Technical Articles
Database Migration Option (DMO) of SUM – Introduction
Scenario:
- You want to migrate your existing SAP ABAP system to the SAP HANA database
Use the database migration option (DMO) of the Software Update Manager (SUM):
it combines SAP upgrade and database migration to SAP HANA in one tool!
Benefits:
- Migration steps are simplified
- System update, Unicode Conversion (*), and database migration are combined in one tool
(*: only possible for a target based on 7.40) - Business downtime is reduced
- The source database remains consistent, so a fast fallback is possible
Motivation
If you want to migrate an existing SAP system (running on anyDB) to a SAP HANA database, required steps may be a dual-stack split, a unicode conversion, a database upgrade of anyDB, an upgrade of your SAP software, and a database migration to SAP HANA. The Software Update Manager (SUM) includes an option that combines the upgrade with the database migration “database migration option” (DMO) for SUM. It is sometimes referred to as the one-step migration procedure, compared to the classical migration (i.e. heterogenous system copy, using Software Provisioning Manager).
The DMO is an inplace-migration (instead of a new installation): it upgrades and migrates the existing system while keeping the system-ID, PAS host name, and connectivity settings stable.
DMO for SAP NetWeaver BW and for SAP Business Suite systems
DMO is available with Software Update Manager 1.0 SP09 and higher, and can be used for systems based on AS ABAP. It can be used for other target database types meanwhile as well, see respective SAP note for details (see below).
DMO processing overview
The processing sequence is based on the shadow system functionality of SUM: the SUM creates the shadow repository on the target database until downtime phase, while in parallel the SAP HANA database is setup (client, schema, …). Then the database connection of the SAP system is switched to the target database, and then the downtime starts. After migration of the application data (including data conversion), the upgrade is finalized and the SAP system runs on the target database. The source database continues to run and the application data in it are not modified, so it remains a fallback throughout the complete process.
Further information
- The respective SAP notes for SUM 1.0 and 2.0 are always listed on the SAP Support Portal page http://support.sap.com/sltoolset
Check the following blog to learn about SUM 2.0:
https://blogs.sap.com/2017/08/10/sum-in-the-family-way/
DMO Guide
- Use quicklink http://support.sap.com/sltoolset and navigate to section System Maintenance
- SAP First Guidance – Migration BW on HANA using the DMO option in SUM
Blogs on DMO
- Migration to SAP HANA: Overview Video of Database Migration Option DMO
- DMO: technical background
- DMO: background on table split mechanism
- DMO: comparing pipe and file mode for R3load
- DMO: table comparison and migration tools
- DMO: Handling table comparison checksum errors
- DMO, SAP HANA and the Target DB SID
- Short history of DMO
Blogs on other DMO use cases:
- DMO without software change
- DMO with System Move – the use case to change PAS host during DMO
- Your SAP on Azure – Part 2 – DMO with System Move
Blogs on DMO downtime aspects:
- Optimizing DMO Performance
- DMO: optimizing system downtime …
- DMO: introducing the benchmarking tool
- DMO: downtime optimization by migrating app tables during uptime (preview)
- Phases behind DMO R3load parallel export/import during UPTIME and DOWNTIME to target HANA DB
Blogs on related topics
- Migration of SAP Systems to SAP HANA
- Comparing SAP migration procedures to SAP HANA database
- A better way to migrate your SAP NetWeaver BW from any database to SAP HANA
- Decision Matrix to Choose Best Migration Option of ABAP Systems to SAP HANA
- Software Update Manager (SUM): introducing the tool for software maintenance
- Best Practice Guide – Classical Migration of SAP NetWeaver AS ABAP to SAP HANA
SAP Education offering
- HA250: “Migration to SAP HANA using DMO” – two days classroom training
Boris Rubarth
Product Management, Software Logistics, SAP SE
Hello Boris,
we have source system SAP ECC6 EHP6 unicode with ase 15.7 on HP-UX 11.31 IA 64 bit
Could we use DMO (upgrade EHP7&migration to suite on hana) with actual DB ase release?
Thanks.
Massimiliano
Hello Boris,
Any advice on migrating to HANA that is in a cluster setup? With a standby-failover node.
Is there something in particular that is different from single-node setup?
I couldn't find yet in the guides or notes any mention on this.
