| Author |
Message
|
| KIT_INC |
 Posted: Tue Apr 08, 2014 6:53 am Post subject: Message still send to stopped Qmgr in the cluster |
 |
|
Knight
Joined: 25 Aug 2006
Posts: 589
|
We are running MQ V701 MQ cluster on Linux.
We have a stopped QMGR (QM1) in our cluster test environment with 5 Qmgrs. A Q (MY.CQ) is shared in the cluster. We found a few messages stuck in the cluster xmitq of a Qmgr in the cluster. The header of these stuck message has the name QM1 and MY.CQ in it. I think the reason for the stuck message was because QM1 is stopped.
I believe QM1 is only stopped but not suspended from the cluster.
Can dis clusqmgr on another Qmgr show me if a QM1 is suspended ? I cannot restart QM1 at this time to check.
Can some one confirmed if MQ cluster logic will still send message to a stopped but not suspended Qmgr ? |
|
| Back to top |
|
 |
| fatherjack |
| Posted: Tue Apr 08, 2014 7:47 am Post subject: Re: Message still send to stopped Qmgr in the cluster |
|
|
Knight
Joined: 14 Apr 2010
Posts: 522
Location: Craggy Island
|
| KIT_INC wrote: |
| Can dis clusqmgr on another Qmgr show me if a QM1 is suspended ? |
Try it and see. Create your own test cluster if you have to.
| KIT_INC wrote: |
| Can some one confirmed if MQ cluster logic will still send message to a stopped but not suspended Qmgr ? |
Read about the BIND option when opening a queue.
_________________
Never let the facts get in the way of a good theory. |
|
| Back to top |
|
|
| PeterPotkay |
| Posted: Tue Apr 08, 2014 10:12 am Post subject: |
|
|
Poobah
Joined: 15 May 2001
Posts: 7722
|
Go to your Full Repositories. They know if any QMs are suspended or not.
There are several cases where MQ will send messages to a queue manager that is suspended, and where it will attempt to send messages to a QM that is stopped. In all these cases its because you gave it no other choice.
_________________
Peter Potkay
Keep Calm and MQ On |
|
| Back to top |
|
|
| KIT_INC |
| Posted: Tue Apr 08, 2014 10:34 am Post subject: |
|
|
Knight
Joined: 25 Aug 2006
Posts: 589
|
I did the cleanup and ensure that QM1 is suspened and removed from the cluster.
DIS CLUSQMGR(*) shows QM1 not in the list anymore. I asume that my clean up is done properly.
But when I display cluster q on a QMGR, I still see
dis qc(*) clusqmgr
AMQ8409: Display Queue details.
QUEUE(MY.CQ.QA) TYPE(QCLUSTER)
CLUSQMGR(QM1)
MY.CQ.QA is an alias to MY.CQ.QL. Both the local Q and the alias were shared in the cluster. After I suspend and remove QM1 from the cluster, MY.CQ.QL no longer shows up in dis qc(*). But this QA still shows. How can I get ride of it ? |
|
| Back to top |
|
|
| fjb_saper |
| Posted: Tue Apr 08, 2014 11:13 am Post subject: |
|
|
Grand High Poobah
Joined: 18 Nov 2003
Posts: 20756
Location: LI,NY
|
That's because although you removed the cluster attribute from the QL you did not do so on the corresponding QA.
If QM1 is still part of the cluster, remove all cluster information from all objects defined in QM1, then remove QM1 from the cluster.
Have fun 
_________________
MQ & Broker admin |
|
| Back to top |
|
|
| KIT_INC |
| Posted: Tue Apr 08, 2014 11:58 am Post subject: |
|
|
Knight
Joined: 25 Aug 2006
Posts: 589
|
| Quote: |
| remove all cluster information from all objects defined in QM1, then remove QM1 from the cluster |
This the point that was not clear in the manual. I was wondering if I first need to remove all cluster Qs specified in QM1 from the cluster before removing QM1 from the cluster. I took the easy way out thinking that removing the entire QMGR from the cluster will also remove the Qs
When I look a little more, I did find other QMGRs in the cluster still show QM1 as the cluster Qmgr when doing a display QC
AMQ8409: Display Queue details.
QUEUE(OTHER.OUT.QL) TYPE(QCLUSTER)
CLUSQMGR(QM1)
If it is true that all cluster Q should be removed from cluster first before removing the cluster Qmgr, now that I have QM1 already removed from the cluster, can I put it back to cluster and removing cluster Qs and remove it again? I guess I'll have to try. |
|
| Back to top |
|
|
| bruce2359 |
| Posted: Tue Apr 08, 2014 12:25 pm Post subject: |
|
|
Poobah
Joined: 05 Jan 2008
Posts: 9469
Location: US: west coast, almost. Otherwise, enroute.
|
Look at the RESET CLUSTER command. If used carefully and wisely, it will delete leftover cluster definitions.
IMS, unused cluster definitions will stale-date in 60 days. Wait patiently, and they will disappear.
_________________
I like deadlines. I like to wave as they pass by.
ב''ה
Lex Orandi, Lex Credendi, Lex Vivendi. As we Worship, So we Believe, So we Live. |
|
| Back to top |
|
|
| fjb_saper |
| Posted: Tue Apr 08, 2014 1:27 pm Post subject: |
|
|
Grand High Poobah
Joined: 18 Nov 2003
Posts: 20756
Location: LI,NY
|
| bruce2359 wrote: |
Look at the RESET CLUSTER command. If used carefully and wisely, it will delete leftover cluster definitions.
IMS, unused cluster definitions will stale-date in 60 days. Wait patiently, and they will disappear. |
I thought that was 90 days?
_________________
MQ & Broker admin |
|
| Back to top |
|
|
| KIT_INC |
| Posted: Tue Apr 08, 2014 7:01 pm Post subject: |
|
|
Knight
Joined: 25 Aug 2006
Posts: 589
|
We do not mind waiting for 90 days. We just concern if there is a left over definition that says
AMQ8409: Display Queue details.
QUEUE(OTHER.OUT.QL) TYPE(QCLUSTER)
CLUSQMGR(QM1)
Will some messages going to OTHER.OUT.QL get into the cluster Xmitq wirh header OTHER.OUT.QL QM1 and they will never get there. |
|
| Back to top |
|
|
|
|
