Skip to Content
avatar image
Former Member

no delta merge for long time

Hello,

yesterday our HANA strucks because of low memory. When I was analyzing this I found a big database but no really big tables, until I recognized large delta areas.

The original DB was 120 GB but now it was 370 GB. I have restartet the DB today in the morning and after 20 Minutes it shrinks to 120 GB again and all Delta Areas were emty.

Can you tell me a reason, why the merge might not have worked? It seems not working since weeks.

Kind Regards

Andreas

Add comment
10|10000 characters needed characters exceeded

  • Get RSS Feed

5 Answers

  • Best Answer
    avatar image
    Former Member
    Jul 17, 2017 at 01:24 PM

    Hello,

    the problem was known and solved by a hana patch as mentioned in note

    https://launchpad.support.sap.com/#/notes/0002301382

    Best Regards

    Andreas

    Add comment
    10|10000 characters needed characters exceeded

  • Apr 04, 2017 at 07:30 AM

    Hi,

    Did you get an OOM trace file? If so can you attach it?

    I would recommend executing the mini check script from 1969700.

    Add comment
    10|10000 characters needed characters exceeded

    • Former Member

      Hello Michael,

      thanks.

      I have got rtedump.oom.trc of indexserver, divserver, webdispatcher, preprocessor, nameserver and xsengine. The only hint I can find are inactive threats, for example:

      5769[thr=25950]: JobexWatchdog is inactive

      5771[thr=25951]: JobexMainDispatcher is inactive

      5773[thr=25952]: JobWrkDiag is inactive

      5775[thr=25953]: JobWrk0002 is inactive

      5777[thr=25954]: JobWrk0003 is inactive

      10951349[thr=47683]: Pool is inactive

      10754699[thr=19253]: Pool is inactive

      5[thr=0]: <SYSTEM> is inactive

      3639[thr=25920]: TimerThread is inactive

      25400[thr=25874]: LocalPostCommitHandler is inactive

      25669[thr=25889]: LoadDataWriter is inactive

      25664[thr=25888]: MinuteCron is inactive

      25660[thr=25887]: MasterTokenLockWriter is inactive

      5642[thr=27508]: WakeupThread is inactive

      But I do not know if they should be active.

      I am still analysing the script and will execute it soon.

      Best Regards

      Andreas

  • avatar image
    Former Member
    Apr 04, 2017 at 08:11 AM

    Hi,

    https://www.sap.com/documents/2012/05/9ec46a16-547c-0010-82c7-eda71af511fa.html

    you can go through the attached guide.

    Check your error code:

    Error 2482: The delta storage is empty or the evaluation of the smart merge cost function indicated that a merge is not necessary.

    No further action required if this occurs occasionally.

    If it happens frequently:

    Check M_DELTA_MERGE_STATISTICS and review smart merge cost function with SAP experts.

    (parameter smart_merge_decision_func)

    Only committed data is moved from delta storage to main storage. Uncommitted records remain in delta storage. It is a good practice (also considered by the auto merge decision function) to avoid merges if a high fraction of records in delta storage is not committed, yet.
    Changes performed during the delta merge will also appear in the new delta storage

    Add comment
    10|10000 characters needed characters exceeded

    • Former Member

      Hello,

      thanks for your reply. I have read the guide, it was very interesting.

      I have checked M_DELTA_MERGE_STATISTICS.

      The latest entry is: hges2349;30.003;MERGE;_SYS_STATISTICS;STATISTICS_ALERTS_BASE;0;FALSE;FALSE;;04.04.2017 07:55:01.928;1.232;AUTO;TRUE;11;0;

      There are a lot of These entries. I think the merge works. So the problem has to be somewhere else.

      Best Regards

      Andreas

  • Apr 04, 2017 at 11:22 PM

    M_DELTA_MERGE_STATISTICS only shows the delta merge information since the last restart.

    To figure out what was going on before the instance got restarted you may want to check the historic records with a SQL like this:

    select top 100 * 
    from
    "_SYS_STATISTICS"."HOST_DELTA_MERGE_STATISTICS_BASE"
    where last_error <> 0
    order by snapshot_id desc;
    Add comment
    10|10000 characters needed characters exceeded

    • Former Member

      Hello Lars,

      thank you for the Statement. I have executed it and found the last unsuccessfull entry at 23.03.2017. So there seems to be no error in merging.

      Best regards

      Andreas

  • avatar image
    Former Member
    Apr 07, 2017 at 06:23 AM

    Hello,

    thanks for all your replies. I got a lot of Information and I think merging works.

    But the Problem occurs again. Since monday the indexserver has grown from 130 to 230 GB. We are not working at the System now, so there are no data that could cause this.

    Any idea where I can find Information what is going on?

    If not I would open a customer call at SAP.

    Best Regards

    Andreas

    Add comment
    10|10000 characters needed characters exceeded