The Q Capture program should allow only one message in its restart queue. In rare cases, IBM has seen multiple messages in this queue. The problem is addressed in V9.7 APAR PM61218 and V10.1 APAR PM60137, which prompt Q Capture to stop with error message ASN7528E if it encounters multiple messages in the restart queue. Until these fixes are applied, customers are advised to look at the restart queue to make sure it contains only one message. You should also alter the restart queue to limit the number of messages to one by setting MAXDEPTH(1). If you have multiple messages in the restart queue, recovering from this problem might require a cold start, which causes a new load (full refresh) of target tables.
The Q Capture restart queue is a WebSphere MQ local queue that is intended to hold a single message that tells the Q Capture program where to start reading in the recovery log on restart. Q Capture should not start or continue to run if the restart queue holds more than one message, but the program is allowing multiple messages in this queue.
This situation can cause Q Capture to send messages with an incorrect timestamp, prompting the Q Apply program to delete messages from the receive queue that should be applied to target tables.
You should alter the restart queue so that it is limited to one message. For example:
If the restart queue contains more than one message, follow these steps:
1. Run the message formatting utility, QMFMT, for the restart queue:
//QMFMT1 EXEC PGM=ASNQMFMT,REGION=0M,TIME=NOLIMIT,
// PARM='QREPL.RESTARTQ CSQ1 -mqmd'
2. Send the output to IBM support. IBM will help you attempt to start Q Capture from a specific point in the log.
3. Clear the restart queue. You can use the following WebSphere MQ command:
CLRMQMQ QNAME('QREPL.RESTARTQ') MQMNAME(CSQ1)
4. In the worst case, you might need to start Q Capture with startmode=cold.
|Information Management||InfoSphere Replication Server||Q Capture||AIX, HP-UX, Linux, Solaris, Windows||9.1, 9.5, 9.7|