Thank you,
Delia
Hi, Can we use this option to migrate Non sap / ABAP system ( in this case ASE from Sybase) to HANA as a pure data base migration tool ( migrate DB objects like procedure, functions, triggers, catalog objects,etc..?
DMO works on ABAP stack systems only, not on non-ABAP systems. Regards, Boris
Thank you Boris for the blog.
We are planning to migrate to Hana from Solaris/Oracle to Redhat/Hana, both the DB and PAS. From the OSS note 2377305, it says we can install an additional Linux apps server on the source system. Therefore, the source and destination R3load will run on the same version. However, since Solaris and Linux are on different code page, would that cause an issue?
I am also not sure if installing an additional Linux apps server helps us. My understanding is that it is an option for source systems like Windows not certified with the system move DMO option? We could potentially run the Hana migration on the source Linux host without the server move and then install the PAS on the "real" target host later. However, is this a supported alternative?
Hi Jonathan,
I can think of two scenarios in which an Additional Application Server (AAS) can be relevant:
1) for plain DMO if you want to use a hardware with more performance for the time of the migration
2) as you say for "DMO with System Move" to fulfil the current requirements on linux OS
Sure it is a valid approach to first run plain DMO, and afterwards install a new PAS on the real target host.
Regards, Boris
Hi,
Nice Blog.
Can you please clarify as below.
During Classical migration when upgrading system i.e. SAP ERP 6.0 without EHP to SAP ERP 6.0 with EHP8. we have dependency to check OS, Database compatibility.
Now if planning to update and migration using DMO i.e. SAP ERP 6.0 without EHP , ORACLE 11.2 to SAP ERP 6.0 with EHP8 on HANA 2.0
Does we require to have same dependency for OS and database?
Reason for asking as Sadow instance get created on PAS which is OLD server which has oracle 11.2 and EHP8 require oracle 11.3 or 12c
Regards,
Nikunj Thaker
Hi Nikunj,
as you say, for an update (e.g. SAP ERP 6.0 without EHP to SAP ERP 6.0 with EHP8) you have to check OS/DB. This is true for a "pure" update with SUM, and true for a combined update and migration (DMO) with SUM.
Not sure why you write "During Classical migration when upgrading system ..." - for me, a classical migration is only the migration process as such, and this classical migration process itself does not include an update. Of course, typically you consider to have an update as a separate step in the same project.
Regards,
Boris
Hi Boris,
Thanks for reply.
I had wrote classical migration means without DMO.
DMO based migration when moving to HANA.
My question is when i am performing migration to HANA and update together still OS/DB dependency are applicable?
EHP8 require Oracle database 12c.
Planning update and migrate existing ERP 6.0 without any EHP with Oracle 11.2 to ERP 6.0 EHP8 with HANA DB.
As Shdown instance get created on CI server which is old (ORACLE DB) hence i requiere to update oracle database to 12c prior EHP update and hana migration?
Regards,
Nikunj Thaker
Hi Nikunj,
this is described in the respective SAP note for DMO. The relevant and actual note is listed on the landing page of the SL Toolset: http://support.sap.com/sltoolset, in section System Maintenance. As of today, it is SAP Note 2428170 for DMO with SUM 2.0 (SUM 2.0 because you target EHP 8). For target database type SAP HANA and source database type ORA, Oracle 11.2 is fine for the source db, in case of a target system based on 7.50 (EHP 8 is based on 7.50).
Regards, Boris
Hello Boris,
We successfully used SUM w/DMO to migrate our BW 7.31 AnyDB system to BW 7.5 on HANA 1.0 on Power (BE). We had great success with the product and achieved transfer rates up to 1 TB per hour in the DMO phase.
We would like to use SUM w/DMO again to apply BW 7.5 Support Packs and migrate to HANA 2.0 on Power (LE). I have heard that SUM currently does not support HANA as a source DB and my need could not be met by SUM. I want to avoid having to apply BW 7.5 Support Packs and then turn right around and complete a HANA 1.0 Big Endian => HANA 2.0 Litte Endian migration using the SDI Data Provisioning Server.
THE ASK: Use SUM w/DMO where HANA is the source DB
What can you do as a product manager for SUM w/DMO to make this a reality?
Thank you for your consideration,
Scott
Hello Scott,
thank you for your suggestion. I understand you motivation.
Nevertheless, I am sorry to tell you that running DMO on a source database SAP HANA will not be enabled.
