IT19794: During Queue Manager failover MQ JMS client may send duplicate message

APAR status

• Closed as program error.

Error description

• Additional Symptom(s) Search Keyword(s):
  While sending a message using the method
  MessageProducer.send(Message), the JMS client detected that
  the queue manager had failed over and performed
  automatic client reconnection. This was done before waiting for
  the queue manager to confirm that the message was put
  successfully.  Because the queue manager failed over after it
  had received the message and before it could send back a reply,
  the message was successfully put using a non-transacted JMS
  session.
  
  After reconnecting, the JMS client sent the message again.
  This resulted in the message arriving on the queue twice.
  

Local fix

Problem summary

• ****************************************************************
  USERS AFFECTED:
  This issue affects users of:
  
  - The WebSphere MQ V7.5 classes for JMS.
  - The IBM MQ V8.0 classes for JMS.
  - The IBM MQ V9.0.0.x classes for JMS.
  
  who have JMS applications that:
  
  - Use a connection factory that has the CLIENTRECONNECTOPTIONS
  property set to either QMGR or ANY to create a JMS connection to
  a multi-instance queue manager.
  - Create a non-transacted JMS session from the JMS connection.
  - Create a JMS MessageProducer from the JMS session for a
  destination hosted on the multi-instance queue manager, where
  the destination has its PERSISTENCE property configured such
  that non-persistent messages are sent to the destination.
  - Use the JMS MessageProducer to send messages to the
  destination.
  
  
  Platforms affected:
  MultiPlatform
  
  ****************************************************************
  PROBLEM DESCRIPTION:
  When an application using the WebSphere MQ classes for JMS
  connects to a multi-instance queue manager, using a Connection
  Factory that has been configured to use automatic client
  reconnection (as documented in the Automatic JMS client
  reconnection topic in the MQ V8 and later sections of IBM
  Knowledge Center :
  https://www.ibm.com/support/knowledgecenter/SSFKSJ_8.0.0/com.ibm
  .mq.dev.doc/q032520_.htm), then the WebSphere MQ classes for JMS
  will receive a notification from the queue manager when it fails
  over. When the notification is received, the WebSphere MQ
  classes for JMS will pause the application the next time it
  tries to communicate with the queue manager and periodically
  attempt to reconnect. When the multi-instance queue manager is
  available again, and the WebSphere MQ classes for JMS have
  reconnected, they will resume the application and allow it to
  continue processing.
  
  When an application using the WebSphere MQ classes for JMS sends
  a message to a JMS destination by calling one of the following
  methods:
  
  - MessageProducer.send(Destination, Message)
  - MessageProducer.send(Destination, Message, CompletionListener)
  - MessageProducer.send(Destination, Message, int, int, int)
  - MessageProducer.send(Destination, Message, int, int, int,
  CompletionListener)
  - MessageProducer.send(Message)
  - MessageProducer.send(Message, CompletionListener)
  - MessageProducer.send(Message, int, int, int)
  - MessageProducer.send(Message, int, int, int,
  CompletionListener)
  
  the WebSphere MQ classes for JMS will:
  
  - Send an MQPUT request to the queue manager.
  - Wait for an MQPUT reply from the queue manager, indicating
  that the message was put successfully.
  
  If a multi-instance queue manager failed over during these
  operations, then the WebSphere MQ classes for JMS would abort
  the current MQPUT operation and reconnect to the queue manager.
  After reconnecting, the WebSphere MQ classes for JMS would
  perform the MQPUT call again, to send the message for a second
  time.
  
  If the application was using a non-transacted JMS session and
  the messages that were being sent were non-persistent,  then if
  the queue manager fail over occurred:
  
  - After the MQPUT request was sent to the queue manager
  - And before the MQPUT reply was received
  
  the message would be put to the destination twice. Once before
  the queue manager failed over, and for a second time after the
  WebSphere MQ classes for JMS had reconnected.
  

Problem conclusion

• The WebSphere MQ classes for JMS have been modified so that, if
  automatic client reconnection occurs while sending a message to
  a JMS destination using one of the methods shown below:
  
  - MessageProducer.send(Destination, Message)
  - MessageProducer.send(Destination, Message, CompletionListener)
  - MessageProducer.send(Destination, Message, int, int, int)
  - MessageProducer.send(Destination, Message, int, int, int,
  CompletionListener)
  - MessageProducer.send(Message)
  - MessageProducer.send(Message, CompletionListener)
  - MessageProducer.send(Message, int, int, int)
  - MessageProducer.send(Message, int, int, int,
  CompletionListener)
  
  the WebSphere MQ classes for JMS will reconnect to the
  multi-instance queue manager and:
  
  - Return from the method normally, if the MessageProducer was
  created from a non-transacted JMS session and was being used to
  put a non-persistent message to a destination.
  - Throw a JMSException containing reason code 2003
  (MQRC_BACKED_OUT) if the application was using a MessageProducer
  created from a transacted JMS session.
  - Throw a JMSException containing reason code 2549
  (MQRC_CALL_INTERRUPTED) in all other situations.
  
  ---------------------------------------------------------------
  The fix is targeted for delivery in the following PTFs:
  
  Version    Maintenance Level
  v7.5       7.5.0.9
  v8.0       8.0.0.7
  v9.0 LTS   9.0.0.2
  
  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 BASE MULTIP

• Fixed component ID

  5724H7241

Applicable component levels