Your WebSphere MQ (WMQ) queue has the current depth greater than its maximum depth.
The WebSphere MQ explorer and runmqsc may show CURDEPTH > MAXDEPTH:
1 : dis q(ET.LQ)
AMQ8409: Display Queue details.
DESCR(WebSphere MQ Default Local Queue)
PROCESS( ) BOQNAME( )
INITQ( ) TRIGDATA( )
CLUSTER( ) CLUSNL( )
If your WMQ application program backs out the unit of work, or if WebSphere MQ has to back it out, the messages in the unit of work will be placed back on the queue and the CURDEPTH value will be increased even if it exceeds the MAXDEPTH value.
Resolving the problem
Make sure your application is committing messages after the unit of work has been completed. This will avoid large numbers of messages being rolled back to the queue and exceeding the MAXDEPTH limit.
You have an application which was getting messages off the queue as a unit of work with the MQGMO_SYNCPOINT flag, but the unit of work was backed out. When a program issues MQGET for a message, WebSphere MQ decrements the CURDEPTH value even if the MQGET is part of a unit of work. If that program backs out the unit of work, or if WebSphere MQ has to back it out, the messages in the unit of work will be placed back on the queue and the CURDEPTH value will be increased even if it now exceeds the MAXDEPTH value.
01322 004 000