The migration from big to little endian is a specific nonrecurring scenario for which a specific tool is offered. Enabling this with SUM would be a huge development effort, and we have important other topics with high priority to work on.
Kind regards,
Boris
Hi Boris,
We are in process of migrating our systems to HANA enterprise cloud.
Our source : CRM ehp3 , sql 2012 and windows 2012 r2.
Is it possible to do a DMO from on premise to HEC ? We know the restriction of DMO system move on windows source.
Can we do this as two step?
1.) Can we run a plain DMO from source .
2.)Now our PAS will be on premise and HANA DB on SAP cloud at the end of the DMO process.
3.) Then uninstall the PAS is on premise and install a new PAS on linux on HEC cloud.
Our databse size is 2 TB
we are from government services orgainsation and we cannot afford two downtimes. Will DMO help on this scenario?
Thanks
Balaji
Hi Balaji,
as listed in the respective SAP notes on DMO, this is not supported.
"Using DMO for a migration across data center is not supported (PAS host and DB host in different data center). There are no technical restrictions, but it comes with a high performance and latency impact. You may use it at your own risk. No support is provided in case of performance issues or broken procedures due to network/latency issues. Consider "DMO with System Move" instead."
Regards,
Boris
Hi Boris,
Thanks for the update.
Since system move with DMO is not supported on windows , shall we consider the DMO is not supported for cross datacenter migration on windows at all.
So is this gives a conclusion only classical migration can be done on if we have PAS on windows and if we try to move to cloud?
Thanks
Balaji
Hello Team
We SAP HCM ECC 6.0 (no enhancement packs) would like to upgrade to HR Renewal 2.0 with SAP HANA BD migration. I'm new to this area, appreciate your advice what best way to plan this upgrade.
Looking forward for your advice
Thank you.
Hello Boris,
We SAP ECC 6.0 EHP4 would like to upgrade to SAP HANA 2.0 migration. I’m new to this area, appreciate your advice what best way to plan this upgrade.
Our OS: windows 2008 and DB: SQL Server 2008 and system is Unicode.
Do we need to upgrade our windows and DB first before conversion to SAP HANA. or SAP had any tool that directly migrate to HANA 2.0
Which is best recommended migration option ?
Looking forward for your advice
Thank you.
Hello Aamir,
Then what you did for this scenario? I am also looking for the solution.
Hello,
I need to upgrade a SCM system with livecache to Hana with DMO tool.
Livecache version of system is 7.7 into a windows 2003 server.
I’ve read the central DMO note (2532577 - Database Migration Option (DMO) of SUM 2.0 SP01) but we can’t find any information/restriction about the minimum version of livecache for the migration to a integrated livecache on hana.
Are there any restriction about the livecache version for the migration to hana?
Thanks a lot in advance
Hi Boris,
When we are Migrating from ERP 6.0 EHP4 to EHP8 from oracle to HANA used below configuration
App Server IBM P8 server : 20 CPU'S ( 180 Logical CPUss)
Database Size : 8TB
Network Connection between Source and HANA : 1 GBps.
Question :
when we are setting R3loads in configuration phase , we found the restriction is to max 1000 for R3loads and no further.
I would like to know if there is any work around to increase this value. Appreciate your help in this regard.
Regards,
Shiva
Hi Shiva,
thanks for asking. Let me see if I got the point: you have already used 1000 R3loads, you have checked the performance of your application server, and you think that it is possible to still increase the number of R3loads above 1000? I'd like to know what the migration rate is that you have achieved (especially as we recommend to use a 10 Gbits network card).
Thanks,
Boris
Hi Boris,
Thanks for your reply and apologies for my delayed response.
Yes, we have used 1000 R3loads and would like to use more than that like 1200 as we have sufficient resources available at hardware level.
We used 1 Gbits network. to clarify more below is the landscape
PAS : on HP-UX with oracle DB
AAS : for migration purpouse we installed an AAS on AIX server with 220 Cpu and 200 GB RAM.
HANA : 4 TB Server
These there servers are connected with 1 Gbits network and we achieved 32.54 MB/sec which is from EUMIGRATEDTRUN.log
iPlease let us know if there is possibility to use R3loads more than 1000.
Thanks,
Shiva
not Analysis as initially writtenin the following blogHi Boris,
Our client is running ECC 6.0 EHP 7 – NON Unicode on Oracle 11.2.0.3.0 on HP-UX with 4 TB of data and the hardware is due to be refreshed. The client also wishes to move to Suite on HANA or S/4Hana.
