| Author |
Message
|
| vmcgloin |
 Posted: Thu Jan 22, 2004 6:47 am Post subject: AMQ9456 cluster error |
 |
|
Knight
Joined: 04 Apr 2002
Posts: 560
Location: Scotland
|
Hi,
Has anyone seen the AMQ9456 error below? Any suggestions what to do about it? We have been seeing these warning on and off for months (since upgrading to 5.3). In one case now we have come across a situation where a clustered queue has actually reached the limits and been removed from a partial repository - running refresh cluster did the trick to bring it back, but we are getting lots of these warnings.
| Quote: |
AMQ9456: Update not received for queue HUB.REQUEST.SVR.EYQUOG06, queue manager
MQQ2.B6BF1A4ADD8F77C0 from full repository for cluster CALQ.
EXPLANATION:
The repository manager detected a queue that has been used in the last 30 days for which updated information should have been sent from a full repository. However, this has not occurred.
The repository manager will keep the information about this queue for a further
60 days.
ACTION:
If the queue is still required, check that:
1) The cluster channels to and from the full repository and the queue manager that hosts the queue, are able to run.
2) The repository managers running on these queue managers have not ended abnormally. |
Mysteriously AMQ9456 is missing from the messages manual.
All the channels and repository managers are OK.
Thanks,
Vicky |
|
| Back to top |
|
 |
| JasonE |
| Posted: Thu Jan 22, 2004 7:37 am Post subject: |
|
|
Grand Master
Joined: 03 Nov 2003
Posts: 1220
Location: Hursley
|
So your qmgr connected to the cluster and someone did an MQOPEN for the queue. At that point it puts a subscription on the Full repos for that queue, which needs to be refreshed every 28 days. No such refresh was made, and hence the error message.
Have you cold started at any point, cleared repos queues, refresh cluster, rebuild full or partial repos etc
I suspect all will be ok for about another 60 days... After 90 days we throw away definitions which have not been refreshed.
Is the qmgr which issues the msg a full or partial? |
|
| Back to top |
|
|
| vmcgloin |
| Posted: Thu Jan 22, 2004 8:00 am Post subject: |
|
|
Knight
Joined: 04 Apr 2002
Posts: 560
Location: Scotland
|
| Quote: |
| So your qmgr connected to the cluster and someone did an MQOPEN for the queue. At that point it puts a subscription on the Full repos for that queue, which needs to be refreshed every 28 days. No such refresh was made, and hence the error message. |
Yes, most of the queues are being regularly used. But, why is the refresh not made?
| Quote: |
| Have you cold started at any point, cleared repos queues, refresh cluster, rebuild full or partial repos etc |
None of the above except refresh cluster commands. However there is definitely one enironment (production) in which the warning is occurring but no refresh cluster has been issued.
| Quote: |
| I suspect all will be ok for about another 60 days... After 90 days we throw away definitions which have not been refreshed. |
We think this has only occurred in on one queue manager were we had some messages going to the DLQ due to UNKNOWN_OBJECT_NAME (2085) for objects that are hosted remotely but shared in the cluster. Refresh cluster seemed to solve that particular problem. Is that what you would recommended at that point?
| Quote: |
| Is the qmgr which issues the msg a full or partial? |
We are getting the warning from both full and partial repositories.
Are you the lucky person that will eventually get passed this when I raise a PMR?
Thanks,
Vicky |
|
| Back to top |
|
|
| JasonE |
| Posted: Thu Jan 22, 2004 8:24 am Post subject: |
|
|
Grand Master
Joined: 03 Nov 2003
Posts: 1220
Location: Hursley
|
| Quote: |
| Refresh cluster seemed to solve that particular problem. |
Yes it would, but its avoiding a problem rather than resolving it.
Are you seeing any of:
1. Messages hung on the system.cluster.transmit.queue
2. Build up of messages on the system.cluster.command.queue
3. Amqrrmfa not running on the qmgrs in question?
This should all happen automatically, it should not drop definitions for objects in use. If it does, it needs to be fixed
Me for the PMR - Maybe, depending on platform and workload etc. |
|
| Back to top |
|
|
| RatherBeGolfing |
| Posted: Thu Jan 22, 2004 8:33 am Post subject: |
|
|
Centurion
Joined: 12 Nov 2002
Posts: 118
Location: Syracuse, NY, USA
|
Vicky,
Just so you know you are not alone, I've been seeing these messages in my clustered queue managers' log files also. What CSD are you running with MQV5.3 - I'm at CSD04. I am running on Win2000 - I have 8 servers in my cluster with 2 full repositories.
Jason,
I too have not issued any refresh commands for a very long time, there are no messages building up on either of the system Qs you mention and the amqrrmfa is running on each of the queue managers where I have seen these log entries.
Here's one from 90 minutes ago.
