IC91711: SERVER MIGHT HALT OR CRASH WHEN PERFORMING DEDUPLICATION OF VERY LARGE FILES DUE TO ACTIVE LOG FILLING

APAR status

• Closed as program error.

Error description

• Deduplication (either client-side or server-side) of very large
  files requires  a lot of active log space. The interaction with
  other server processes which consume log space as well
  might cause the server to appear to hang, to halt unexpectedly
  or even crash when the active log space is totally consumed.
  The size of files that can be processed without experiencing
  this APAR varies depending on the amount of work being
  performed simultaneously on the Tivoli Storage Management
  server, but the failure has been witnessed with files of size
  400G or greater. This failure is directly correlated to the
  size of the file being processed, so customers with larger
  files will tend to experience the APAR more often.
  This failure can be be diagnosed by examining SHOW THREADS and
  SHOW TXNT output for the following entries:
  SHOW THREADS will display a call stack similar to the following:
  Thread 38, Parent 1: BfCleanupPendingInsertionsThrea, Storage
  25788, AllocCnt 46 HighWaterAmt 27196
   tid=1126, ptid=1, det=1, zomb=0, join=0, result=0, sess=0
    Stack trace:
      0x0900000000255c60 semop
      0x090000000449358c sqloSSemP
      0x0900000004492f64 .sqlccrecv.fdpr.clone.739
      0xffffffff89000017 *UNKNOWN*
      0x090000000449282c sqljcReceive__FP10sqljCmnMgr
      0x09000000044a0440
  sqljrDrdaArExecute__FP14db2UCinterfaceP9UCstpInfo
      0x0900000004737520
  CLI_sqlExecute__FP17CLI_STATEMENTINFOP19CLI_ERRORHEADERINFO
      0x09000000047a671c
  SQLExecute2__FP17CLI_STATEMENTINFOP19CLI_ERRORHEADERINFO
      0x09000000047b8c64 SQLExecute
      0x0000000100157e14 tbRegExecEx
      0x00000001007a6618 BfCleanupPendingInsertionsThread
      0x000000010000c0e0 StartThread
  Searching the SHOW TXNT output to find the associated
  transaction for that thread(in this
  example the ThreadId is 38) will show information similar to
  the following:
  Tsn=0:15673792, Resurrected=False, InFlight=True,
  Distributed=False, Persistent=True, Addr 1111d9c58
    Start ThreadId=38, Timestamp=04/05/13 20:14:23,
  Creator=bfddedup.c(4096)
    Last known in use by ThreadId=38
    Participants=1, summaryVote=ReadOnly
    EndInFlight False, endThreadId 0, tmidx 0 0,
  processBatchCount 0, mustAbort False.
      Participant DB: voteReceived=False, ackReceived=False
        DB: Txn 112517218, ReadOnly(NO), connP=1123cd398,
  applHandle=338, openTbls=1:
        DB: --> OpenP=11b2a66f8 for table=BF.Bitfile.Extents.
        DB: --> RegSqlId=0x01000034 DELETE for
  table=BF.Bitfile.Extents, executed(No).
  If the timestamp in the SHOW TXNT output indicates that the
  transaction has been running for many hours, and QUERY LOG
  commands being issued over time indicates that more and more
  active log space is being consumed, this APAR does potentially
  apply.
  Versions Affected:
  All supported Tivoli Storage Manager server releases on all
  supported platforms.
  Initial Impact:
  High
  Additional Keywords:
  dedup tsm hang activelog crash zz61 zz62 zz63
  

Local fix

• Adjust Tivoli Storage Management policies such that very large
  files are not stored in deduplicated storage pools.
  

Problem summary

• ****************************************************************
  * USERS AFFECTED: All Tivoli Storage Manager server users.     *
  ****************************************************************
  * PROBLEM DESCRIPTION: See error description.                  *
  ****************************************************************
  * RECOMMENDATION:                                              *
  ****************************************************************
  This problem is projected to be fixed in a future version of
  the Tivoli Storage Manager server.  Note that this is subject
  to change at the discretion of IBM.
  Affected platforms:  AIX, HP-UX, Solaris, Linux, and Windows.
  

Problem conclusion

• *
  

Temporary fix

Comments

APAR Information

• APAR is sysrouted FROM one or more of the following:

• APAR is sysrouted TO one or more of the following:

Fix information

• Fixed component name

  TSM SERVER

• Fixed component ID

  5698ISMSV

Applicable component levels

• R63A PSN

     UP

• R63H PSN

     UP

• R63L PSN

     UP

• R63S PSN

     UP

• R63W PSN

     UP