Since a pre-requisite for upgrading to EHP 8 is to have the system already on Unicode so we can’t go to EHP 8 on a single step.
Rather, migration to Unicode and Hana database on new set of servers using DMO option in SUM seems a good option and it will be Suite on HANA with EHP 7 and Hana 2.0. -> Can this be done in a single step?
And are there any other things which I need to keep in mind for this transformation?
Do I need to split the steps for Unicode and HANA migration?
The HP-UX servers are pretty old and almost out of circulation.
Could you please suggest a roadmap that we can follow with this landscape.
Can we perform UC using SUM on the same servers as a single step considering the size and server state and then migrate to HANA in the different phase.
Thanks in advance.
Hi Shitiz,
my view is that these questions are answered in the respective notes for DMO.
And I think the decision about the target level (EHP 7 or EHP 8) should rather be a business decision and not driven by technology considerations.
Anyhow let me see if I can comment on your two questions, assuming you stay on EHP 7 (which means using DMO of SUM 1.0):
1) One step approach: migrate to HANA + convert to Unicode + switch server: possible.
You may use plain DMO, and will have to apply SPs, or use "DMO without System Update" to have the migration only. For the server switch, you combine this with approach "DMO with System Move".
2) Two steps approach: a) convert to Unicode, then b) migrate to HANA: possible as well.
Keep in mind that you will have to check the requirements on OS and db version in the respective notes.
Regards, Boris
Hi Boris,
Thank you for info so far. We are planning to migrate ECC 6.0 EHP 5 running on Windows Server 2008 R2 and SQL Server 2008 TO S/4 Hana 1709. We are planning for DMO with System move for migrating it to S/4 1709. We are planning to move it to Linux based App servers.
As per note 2537452 - Database Migration Option (DMO) of SUM 2.0 SP02, The conversion is supported for Sql 2005 or Sql server 2012 only. For S/4 1709, I guess the SAP_BASIS will be higher than 751
Could you please confirm
Thank you in advance.
Best Regards,
Hi Shitiz,
SAP Note 2537452 states the requirement:
> MS SQL 2012 or higher if target SAP_BASIS is 751 or higher
Sorry to say that it is beyond my knowledge if such in place Windows upgrades are possible.
Best regards, Boris
Hi Borris,
Thank you for pointing this out. We'll be upgrading MS SQL from 2008 to 2012.
As Note 2537452, it states as
Requirements for Source System and Source Database:
Database: All source databases excluding SAP HANA
Operating System (OS) of source primary application server (PAS) host: Any Unix-based or Windows operating system.
So, for Converting ECC 6.0 EHP 5 to S/4 1709, we should be good if we have source server as MS Windows 2008 and we intend to use DMO with System move to have target PAS on Linux?
Could you please confirm if SUM 2.0 SPS 2 supports DMO with System Move from source OS as Windows 2008 and target as S/4 1709.
Best Regards,
Shitiz
Hi Shitiz,
target 1709 requires Windows Server 2012. This requirement is valid also for "DMO with System Move", as the shadow system is being created on the source system.
Best regards, Boris
Hello Boris,
Thank you for your guidance so far, now we have uninstalled the CI from Windows 2008 VM and moved it to a Windows 2012 server based VM, so now our CI is running on Windows 2012 and DB which is MSSQL 2012 is running on Windows 2008.
The Initial check phase of SUM where it verifies the host OS and DB version has passed, but we are not sure if we should continue as our DB MSSQL 2012 is still hosted on Windows 2008.
Can we go ahead with the conversion to 1709 with our MSSQL 2012 still hosted on Windows 2008, or should we do a system copy instead to have the DB server also on Windows 2012?
As per PAM EHP 5 is supported on MSSQL 2012 running on Windows 2008, but we don't want high R3load times or any other OS related issues.
Thank you in advance.
Regards,
Shitiz
Hi Shitiz,
thanks for asking, but unfortunately I do not have more insight than what is stated in the respective notes for SUM, for DMO and for the conversion.
Regards,
Boris
Hi Boris,
great blog! Thanks for sharing so much information.
We have to plan the following migration and it seems quite challenging.
