMPTGT
Instance
Kernel

PRD REP TGT REP


SHD REP
Application Data Application Data

Source DB Host SAP HANA DB Host * Cluster table replication via R3load
© 2019 SAP SE or an SAP affiliate company. All rights reserved. ǀ PUBLIC | downtime-optimized DMO with SUM 2.0 SP 04 26
Downtime optimized DMO: migration of remaining tables

PAS Host
Downtime:
PRD Instance
Non-triggered
PAS
application tables
Source
are migrated
Kernel

R3load R3load

SHD Instance Writer


SHD Kernel TMPTGT
Instance
Kernel

PRD REP TGT REP


SHD REP
Application Data Application Data

Source DB Host SAP HANA DB Host


© 2019 SAP SE or an SAP affiliate company. All rights reserved. ǀ PUBLIC | downtime-optimized DMO with SUM 2.0 SP 04 27
Downtime optimized DMO: update

PAS Host
Downtime:
PRD Instance
Target kernel
PAS
for PRD instance
TGT
Kernel

PRD REP TGT REP


SHD REP
Application Data Application Data

Source DB Host SAP HANA DB Host


© 2019 SAP SE or an SAP affiliate company. All rights reserved. ǀ PUBLIC | downtime-optimized DMO with SUM 2.0 SP 04 28
Downtime optimized DMO: procedure finished

PAS Host
Downtime:
PRD Instance Application tables
PAS are updated
TGT
Kernel Procedure finished

PRD REP TGT REP


SHD REP
Application Data Application Data

Source DB Host SAP HANA DB Host


© 2019 SAP SE or an SAP affiliate company. All rights reserved. ǀ PUBLIC | downtime-optimized DMO with SUM 2.0 SP 04 29
Boris.Rubarth@sap.com

Potrebbero piacerti anche