IT14138: DUPLICATE MESSAGE IDS MAY BE SEEN WHEN USING ASYNCHRONOUS PUT

APAR status

• Closed as program error.

Error description

• When using asynchronous put (MQPMO_ASYNC_RESPONSE) with a C or
  C++ client it is possible that duplicate message IDs will be
  seen. An MQPUT1 in syncpoint will put messages asynchronously by
  default, and asynchronous puts will also occur where by default
  where a queue has the attribute DEFPRESP(ASYNC).
  Java and .NET clients are not susceptible to this problem, but
  will instead report an error. A Java client will return
  MQRC_UNEXPECTED_ERROR and log the following FDC:
  Source Class   :- com.ibm.mq.jmqi.remote.impl.RemoteTagPool
  Source Method  :- getTags(RemoteTls,int,RemoteSession)
  ProbeID     :- 06
                     Data
                     ----
  |  Description     :- Queue manager returned counter which will
  wrap
  |  maxCounterVal    :- 256
  |  tagsInCounterRange :- 255
  A Java client will also throw an exception with reason code
  MQRC_UNEXPECTED_ERROR, and a stack similar to the following:
  Caused by: com.ibm.mq.jmqi.JmqiException: CC=2;RC=2195
   at com.ibm.mq.jmqi.remote.impl.RemoteTagPool.getTags(RemoteTagP
  ool.java:398)
   at com.ibm.mq.jmqi.remote.api.RemoteFAP.querySyncDelivery(Remot
  eFAP.java:858)
   at com.ibm.mq.jmqi.remote.api.RemoteFAP.jmqiPutMessageWithProps
  (RemoteFAP.java:8942)
   at com.ibm.mq.jmqi.remote.api.RemoteFAP.jmqiPut1(RemoteFAP.java
  :8195)
   at com.ibm.mq.ese.jmqi.InterceptedJmqiImpl.doPut1(InterceptedJm
  qiImpl.java:1122)
   at com.ibm.mq.ese.jmqi.InterceptedJmqiImpl.jmqiPut1(Intercepted
  JmqiImpl.java:970)
   at com.ibm.mq.ese.jmqi.ESEJMQI.jmqiPut1(ESEJMQI.java:670)
   at com.ibm.msg.client.wmq.internal.WMQMessageProducer$SpiUniden
  tifiedProducerShadow.sendInternal(WMQMessageProducer.java:995)
   ... 11 more
  A .NET client will also throw an exception with reason code
  MQRC_UNEXPECTED_ERROR and the following text:
  Queue manager returned counter which will wrap
  MQRC_UNEXPECTED_ERROR
  

Local fix

Problem summary

• ****************************************************************
  USERS AFFECTED:
  All users of asynchronous put on a little endian platform,
  including users performing MQPUT1 in syncpoint.
  
  
  Platforms affected:
  Linux on Power, Linux on x86, Linux on x86-64, Solaris x86-64,
  Windows
  
  ****************************************************************
  PROBLEM DESCRIPTION:
  The queue manager logic which generates the message ID for a new
  message did not correctly account for the byte order on little
  endian platforms. Because of this is was possible under certain
  conditions for a message ID to be set which had already been
  used.
  
  This problem can only occur on little endian platforms, and only
  where asynchronous put is used.
  

Problem conclusion

• The miscalculation in the queue manager logic has been rectified
  and duplicate message IDs will no longer be generated in this
  scenario.
  
  ---------------------------------------------------------------
  The fix is targeted for delivery in the following PTFs:
  
  Version    Maintenance Level
  v7.1       7.1.0.8
  v7.5       7.5.0.7
  v8.0       8.0.0.5
  
  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

• R710 PSY

     UP