IC66603: TOPIC LISTENERS RECEIVING SAME MESSAGE TWICE AFTER QUEUE MANAGER FAILOVER WHEN USING MQ V7 IN JBOSS

WebSphere MQ V7.0.1 for i5/OS Fix Pack 7.0.1.3
WebSphere MQ V7.0 Fix Pack 7.0.1.3

APAR status

• Closed as program error.

Error description

• Two instances of the same message are seen by an application
  running in JBoss when using the WebSphere MQ v7 Resource Adapter
  (RA).  The duplicate messages are seen shortly after the queue
  manager fails over.
  
  The issue is seen when using two applications running in
  separate instances of an application server.  One sends a
  message which the other picks up using a topic listener.
  Shortly after the queue manager in use has failed over to
  another queue manager running on a different server the
  receiving application receives a duplicate message.  That is, it
  sees the same message twice.
  
  After a restart of the partition with the sending application,
  no further duplicate messages are seen.
  

Local fix

Problem summary

• ****************************************************************
  USERS AFFECTED:
  This issue affect users of the WebSphere MQ v7 Resource Adapter
  connecting to a queue manager who restart their queue manager
  twice as part of a failover, or otherwise break a connection as
  it is being rebuilt.
  
  Platforms affected:
  All Distributed (iSeries, all Unix and Windows) +Java
  ****************************************************************
  PROBLEM SUMMARY:
  When a connection from the resource adapter to the queue manager
  was broken, the resource adapter created a reconnection helper
  thread to rebuild this connection and its associated consumer.
  
  If this new connection was broken after the connection was
  created but before the consumer was created, a new reconnection
  helper thread was created to reconnect what the resource adapter
  thought was a new broken connection.
  
  However, as the original reconnection helper thread treated this
  break in the connection as a failure to rebuild the connection,
  it would continue in its attempt to reconnect.
  
  This meant we could end up with two reconnection helper threads
  active for a connection, and so when they were both eventually
  successful in reconnecting, we would have two instances of a
  consumer, where before the break in the connection there was
  only one.
  

Problem conclusion

• A broken connection when recreating a consumer now does not
  create a new reconnection helper thread, instead the original
  thread handles the reconnection. Only one reconnection helper
  thread can now be active for a consumer at a time.
  
  ---------------------------------------------------------------
  The fix is targeted for delivery in the following PTFs:
  
                     v7.0
  Platform           Fix Pack 7.0.1.3
  --------           --------------------
  Windows            U200320
  AIX                U834987
  HP-UX (PA-RISC)    U834414
  HP-UX (Itanium)    U834413
  Solaris (SPARC)    U834986
  Solaris (x86-64)   U834210
  iSeries            tbc_p700_0_1_3
  Linux (x86)        U834415
  Linux (x86-64)     U834985
  Linux (zSeries)    U834412
  Linux (Power)      U835662
  
  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 WINDOWS V7

• Fixed component ID

  5724H7220

Applicable component levels

• R700 PSY

     UP