IC70497: SERVER CAN HANG DURING BACKUP DB PROCESS

APAR status

• Closed as unreproducible.

Error description

• A Tivoli Storage Manager server can  hang during backup db
  process.
  
  
  Customer/L2 Diagnostics:
  
  Force the server to core/dump.
  From the forced tsm server core/dump file, a thread has the
  following callstack:
  
  net_stubs.recv(??, ??, ??, ??) at 0x900000000103788
  .tcprecv(SQLCC_COMHANDLE_T*,int,char*,int,unsigned
  short,unsigned
  short,SQLCC_TCPCONNHANDLE_T*,SQLCC_COND_T*,unsigned
  int,int*)(0x0, 0x5100000051, 0x200000400162110, 0x100000000,
  0x200000400162110, 0x172727079, 0x36, 0x73716c6a72727079) at
  0x900000002f97a74
  cliutl.sqlcctcprecv(SQLCC_COMHANDLE_T*,SQLCC_COND_T*)(??, ??) at
  0x900000002f995a8
  sqleu_common.sqlccrecv.fdpr.clone.21.chunk.6(0x11cea3aa0, 0x0,
  0xffff0000ffff, 0x111649090, 0x0, 0x9000000000000090,
  0x1000000000001, 0x1000000000001) at 0x90000000355c998
  sqleu_common.sqljcReceive(sqljCmnMgr*)(??) at 0x900000002dc509c
  sqleu_common.sqljrDrdaArDisconnect(db2UCinterface*)(??) at
  0x900000002e11a58
  sqltraceflags.sqleUCdisconnect(??) at 0x900000002e11f90
  sqltraceflags.sqleUCappConnectReset(??, ??, ??, ??) at
  0x900000002e124e4
  sqltraceflags.CLI_sqlDisconnect(CLI_CONNECTINFO*,sqlca*,CLI_ERRO
  RHEADERINFO*)(??, ??, ??) at 0x900000002e1af7c
  sqltraceflags.SQLDisconnect2(CLI_CONNECTINFO*)(??) at
  0x900000002e1b31c
  sqltraceflags.SQLDisconnect.fdpr.chunk.9(0x2800000028) at
  0x900000002e1b598
  dbiconn.RdbCloseConnection(??, ??, ??) at 0x1001ae358
  dbiconn.DbiReleaseConnection(??, ??, ??) at 0x1001ae958
  bftxn.FreeTxnDesc(??, ??, ??) at 0x1001b7e0c
  dbitxn.dbiEndTxn(??, ??, ??) at 0x1001b7a84
  tmutil.tmDiscardTxn(??, ??) at 0x1000c1560
  pkthread.FinishThread(??, ??) at 0x10000b760
  pkthread.pkExitThread(??) at 0x10000b414
  pvr.AgentThread(??) at 0x100262e84
  pkthread.StartThread(??) at 0x100009d64
  
  The key routine is tmDiscardTxn, where we acquire a global
  mutex. This thread is hanging in db2 calls.
  So, it blocks any threads which attempt to acquire the same
  global mutex.
  For example, routine tmBeginNamed attempt to acquire this global
  mutex, so, any thread calls tmBeginNamed will be blocked.
  Many threads will be found which has tmBeginNamed in the
  callstack.
  The following thead is one of them.
  
  pth_spinlock._global_lock_common(??, ??, ??) at
  0x9000000006587b4
  pth_mutex._mutex_lock(??, ??, ??) at 0x900000000665638
  pkmon.pkAcquireMutexTracked(??, ??, ??) at 0x100007444
  tmtxn.tmBeginNamed(??, ??, ??, ??) at 0x1000bd40c
  bfdedup.bfDerefQueueThread(??) at 0x1005d61d0
  pkthread.StartThread(??) at 0x100009d64
  
  
  
  Initial Impact:
  Medium
  .
  Tivoli Storage Manager Versions Affected:
  Tivoli Storage Manager Server versions 6.1.0, 6.1.1, 6.1.2,
  6.1.3 and 6.2.0 on all platforms.
  .
  Additional Keywords:
  TSM zz61 zz62 62379 halt hang database backup db
  

Local fix

• Stop the server process/service and restart it.
  

Problem summary

• ****************************************************************
  * USERS AFFECTED: All Tivoli Storage Manager server users of   *
  *                 the BACKUP DB command.                       *
  ****************************************************************
  * PROBLEM DESCRIPTION: See error description.                  *
  ****************************************************************
  * RECOMMENDATION:                                              *
  ****************************************************************
  This problem was fixed at levels 6.1.4.0 and 6.2.1.0.
  Affected platforms:  AIX, HP-UX, Sun 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

Applicable component levels