cancel
Showing results for 
Search instead for 
Did you mean: 

Delta Merge failing during data conversion.

bryan_bingay
Participant
0 Kudos

We are doing a mock data conversion test of our legacy data, to our SAP PRE-PROD ECC system. I'm noticing a lot of HANA alert 88. Delta merge is failing while the data load is happening. I have a couple of questions.

1. Is it possible to turn off the delta merge system wide during the load? In our mock test it is taking 4 days to complete. The system will be up and available to the business during the load. I have concerns about turning it off for that long.

2. What are the best practices around delta merge during a large data load?

Accepted Solutions (1)

Accepted Solutions (1)

KMS71
Participant
0 Kudos

Hi Bryan,

Not sure if there is a best practice around this, but I understand (not done this myself) that you can disable the Automerge per loaded table while you are loading and the enable the Automerge again once the load is done. You may have to break your load into chunks and then trigger a Memory Merge after each chung, then a 'normal' MERGE before enabling Automerge again (credit to Lars and Richard and their book for this!).

Would also have a look at view M_DELTA_MERGE_STATISTICS for any additional info as to whey the merge is failing...hitting Out of memory maybe?

Hope this helps.

Kris

bryan_bingay
Participant
0 Kudos

Hi Kris. Thanks for the reply. M_DELTA_MERGE_STATISTICS shows this error. I will keep searching on SAP support for more information on this error.

columnstore error: [2486] table optimization did not have an effect because all rows in deltawere not optimizable

Former Member
0 Kudos

See SAP Note 2057046 for a lot of information related to merges, including info message 2486.

bryan_bingay
Participant
0 Kudos

The error messages in the view M_DELTA_MERGE_STATISTICS does not have a good enough error to troubleshoot this. I have engaged SAP support. I increased the trace level of the  mergedog and merge monitor to get get more detailed error messages. I then did a full system dump and uploaded it to SAP.

Thank you Martin and Kris taking the time to help.

Bryan

http://help.sap.com/saphelp_hanaplatform/helpdata/en/ca/8d766d538946f0b0ca9d1fd3901609/content.htm?f...

KMS71
Participant
0 Kudos

Hi Bryan,

Let us know what the solution was once the wizards reply 🙂

Cheers

Kris

bryan_bingay
Participant
0 Kudos

SAP support confirmed this is a bug in 112.02. It is fixed in 112.03. A workaround is to unregisterthe secondary and then register it with the "--force_full_replica" option. I tried the workaround in pre-prod, but it did not work. I tried the same workaround in prod and it worked.

The problem is, the delta merge is trying on the secondary and it should only be doing this onthe primary. When I query the view M_DELTA_MERGE_STATISTICS the host it fails on is the secondary, the primary succeeds.

I also had an index server crash in prod Saturday. There is a memory leak in the merge process. None of these issues are in the release notes for 112.03. I would avoid 112.02 and move to 112.05 instead.

Bryan

bryan_bingay
Participant
0 Kudos

After more back and forth with SAP support, they identified two separate problems, both related to system replication operationMode, logreplay. One with delta merge and the other with system replication. Two different development teams had to get involved.SAP has created two KB notes, 2348480 and 2351926. One is fixed with 112.03 and the other is fixed with 112.05.

logreplay is a new mode. Delta data shipping is still supported but the takeover times are supposed to be longer than logreplay. I asked SAP support if these problems can be avoided with delta data shipping if patching is not possible. It's pretty quick to change the replication mode without impacting the systems. I'll provide an update once I hear back from support.

Bryan

bryan_bingay
Participant
0 Kudos

SAP support confirmed that this only happens in operation mode logreplay. It does not happen in delta_datashipping.

Answers (2)

Answers (2)

bryan_bingay
Participant
0 Kudos

SAP has created note 2348480 for this problem.

Former Member
0 Kudos

Hi Bryan,


It is the only store that divides the Colum tables into a "Main" and a Delta" part.

Therefore Column Store divides each columns into two parts:


Read Main optimized


DeltavWrite optimized


  • Delta: Less Compressed, Write Optimized, Not Sorted Dict.
  • Main: Very Compressed, Read Optimized, Sorted Dictionary

Types of Merge:

Auto merge: A background thread, called MergedogMonitor, checks every 60 seconds all auto merge enabled columns

  Smart merge

If an application powered by SAP HANA requires more direct control over the merge process, SAP HANA supports a function that enables the application to request the system to check whether or not a delta merge makes sense now. This function is called smart merge. For example, if an application starts loading relatively large data volumes, a delta merge during the load may have a negative impact both on the load performance and on other system users. Therefore, the application can disable the auto merge for those tables being loaded and send a “hint” to the database to do a merge once the load has completed. When the application issues a smart merge hint to the database to trigger a merge, the database evaluates the criteria that determine whether or not a merge is necessary. If the criteria are met, the merge is executed.


Hard merge : triggers a Delta Merge independently of Delta size

Forced merge: is a "do-it-now" option that starts a Delta Merge independently of Delta size or system recourses.


Critical merge :is there to prevent disastrous situations that could occur if all other merge triggers not manage to start a merge


Hope you  if you use Smart Merge will fix the issue....


Please let me know  if im wrong.......


Thanks in  Advance


Regards

Prasanth

bryan_bingay
Participant
0 Kudos

Smart Merge seems like it is for a custom app. This is a suite on HANA install, ERP, CRM and BW. SAP confirmed it is a bug in the version I have installed.

Bryan