Source-Release: SAP ERP with EHP7 / NW 7.40 non Unicode
Source-OS/DB: Windows2008
Source-DB: Oracle 12
Source-OS/PAS: Windows2008
Target-Release: SAP ERP with EHP8 / NW 7.50
Target-OS/DB: RHEL7
Target-DB: HANA 2.0
Source-OS/PAS: Windows2012
As far as I could understand the documentation and restrictions, there is no "one step" option for this migration, because source system is non Unicode.
So we came up with at least 2 options with DMO (two steps).
Option A.)
A1.) Migration Non Unicode 7.40 --> Unicode 7.40 (new system?)
A2.) DM0 (SUM 2.0) inkl. Upgrade EHP7/7.40 --> EHP8/7.50
Option B.)
B1.) DMO (SUM 1.0) with Uncode, System Move and No System Update) EHP7/7.40/HANA
B2.) Standard Upgrade EHP7/7.40 --> EHP8/7.50
And we will have to install an AAS on Linux in Source and Target to be able to run DMO SUM 1.0 oder SUM 2.0 with System Move Option, correct?
What do you suggest?
Thanks in advance
Best Regards
Kuni
ps. We could really need some kind of "SAP-magic" here 🙂
Hi Kuni,
agree to your two options, I do not see anything that should prevent one of the options.
For the AAS on Linux, maybe we have at least a small portion of magic: with SUM 2.0 SP 02, "DMO with System Move" allows Windows as source PAS operating system. So no AAS on source required - but on target, as Win to Win is not supported.
Best regards, Boris
Hi Boris,
thanks for you answer.
We decided to use option B1.) DMO (SUM 1.0) with Uncode, System Move and No System Update)EHP7/7.40/HANA
1) As DEV and TST have no seperate ASCS instances we would have to split the ASCS beforehand, correct?
2 It is clear we need the AAS instance on linux for the source system, and the PAS on the target system.
Is it possible to
2a.) install just the new PAS on linux for the target system
or
2b.) do we have to install one AAS on linux for the target + PAS on linux for the target system?
2c.) if 2b. As we use VM, could this instance be another AAS instance on the same VM?
3) What are the recommended resource requirements (CPU, RAM, ..) for the AAS?
Thanks in advance
Best Regards
Kuni
Hi Kuni,
sorry, now you lost me. I was under the assumption that the OS of your target PAS shall be Windows.
What I can say is:
- "You may run the SUM on an AAS even if ASCS is not yet running separately." (taken from note 2537452)
- I am not aware of recommendation for the AAS, but as SUM & R3load are running on it, performance is relevant
Best regards,
Boris
Hi Boris,
sorry, my bad.
Of course you are right, PAS on target shall be Windows.
So what would you recommend, installing the AAS for source and AAS for target in one VM, or separated?
Best Regards
Kuni
Hi Kuni,
sorry, I am not a VM expert. What I can say is that performance matters for the migration, so the better your machines perform, the faster the migration (export and import) works.
Not sure if I got your sentence correct: you do not intend to install "AAS for source" and "AAS for target" on the same VM, or do you? Then I wonder why you use "DMO with system move" at all ...
Best regards,
Boris
Hi Boris,
thanks again, also for your patience.
So finally I think we'll have to execute the following steps:
1) Install a new AAS on Linux for DMO
2) On the new AAS we run DMO "Without System Move"
3) After DMO ist finished, we'll have to install a new PAS on Windows for the new HANA DB
4) Delete AAS on Linux
Correct?
Best Regards
Kuni
ps. maybe it would be better to hire a certified consultant and perform an os/db migration 🙂
Hi Kuni,
sorry that I struggle again, now with the sequence: you first use an AAS, and afterwards install a PAS?
Shouldn't it be the other way around: first install the PAS for target SAP HANA database, and then install the AAS? (AAS: additional).
Best regards,
Boris
Hi Boris,
after all, I still struggle again with the basisc question which option to use, "System Move" or not?
Source: PAS on Windows2008
Target: PAS on Windows2012 (new host)
That's why I thought we had to use "System Move Option", because PAS has to be changed.
In any case we need
- a new AAS on linux for the (SUM1.0) connected to the old Oracle DB
- to install a new HANA DB and a new PAS on Windows
So, my question is: "System Move", or not? and depending on the option, what steps do we need?
Thanks for your time and patience.