-------------------------------------------------------------------------------
01/22/2004 09:55:37
AMQ9456: Update not received for queue FA.COM.INQUIRY, queue manager
MOSY1251_2003-07-18_14.26.02 from full repository for cluster DAASSTP.
EXPLANATION:
The repository manager detected a queue that has been used in the last 30 days
for which updated information should have been sent from a full repository.
However, this has not occurred.
The repository manager will keep the information about this queue for a further
60 days.
ACTION:
If the queue is still required, check that:
1) The cluster channels to and from the full repository and the queue manager
that hosts the queue, are able to run.
2) The repository managers running on these queue managers have not ended
abnormally.
-------------------------------------------------------------------------------
_________________
Cheers,
Larry
MQ Certifiable |
|
| Back to top |
|
|
| vmcgloin |
| Posted: Thu Jan 22, 2004 8:57 am Post subject: |
|
|
Knight
Joined: 04 Apr 2002
Posts: 560
Location: Scotland
|
Good to hear we are not alone.
Ditto, no messages building up on either of those queues, and the repository manager processes are running.
The only suspicious thing I have seen is that the summary option of amqrfdm shows a warning:
**** WARNING **** Some data has gone missing **** WARNING ****
I don't know if this is usual or not.
Larry, I am just in the process of raising the PMR. So I'll report the outcome back here. We are at 5.3 CSD04 on AIX.
Cheers,
Vicky |
|
| Back to top |
|
|
| JasonE |
| Posted: Thu Jan 22, 2004 9:45 am Post subject: |
|
|
Grand Master
Joined: 03 Nov 2003
Posts: 1220
Location: Hursley
|
| No, that warnng can be ignored. If the queue hasnt been used, it wouldnt suprise me for the information not to be sent, but if you are seeing this for queues actively in use it sounds more worrying. I know a refresh will avoid it, but you really want to push for a solution. Otherwise the first you will see of it is 2085 in a further 60 days time |
|
| Back to top |
|
|
| RatherBeGolfing |
| Posted: Thu Feb 05, 2004 8:15 am Post subject: |
|
|
Centurion
Joined: 12 Nov 2002
Posts: 118
Location: Syracuse, NY, USA
|
After I reported another symptom of my problem - namely my SYSTEM.CLUSTER.TRANSMIT.QUEUEs were clogging because some of my auto-defined cluster sender channels were not starting (they were in an inactive state) - the change team sent my tech contact a fresher version of amqknl2.dll.
I'm planning on pushing this newer dll into my 24 servers over the next 4 days. Tech said that because the dll corrects intra-cluster communication problems, it should correct the AMQ9456 errors as well. I hope he is correct!
_________________
Cheers,
Larry
MQ Certifiable |
|
| Back to top |
|
|
| vmcgloin |
| Posted: Thu Feb 05, 2004 8:27 am Post subject: |
|
|
Knight
Joined: 04 Apr 2002
Posts: 560
Location: Scotland
|
Thanks for the update Larry.
Our problem appears to be different, since there is no build up of messages on the cluster transmission queues, and channels are starting fine (& the platforms are different).
It seems that we get the same errors on z/OS, and support are aware of problems from at least 2 other PMRs but have not got to the bottom of them yet. At the moment it seems like the advice is to run a refresh cluster on all partial repositories every 30 days, but we are waiting for clarification on that.
Cheers,
Vicky |
|
| Back to top |
|
|
| andye |
| Posted: Tue Mar 02, 2004 11:17 am Post subject: |
|
|
Newbie
Joined: 11 Nov 2003
Posts: 7
Location: HURSLEY
|
Hello,
The problem relating to AMQ9456 has now been resolved by APARs IC39601 (distributed platforms) & PQ85250 for z/OS. Fixes could be made available but you would need to go through support channels to obtain them, otherwise, the fix will be available in CSD08.
Hope this info is useful  |
|
| Back to top |
|
|
| offshore |
| Posted: Tue Mar 02, 2004 12:33 pm Post subject: |
|
|
Master
Joined: 20 Jun 2002
Posts: 222
|
We are getting this same message at my shop. We are running MQ 5.3 for Windows along with MQ 5.3 for z/OS.
So far I've only seen on it the partial repos, which happen to be on the NT boxes. |
|
| Back to top |
|
|
| RatherBeGolfing |
| Posted: Tue Mar 02, 2004 1:13 pm Post subject: |
|
|
Centurion
Joined: 12 Nov 2002
Posts: 118
Location: Syracuse, NY, USA
|
For what it's worth, no harm has befallen any of my applications because of the warnings issued by AMQ9456.
And, despite the wording that "... the repository manager will keep the information about this queue for a further 60 days...", we are well beyond 60 days since we first started seeing the AMQ errors and nothing has happened to the information about the queues.
I'm going to pursue getting whatever DLLs I need to replace under APAR IC39601, but I'm not sweating it out any longer!
_________________
Cheers,
Larry
MQ Certifiable |
|
| Back to top |
|
|
|
|
