| |
|
RSS Feed - WebSphere MQ Support
|
RSS Feed - Message Broker Support
|
  |
|
| When Your Repository Manager Quits, What To Do? |
« View previous topic :: View next topic » |
| Author |
Message
|
| Gerald Ball |
 Posted: Thu Feb 17, 2005 1:08 pm Post subject: When Your Repository Manager Quits, What To Do? |
 |
|
Newbie
Joined: 09 Feb 2005
Posts: 6
|
Looking at the Websphere MQ clustering manual, they call this sad situation:
The repository manager process is not processing repository commands.
When this happens, messages to update the cluster (whether the update was on the local or remote queue manager and whether full or partial repository) just sit on the SYSTEM.CLUSTER.COMMAND.QUEUE. Deleting/adding cluster objects, refresh, reset, repos(' ') etc. does not work because the messages never get processed by the repository manager.
This appears to be the root cause of the problem to which I have posted several request for help the past two weeks and why the helpful suggestions were tried to no avail.
The MQ clustering manual provides absolutely no information on how to get out of this; the only way I have gotten out of this mess in the past is to completely delete the queue manager and build an identical new one. Is there any other way? |
|
| Back to top |
|
 |
| EddieA |
| Posted: Thu Feb 17, 2005 1:49 pm Post subject: |
|
|
Jedi
Joined: 28 Jun 2001
Posts: 2453
Location: Los Angeles
|
Is the Repository process, AMQRRMFA, running. I've seen instances in the past where this just quietly died.
Cheers,
_________________
Eddie Atherton
IBM Certified Solution Developer - WebSphere Message Broker V6.1
IBM Certified Solution Developer - WebSphere Message Broker V7.0 |
|
| Back to top |
|
|
| PeterPotkay |
| Posted: Thu Feb 17, 2005 2:00 pm Post subject: Re: When Your Repository Manager Quits, What To Do? |
|
|
Poobah
Joined: 15 May 2001
Posts: 7722
|
| Gerald Ball wrote: |
| The MQ clustering manual provides absolutely no information on how to get out of this; the only way I have gotten out of this mess in the past is to completely delete the queue manager and build an identical new one. Is there any other way? |
Bouncing the QM didn't work? That should cause the QM and all its children processes to come up clean.
By the way, we monitor the Cluster Command Queue for depth and IPROCS = 0.
_________________
Peter Potkay
Keep Calm and MQ On |
|
| Back to top |
|
|
| Nigelg |
| Posted: Fri Feb 18, 2005 1:09 am Post subject: |
|
|
Grand Master
Joined: 02 Aug 2004
Posts: 1046
|
EddieA, the repos mgr process NEVER 'quietly dies'. It either cuts an FFST because it trapped (usually with probe XC130003/4, SIGSEGV or SIGBUS, an APARable problem), or it writes a sequence of msgs to the error logs with the reason for it ending, usually a configuration problem.
Here is an example of a config problem:
| Quote: |
AMQ9511: Messages cannot be put to a queue ... reason code 2024.
AMQ9448: Repository manager stopping because of errors. Restart in 600 seconds.
...
AMQ9449: Repository manager restarted
AMQ9511: Messages cannot be put to a queue ... reason code 2024.
AMQ9409: Repository manager ended abnormally.
|
The AMQ9511 msg may be varied, as may the reason code. (In this case, 2024 is MQRC_SYNCPOINT_LIMIT_REACHED, and so the remedy is to increase MAXUMSGS.)
The repos mgr can be restarted as Peter says by restarting the qmgr. Also, although it is not recommended, it can be restarted from the command line with its original arguments.
However, if the condition causing it to end is not fixed, it will probably just end again, with one of the above symptoms.
Check the symptoms, and either contact IBM (if it ended with a trap, much as it pains me to recommend this course of action), or deduce the configuration problem behind the reason code mentioned in the error log msg and fix it. |
|
| Back to top |
|
|
|
|
|
| |
|
Page 1 of 1 |
|
You cannot post new topics in this forum
You cannot reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum
You cannot vote in polls in this forum
|
|
|
|