Best regards
Kuni
Hi Kuni
We have discussed a couple of options and details. Let me see if I got this right:
– your source system is SAP ERP 6.0 EHP 7 on Oracle 12 with Windows 2008
– your target is to run EHP 8 on SAP HANA with Windows 2012
I see two possible ways:
You need two steps:
1 Bring source to HANA on Windows 2012 (on Unicode)
2 SUM update EHP 7 -> EHP 8
For step 1, several options are possible:
1a Use SWPM to new host.
1b Setup a new PAS on Windows 2012 and switch manually to new host, then do “DMO without System Update”.
1c Use “DMO without System update” combined with “DMO with System Move” – but then you have to set up not only PAS on Windows 2012 but also AAS on Linux to allow “DMO with System Move” (win to win not supported) … seems to much work.
Regards,
Boris
Hi Boris,
1a) no OS/DB migration certification ...
1b) this option seems to be less work ... we will consider this approach
1c) our reasons for “DMO with System Move” in the first place:
Many thanks
and
best regards
Kuni
Hello Boris,
Is it possible to migrate SAP CRM data to SAP S/4HANA-Customer Management using DMO ? Here I am talking about :
Thanks in advance.
Regards
Banibrata
Hi Banibrata,
no - guess that Oliver Frick has already provided this answer, in his reply to your request on the blog https://blogs.sap.com/2018/02/28/s4hana-for-customer-management-1.0-introduction-from-technical-point-of-view/
Regards, Boris
Hi Boris,
We are converting our system which is ECC 6.0 EHP 8 on HANA (ASCS running on a different host and SUSE HA ) to S/4 HANA 1809. Need your help on some doubts below, which are not clear with the SUM guide
Regards,
Shitiz
Hello Boris,
we want to to an DMO to target release SAP NW AS for ABAP 7.52. Source DB is DB6, target HANA2.0.
Problem is, that this scenario is not possible, see: 2644872 - Database Migration Option (DMO) of SUM 2.0 SP04
When is this option available? And is there, maybe a way to do this togehter in a test/beta sceanrio?
Regards
Thomas
Hello Thomas,
as far as I know, there are no plans to support this scenario. What kind of system are you considering here?
Regards,
Boris
Hello Boris,
system type is a Fiori Frontend system, stand alone SAP NW AS for ABAP 7.52.
Regards
Thomas
Hi Boris,
I have the same question , I am trying to move (DMO) from Oracle to HANA , NW 7.52 system.
Is this not supported ?
regards
Vagesan
Hello,
We would now like to migrate our DB2 DB to HANA DB.
Our current landscape:
ECC 6.0 Ehp4
DB2 Database
Windows Server
Can I use DMO Tool for the Migration?
Hi Tatjana,
DMO is an option as it works for source system SAP ERP 6.0 with or without EHPs. Please check the requirements in the respective SAP Notes which are listed in the blog above.
Regards, Boris
Hello Boris,
Thanks for reply.
We would like to migrate to SAP HANA 2.0. I am new to this area, and need your advice what is the best way to plan this migration.
Source Database System:
Windows 2003
DB2 9.7 Database
ERP 6 EHP 4
Target Database System
Linux
Hana DB
ERP 6 EHP 8
(Applikation Server should also be migrated to a linux server)
Is that the right migration steps?
I do not understand if I need to install ERP6 EHP8 on the new Linux Server before I migrate the Database form DB2 to HANA or will it be automatically installed/copied during the DMO process? And which is the best migration option for this ? DMO with System move ?
Many Thanks
Hello Tatjana,
yes, your scenario is then "DMO with System Move", and your sequence is correct. After installation of SAP HANA database, you will have to install the AppServer (as described in the DMO guide, section 3.12 System Move: Preparing Target System Landscape). SUM does not install an AppServer, and for this scenario, it is required to setup the new PAS in the target landscape.
Targeting EHP 8 means that you will use SUM 2.0, and you will have to check the requirements in the respective SAP Note for DMO with SUM 2.0 (currently it is Note 2644872 for SUM 2.0 SP 04). You will find that DB2 9.7 is not sufficient (DB2 version 11.1 or higher required if target SAP_BASIS is 750 or higher).
Kind regards, Boris
Hello Boris,
Thanks for the detailed answer. That helped me very well.
I looked at the SAP note 2644872 . There is the following:
I) Target Database SAP HANA
4. DB2 (Db2 for z/OS)
6. DB6 (Db2 for Linux, UNIX, and Windows) -
We have DB2 9.7 on Windows 2003.
