IT25090: A BACKUP OR ARCHIVE OPERATION MAY HANG ON LINUX PLATFORM

IBM Spectrum Protectâ„¢ Client 8.1.6 Downloads and READMEs

APAR status

• Closed as program error.

Error description

• An IBM Spectrum Protect client backup or archive operation may
  hang when running on a Linux platform.
  No error is reported in the client's error log file.
  
  IBM Spectrum Protect Versions Affected: 7.1.x or 8.1.x on Linux
   platform
  
  Customer/L2 Diagnostics:
  
  - Check with the command below if there is a parent + child
  client process created for the hanging dsmc command:
      "ps -ef |grep dsmc"
  
  UID   PID  PPID  C STIME TTY  TIME CMD
  
  root 27375 16547 0 15:05 ?    00:00:02 dsmc i /fsxx/path/
    -sub=yes
  root 27390 27375 0 15:05 ?    00:00:00 dsmc i /fsxx/path/
    -sub=yes
  
  - Note the PID for the parent and child process ( here 27375 /
    27390 ) and collect output for command  "gstack <PID>" for
    both.  The parent will show a thread stack similar to the
    one below:
  
  Thread 4 (Thread 0x7fc573f09700 (LWP 27381)):
  #0  0x00007fc57f8ee189 in waitpid () from /lib64/libpthread.so.0
  #1  0x0000000000c85639 in GpfsClusterInfo::processPopenCall(char
  const*, std::string&) ()
  #2  0x0000000000c88a4e in GpfsClusterInfo::requestLocalNodeId()
  ()
  #3  0x0000000000c8907e in GpfsClusterInfo::getInstance() ()
  #4  0x00000000004c5404 in HsmFsEntry::HsmFsEntry() ()
  #5  0x0000000000582215 in doBackMigr(txnProducerObject_t*,
  backupSpec*, int, char*) ()
  #6  0x0000000000550f08 in baIncremental(backupPrivObject_t*,
  backupSpec*) ()
  #7  0x0000000000558787 in baProcessRequest(backupPrivObject_t*,
  backupSpec*) ()
  #8  0x000000000055a1d7 in baRequestHandler(backupPrivObject_t*,
  backupSpec*) ()
  #9  0x00000000005a1453 in DccTxnProducer::HandleQueue() ()
  #10 0x00000000005a1b32 in DccTxnProducer::Run(void*) ()
  #11 0x00000000005a1bf0 in DccTxnProducer::DoThread(void*) ()
  #12 0x00000000005a1c17 in callDoPThread2(void*) ()
  #13 0x00000000006dcda7 in startThread2(Threadstart*) ()
  #14 0x00000000006dcddc in startThread(Threadstart*) ()
  #15 0x00007fc57f8e6e25 in start_thread () from
  /lib64/libpthread.so.0
  #16 0x00007fc57e5cb34d in clone () from /lib64/libc.so.6
  
     The child process will show that it is waiting on a "Mutex":
  
  #0  0x00007f79ddb2d945 in pthread_cond_wait@@GLIBC_2.3.2 () from
  /lib64/libpthread.so.0
  #1  0x00000000006eaf59 in psWaitCondition(pthread_cond_t*,
  pthread_mutex_t*) ()
  #2  0x00000000006db4ae in pkWaitCb(conditionBundle*) ()
  #3  0x000000000070c07b in fifoObject::fifoQgetNextWaitNoTS(int)
  ()
  #4  0x000000000070c409 in fifoObject::fifoQgetNextWait(int) ()
  #5  0x00000000005a126c in DccTxnProducer::HandleQueue() ()
  #6  0x00000000005a1b32 in DccTxnProducer::Run(void*) ()
  #7  0x00000000005a1bf0 in DccTxnProducer::DoThread(void*) ()
  #8  0x00000000005a1c17 in callDoPThread2(void*) ()
  #9  0x00000000006dcda7 in startThread2(Threadstart*) ()
  #10 0x00000000006dcddc in startThread(Threadstart*) ()
  #11 0x00007f79ddb29e25 in start_thread () from
  /lib64/libpthread.so.0
  #12 0x00007f79dc80e34d in clone () from /lib64/libc.so.6
  
  
  Initial Impact: Medium
  
  Additional Keywords:   TSM hung hang freeze stuck TS000848315
  

Local fix

• - None
  

Problem summary

• ****************************************************************
  * USERS AFFECTED:                                              *
  * IBM Spectrum Protect Backup-Archive Client versions 7.1 and  *
  * 8.1 running on all Linux platforms                           *
  ****************************************************************
  * PROBLEM DESCRIPTION:                                         *
  * see ERROR DESCRIPTION                                        *
  ****************************************************************
  * RECOMMENDATION:                                              *
  * Apply fixing level wen available. This problem is projected  *
  * to be fixed in level 8.1.6. Note that this is subject to     *
  * change at the discretion of IBM.                             *
  ****************************************************************
  *
  

Problem conclusion

• The code has been changed and IBM Spectrum Protect for
  Backup-Archive Client will not hang.
  

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:

Modules/Macros

• dsmc
  

Fix information

• Fixed component name

  TSM CLIENT

• Fixed component ID

  5698ISMCL

Applicable component levels

• R81L PSY

     UP