IC54039: ISSUING TSM SERVER COMMAND CANCEL SESSION ON A RESTORE SESSION MAY CAUSE A PARTIAL SERVER HANG UNDER THE RIGHT.

APAR status

• Closed as program error.

Error description

• During the partial hang, client sessions will stop transferring
  data and no error messages are generated. All sessions will be
  hung and cannot be cancelled.  The partial hang is caused by an
  incorrect locking algorithm used by the CANCEL RESTORE function.
  .
  Versions Effected:  All TSM versions
  .
  Customer/L2 Diagnostics: Capture the following commands just
  before and during the partial hang
  
  1) SHOW TXNT
  2) SHOW LOCKS
  3) SHOW THREADS
  
  If the above commands cannot be issued to the TSM server,
  forcibly end the application.
  
  If the server platform is AIX, issue a kill -11 on the process
  and collect the dump using the AIX utility snapcore.
  
  The condition will be caused by these 2 threads, which are
  waiting on one another.  Please review SHOW THREADS and the TSM
  Server Activity Log for CANCEL RESTORE (or CANCEL SESSION) to
  determine if the condition is being observed.
  
  Thread ###: SmHttpCommandThread
   tid=87987, ktid=166021, ptid=342, det=0, zomb=0, join=0,
  result=0, sess=0
    Acquiring mutex IMV->chainMutex (0x35c52c04) at 1038e48c
  (imCancelRestore_82_66)
    Stack trace:
      0xd004b630 _global_lock_common
      0xd0049eb8 pthread_mutex_lock
      0x1000ded4 pkAcquireMutex
      0x1038e490 imCancelRestore_82_66
      0x10396524 CancelSessionNum
      0x1038bd68 imCancelRestoreCmd
      0x1020bc08 AdmCommandLocal
      0x1020cd74 admCommand
      0x105581a8 HtRunCommands
      0x10558a2c htPostForm
      0x1055599c SmHttpCommandThread
      0x1000ea00 StartThread
      0xd004c7e8 _pthread_body
  
  This thread is waiting to acquire the IMV mutex and holding a
  lock on the RESTORE SESSIONS table.
  
  Thread ###: ImBuildRestoreObjectsThread
   tid=76409, ktid=61639, ptid=300, det=1, zomb=0, join=0,
  result=0, sess=0
    Holding mutex IMV->chainMutex (0x35c52c04), acquired at
  103805dc (SetSourceFlag)
    Awaiting cond waitP->waiting (0x48b09780), using mutex
  TMV->mutex (0x337f2384), in tmLock (0x100425f4)
    Stack trace:
      0xd0064104 _cond_wait_global
      0xd0064cb8 _cond_wait
      0xd0065884 pthread_cond_wait
      0x1000d970 pkWaitCondition
      0x100425f8 tmLock
      0x1037f3f8 LockSession
      0x103805f0 SetSourceFlag
      0x10387434 ImBuildRestoreObjectsThread
      0x1000ea00 StartThread
      0xd004c7e8 _pthread_body
  
  This thread is waiting for the lock on the RESTORE SESSIONS
  table and holding the IMV mutex.
  
  Initial Impact:  Medium
  
  Additional Keywords:  hang hung frozen freeze freezing nqr
  

Local fix

• 1)  Issue the SHOW COMMANDS above or kill the server process to
      generate a core dump for support analysis
  2)  Avoid issuing CANCEL SESSION or CANCEL RESTORE
  3)  Add TESTFLAG DISABLENQR to dsm.opt for all clients running
      restore
  

Problem summary

• ****************************************************************
  * USERS AFFECTED: All TSM Server users.                        *
  ****************************************************************
  * PROBLEM DESCRIPTION: See Error Description.                  *
  ****************************************************************
  * RECOMMENDATION: Apply fixing level when available. This      *
  *                 problem is currently projected to be fixed   *
  *                 in level 5.4.3. Note that this is subject    *
  *                 to change at the discretion of IBM.          *
  ****************************************************************
  *
  

Problem conclusion

• The problem has been fixed.
  

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:

  PK55476 PK66605

Fix information

• Fixed component name

  TSM SERVER

• Fixed component ID

  5698ISMSV

Applicable component levels

• R53A PSY

     UP

• R53H PSY

     UP

• R53L PSY

     UP

• R53S PSY

     UP

• R53W PSY

     UP

• R53Z PSY

     UP

• R54A PSY

     UP

• R54H PSY

     UP

• R54L PSY

     UP

• R54S PSY

     UP

• R54W PSY

     UP

• R54Z PSY

     UP