Does that mean that we can migrate directly to HANA (without upgrading to DB2 11.1)?
Thank you
Kind regards, Tatjana
Hello Tatjana,
if your source system runs on DB6 (DB2 for LUW) and your target is EHP 8 (which is based on 7.50), it looks as if DB2 9.7 is sufficient if it has at least FP5.
Kind Regards, Boris
Hi Boris,
Thank you very much.
I'll try to ask again.
Do I have to split the enqueue server from the central instance, before migrating to HANA?
Many Thanks
Hi Tatjana,
as far as I can see, you do not have to split the enqueue server before starting DMO with system move - as long as you do not start SUM on an additional application server (as listed in the respective note listed above).
Regards, Boris
Hi Boris,
Thanks for so much information on .
We have to plan the following migration and it seems quite challenging.
Source-Release: NW 7.40 SP08 Unicode
Source-OS/DB: Windows2012 R2
Source-DB: Oracle 11g SP2
Source-OS/PAS: Windows2012 R2
Target-Release: NW 7.40 SP21
Target-OS/DB: SLES12
Target-DB: HANA 2.0
Target-OS/PAS: Windows2016
we came up with at least 2 options with DMO .
Option A.)
A1.) Prepare target Server ( Install one Vanila system) both 1. Database on hana on Linux Host and 1. PAS on Windows.
A2.)Run DMO (SUM 1.0 SP23) System Move option.
So one Shot System Upgrade and Move to terget version. < But as per guide look like there is restiction of PAS move from WIN to WIN.
So is it possible this option ?
Option B.)
B1.) Prepare target Server ( Install one Vanila system) both 1. Database on hana on Linux Host and 1. PAS on Windows. and One AAS on Linux on terget DB Server.
B2.) Run DMO (SUM 1.0 SP23) System Move option. During DMO run on target phase give details of AAS on Linux.
B3.) After DMO complete, Uninstall AAS on Linx and connect back PAS on Windows.
What do you suggest?
Thanks in advance
Best Regards
Kousik Maity
Hi Kousik,
as you have already stated, for targeting a system based on 7.40, you have to use SUM 1.0, and for this SUM version, the scenario "DMO with system move" has restrictions concerning the OS of source and target.
So option A) is not possible, as Win to Win is not supported.
Option B) seems possible. Not sure if I understand your point B3: you may uninstall AAS on linux, but you do not have to connect PAS back.
Best regards,
Boris
Dear Boris,
Thank you for your reply.
----------------
Option B) seems possible. Not sure if I understand your point B3: you may uninstall AAS on linux, but you do not have to connect PAS back.
----------------
Ans: Actually our customer want target PAS on Windows.
Dear Boris,
We want your expert advice 🙂 about below scenario we have decided to execute:
Kindly advice whether below steps visible
Option C:
C1.) Prepare target Server ( Install one Vanila system) both 1. Database on hana on Linux Host and 1. PAS on Windows. < As customer want PAS on Windows)
C2.) Run DMO (SUM 1.0 SP23) without System Move option.
C3.) After DMO complete, Source system PAS will connect to Target Hana Db as it database.
C4.) Now Stop and uninstall Source PAS. Now Run SWPM System Move option in Target Windows Server(PAS) to connect back to Hana DB.
Thanks & Regards,
Kousik Maity
Hello Boris,
DMO for Java Systems >NW740, is this planed in future?
Kind regards
Christoph
Hello Christoph,
no, afaik there is no plan to offer DMO for Java systems. Classical migration can be used.
Kind regards, Boris
Hi, Boris,
we had been already on NW_7.50 SPS09 with EHP8_SPS08, we only made DMO, no upgrade
SUM 2.0 SP07 PL06
It is always extremly annoying, when the SUM (only DMO action, no upgrade) independently enters new system parameters in profile files, especially if these are completely wrong entries, (too small value)
in our SUM-DMO execution (move from DB2-for-i to HDB) SUM added this entry to the PAS instance profile (and only to PAS profile, not to 3 more AAS)
# *** PERMANENT 750 CHANGE ***
ztta/roll_extension = 200000000 <== that is too small !!!
the kernel-default is 10 times higher
ztta/roll_extension = 2000000000
regards
Christoph Ostrop
Hi Christoph,
thank you for the notification. We will check this, make take a while.
