IZ54313: MULTI-THREADED MQ CLIENT APPLICATION MAY FAIL WITH SIGABRT IF CONNECTION TO QMGR IS INTERRUPTED AND CONNECT ATTEMPTED AGAIN

WebSphere MQ V6.0 Fix Pack 6.0.2.8
WebSphere MQ V6.0 for iSeries Fix Pack 6.0.2.8

APAR status

• Closed as program error.

Error description

• Consider a multi-threaded WebSphere MQ client application that
  has non-shared connection handles open with a queue manager.
  Suppose the application is designed to reconnect to the queue
  manager when the connection is broken. In such a scenario, the
  thread attempting to reconnect, following a broken connection,
  terminates with SIGABRT signal.
  
  The core file thus generated, shows the below stack trace:
  #0  0xffffe410 in __kernel_vsyscall ()
  #1  0xb7c2e7d0 in raise () from /lib/libc.so.6
  #2  0xb7c2fea3 in abort () from /lib/libc.so.6
  #3  0xb7c63f8b in __libc_message () from /lib/libc.so.6
  #4  0xb7c69911 in malloc_printerr () from /lib/libc.so.6
  #5  0xb7c6ab70 in _int_free () from /lib/libc.so.6
  #6  0xb7c6af84 in free () from /lib/libc.so.6
  #7  0xb7bc9c3c in FreeConn () from /opt/mqm/lib/libmqiz_r.so
  #8  0xb7bc9ce1 in ClientThreadDestructor ()
  from /opt/mqm/lib/libmqiz_r.so
  #9  0xb7dc5d0e in xppRunDestructors ()
  from /opt/mqm/lib/libmqmcs_r.so
  #10 0xb7dc3419 in destroy_thread ()
  from /opt/mqm/lib/libmqmcs_r.so
  

Local fix

Problem summary

• ****************************************************************
  USERS AFFECTED:
  Multi-threaded WebSphere MQ v6.x client applications having non-
  shared connections (MQCNO_HANDLE_SHARE_NONE) open with the
  queue manager are affected. WebSphere MQ v7.x clients are not
  affected.
  
  Platforms affected:
  All Distributed (iSeries, all Unix and Windows)
  ****************************************************************
  PROBLEM SUMMARY:
  While connecting to the queue manager, a thread-cleanup
  function is registered. This is unregistered when the queue
  manager disconnects. Registration and de-registration are
  controlled by a flag.
  
  However, while reconnecting to the queue manager following a
  broken connection, there was a code defect that incorrectly
  updated the flag, causing the thread-cleanup function to be
  invoked during thread termination, though the cleanup had
  already taken place once. This double cleanup operation caused
  SIGABRT.
  

Problem conclusion

• The code that was not setting the flag that controlled
  registration and de-registration of the thread-cleanup
  function has been corrected.
  
  ---------------------------------------------------------------
  The fix is targeted for delivery in the following PTFs:
  
                     v6.0
  Platform           Fix Pack 6.0.2.8
  --------           --------------------
  Windows            U200309
  AIX                U825517
  HP-UX (PA-RISC)    U824678
  HP-UX (Itanium)    U825875
  Solaris (SPARC)    U825511
  Solaris (x86-64)   U825872
  iSeries            SI36108
  Linux (x86)        U825181
  Linux (x86-64)     U825874
  Linux (zSeries)    U825516
  Linux (Power)      U825182
  Linux (s390x)      U825873
  
  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

  WMQ LIN X86 V6

• Fixed component ID

  5724H7204

Applicable component levels

• R602 PSY

     UP