IT37359: IBM MQ channel process amqrmppa hangs after memory exception

APAR status

• Closed as program error.

Error description

• The IBM MQ channel process amqrmppa was not responding after
  reporting memory exception (e.g. SIGSEGV)". The problem could
  also occur on other IBM MQ processes (e.g. runmqchl, amqcrsta)
  and MQ client applications. The stack trace captured from the
  affected process is likely to show functions ctime and
  cccFormatStatus.
  
  Thread stack:
      #0  0x00007f9c263456ec in __lll_lock_wait_private () from
  /usr/lib64/libc.so.6
      #1  0x00007f9c262c1ba2 in _L_lock_16654 () from
  /usr/lib64/libc.so.6
      #2  0x00007f9c262be7e3 in malloc () from
  /usr/lib64/libc.so.6
      #3  0x00007f9c262c5b8a in strdup () from
  /usr/lib64/libc.so.6
      #4  0x00007f9c262ef1d1 in tzset_internal () from
  /usr/lib64/libc.so.6
      #5  0x00007f9c262efb93 in __tz_convert () from
  /usr/lib64/libc.so.6
      #6  0x00007f9c262edae9 in ctime () from /usr/lib64/libc.so.6
      #7  0x00007f9c27f3feff in cccFormatStatus () from
  /opt/mqm/lib64/libmqmr_r.so
      #8  0x00007f9c27f429a3 in cccFFSTWorkDump () from
  /opt/mqm/lib64/libmqmr_r.so
      #9  0x00007f9c271050f8 in ?? () from
  /opt/mqm/lib64/libmqmcs_r.so
      #10 0x00007f9c2710105e in xcsFFSTFn () from
  /opt/mqm/lib64/libmqmcs_r.so
      #11 0x00007f9c270f2db9 in xehExceptionHandler ()from
  /opt/mqm/lib64/libmqmcs_r.so
      #12 <signal handler called>
      #13 0x00007f9c262ba584 in _int_free () from
  /usr/lib64/libc.so.6
      #14 0x00007f9c27181015 in xcsFreeMemFn () from
  /opt/mqm/lib64/libmqmcs_r.so
  

Local fix

Problem summary

• ****************************************************************
  USERS AFFECTED:
  Users of IBM MQ code (running in client or server installations)
  in the unlikely event that a memory exception is generated
  within heap management routines (a possible cause of this is
  corruption of the heap memory).
  
  
  Platforms affected:
  MultiPlatform
  
  ****************************************************************
  PROBLEM DESCRIPTION:
  To format a date the IBM MQ signal handler called C library
  function ctime marked as non-async-signal-safe. This caused a
  hang in malloc when acquiring a lock used
  in memory management as the lock had already been acquired by
  the same thread when freeing memory but then received a memory
  exception.
  

Problem conclusion

• The IBM MQ code has been modified to use the thread-safe ctime_r
  instead of ctime to resolve the problem.
  
  ---------------------------------------------------------------
  The fix is targeted for delivery in the following PTFs:
  
  Version    Maintenance Level
  v9.1 LTS   9.1.0.9
  v9.2 LTS   9.2.0.3
  v9.x CD    9.2.4
  
  The latest available maintenance can be obtained from
  'WebSphere MQ Recommended Fixes'
  http://www-1.ibm.com/support/docview.wss?rs=171&uid=swg27006037
  
  If the maintenance level is not yet available information on
  its planned availability can be found in 'WebSphere MQ
  Planned Maintenance Release Dates'
  http://www-1.ibm.com/support/docview.wss?rs=171&uid=swg27006309
  ---------------------------------------------------------------
  

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

  MQ BASE V9.2

• Fixed component ID

  5724H7281

Applicable component levels

• R920 PSY

     UP