Regards,
Boris Rubarth
Follow-up:
with a future SP version of SUM 2.0, this behavior will most probably be adapted, e.g. parameter will only be modified if it was set beforehand and value was < 200000000.
Thanks and regards, Boris
Hi Boris
We are looking at migrating to sybase ASE. We are using SUM tool and was wondering if there is a way for us to skip stats and do it manually if required.
Regards
Krish
Hi Krish,
sorry, stats are required and I can't offer you a way to skip it.
Regards, Boris
Thanks, Boris.
The reason we asked is that Sybase ASE performs stats generation during index creation, so it seemed like we are doing this twice.
Also, the stats when it fails, restarts from beginning , adding to the time load. Our import/export for 4TB source system takes around 40-48 and then if you stats to the top it, it makes for a very long downtime.
Boris, what does this documentation really mean? Are they talking about the 'Update statistics' step in SUM or is this a different step altogether?
https://help.sap.com/viewer/b22e5b44184242ea80802ebb427a7936/03/en-US/2ffc46ac46d440e686c7f2cf8597f95e.html
As of SAP ASE versions 15.7 SP130 and 16.0 SP01, the creation of optimizer statistics during parallel index creation has become stable. At the same time, the use of multiple parallel workers for index creation can be globally limited to a minimum table size. As a result, the best practice is to:
Also there is a note 3166808 - Slow update statistics during SUM DMO migration SYB - SAP ONE Support Launchpad that talks about the workaround to skip stats. we did not have success with this workaround.
Hello Boris
Thanks for the detialed blog.... Can we setup new (System B) on SAP S4 HANA based on ECC Non HANA (System A) Non Uni code system
Which option will be better ,Iam thinking of the first option.
Thanks and Reagrds
Kiran.
Hi Kiran,
thank you for the question.
Unfortunately, I do not know what exactly you are thinking of when stating "create system" in the various steps. Anyway, selected data migration is not possible with DMO.
Thanks and regards,
Boris
Hi Boris,
can I run DMO with System Move Serial mode on the PRD system and migrate this system to Linux/HANA, but then continue to use the old PRD system productively until the project ends?
Many Thanks
Hi Tatjana,
thanks for your question. Let me see if I understand it correct:
In general, "DMO with system move" will include a downtime for the source system (PRD, in your case). So you can't continue to use the old PRD system until project end. Only during this downtime, the application table content is transferred from source to target landscape. Without downtime, table changes by end-user activity in uptime potentially get lost if they happen after the respective table was migrated to the target system. "DMO with system move" does not offer any replication technology that considers such end-user changes. And it can't be combined with downtime-optimized DMO.
This statement is independent on whether you choose serial or parallel mode for "DMO with system move": downtime is required for the source system. It is important to note that the serial mode leads to drastically longer downtime.
Does that hit your question?
Regards, Boris
Hello Boris
Many thanks for your response.
I think I worded my question incorrectly earlier.
Can the source system (.e. PRD) still be used productive after the DMO run ?
Is it possible to continue working on the old system after the migration with DMO System Move ?
Thanks very much
Regards
Hello Tatjana,
OK, got it: different use case.
Technically this is possible:
If you did this on your PRD system: which of those two is now the real PRD? Any changes in the source are missing in the target, as there is no synchronization between the systems.
Leaving aside license aspects.
Regards, Boris
Hi Boris,
We are trying different scenarios for the S4H conversions to see which would give us lowest downtime.
Right now one of our SAP system is running on Windows with Oracle DB. We have created an AAS on Linux in the AWS where SUM will be started with the option "downtime-optimized, uptime migration".
Is there any documentation available for this scenario to prepare the AAS or to run the whole conversion?
Thanks and regards,
Navneet
Dear Boris,
We make SUM runs on sandbox systems.
But export phase lasts around 20 or so hours for 5 TB data on SQL (1.5 TB export size).
I have used 3rd party SQL monitor to get some insights.
It seems that row count phase last around 2.5 hours (check attached images) select count_big (*),
and subtree cost for other selects for splitting seems too high.
Would adding index help?
Does adding index has any drawback other than increasing DB size on SQL server?
After all they will be discarded in HANA DB.
Another thing,
MDOP value is 2 for (98 CPUs), can we try to experiment with 4 or 8.
Regards,
(ROW COUNT phase around 2.5 hours)
General select for exports has high subtree cost as well for all tables.