| Author |
Message
|
| sammy123 |
 Posted: Fri Apr 23, 2010 11:58 pm Post subject: removing PR Qmgrs from Cluster |
 |
|
Newbie
Joined: 23 Apr 2010
Posts: 5
|
Hi All,
I have a cluster of four queue managers(2 FR & 2 PR) and it's running fine. Now I have to remove 2 partial repository queue managers from the clusters and I don't have access to the full repository queue managers. Here is my steps to remove the partial repository QM.
1) SUSPEND QMGR CLUSTER(<clusName>)
2) Remove the CLUSRCVR channel from the cluster
3) Stop the CLUSRCVR channel
4) Stop and delete the CLUSSDR channel
5) REFRESH CLUSTER(<ClusName>) REPOS(YES)
are the above steps perfect? will there can be any impact on the other clusters with these queue managers.
Please advise.
Thanks in advance... |
|
| Back to top |
|
 |
| exerk |
| Posted: Sat Apr 24, 2010 12:52 am Post subject: |
|
|
Jedi Council
Joined: 02 Nov 2006
Posts: 6339
|
Just follow the manual and remember to give channels time to stop of their own accord, only force them down if you have to. You missed one step at the beginning, which is to remove any queues from the cluster and wait for any in-flight messages (in-flights will DLQ) and as to why you'd want to issue the refresh command is beyond me - look more closely at the manual to see what exactly that command does, and what effect it will have on a queue manager that is not in a cluster.
_________________
It's puzzling, I don't think I've ever seen anything quite like this before...and it's hard to soar like an eagle when you're surrounded by turkeys. |
|
| Back to top |
|
|
| sammy123 |
| Posted: Sun Apr 25, 2010 11:13 pm Post subject: |
|
|
Newbie
Joined: 23 Apr 2010
Posts: 5
|
thanks for your reply !!!
you are right that I have missed the one step in the starting ....
I think, refresh cluster command shlud be executed from the repositiry queue manager ? |
|
| Back to top |
|
|
| exerk |
| Posted: Sun Apr 25, 2010 11:29 pm Post subject: |
|
|
Jedi Council
Joined: 02 Nov 2006
Posts: 6339
|
| sammy123 wrote: |
thanks for your reply !!!
you are right that I have missed the one step in the starting ....
I think, refresh cluster command shlud be executed from the repositiry queue manager ? |
If you mean the Full Repositories yes, and as for the queue manager you will have removed from the cluster, I refer you to my previous statement; "...look more closely at the manual to see what exactly that command does, and what effect it will have on a queue manager that is not in a cluster...".
_________________
It's puzzling, I don't think I've ever seen anything quite like this before...and it's hard to soar like an eagle when you're surrounded by turkeys. |
|
| Back to top |
|
|
| PeterPotkay |
| Posted: Mon Apr 26, 2010 1:45 am Post subject: |
|
|
Poobah
Joined: 15 May 2001
Posts: 7722
|
There is no need to execute Refresh Cluster, from anywhere, just to remove a Partial QM from the cluster.
The Refresh Cluster command does not refresh the cluster.
_________________
Peter Potkay
Keep Calm and MQ On |
|
| Back to top |
|
|
| sammy123 |
| Posted: Mon Apr 26, 2010 8:51 pm Post subject: |
|
|
Newbie
Joined: 23 Apr 2010
Posts: 5
|
You mean just by removing cluster queues, CLUSSDR and CLUSRCVR channels will completly remove the queue manager from the cluster?
There is no need to refresh cluster... |
|
| Back to top |
|
|
| exerk |
| Posted: Mon Apr 26, 2010 9:15 pm Post subject: |
|
|
Jedi Council
Joined: 02 Nov 2006
Posts: 6339
|
Once more, and with feeling...
If you have followed the instructions detailed in the manual and removed all cluster-related objects, i.e. the CLUSSDR/CLUSRCVR channels, the queue manager is no longer in the cluster, so the REFRESH command would have no effect! Follow the logic - how could a queue manager which is no longer in the cluster update its information about that cluster when it no longer has any means of communicating with that cluster?
_________________
It's puzzling, I don't think I've ever seen anything quite like this before...and it's hard to soar like an eagle when you're surrounded by turkeys. |
|
| Back to top |
|
|
| sammy123 |
| Posted: Wed Apr 28, 2010 6:44 pm Post subject: |
|
|
Newbie
Joined: 23 Apr 2010
Posts: 5
|
Here is what I found out about the refresh cluser with repository(YES)
Using REFRESH CLUSTER(clustername) REPOS(YES) specifies that in addition to the default behavior, objects representing full repository cluster queue managers are also refreshed. This option may not be used if the queue manager is itself a full repository. If it is a full repository, you must first alter it so that it is not a full repository for the cluster in question. The full repository location will be recovered from the manually defined CLUSSDR definitions. After the refresh with REPOS(YES) has been issued the queue manager can be altered so that it is once again a full repository, if required.
It seems that this command is basically to update the repository qmgr inofrmation in the PR qmgr. So if we are trying to remove a PR from the cluster, this command should be executed to update that there is no repository queue manager for this cluster and it should not try to send any info to the repository qmgr.
Please correct me if I am wrong |
|
| Back to top |
|
|
| Vitor |
| Posted: Wed Apr 28, 2010 7:16 pm Post subject: |
|
|
Grand High Poobah
Joined: 11 Nov 2005
Posts: 26093
Location: Texas, USA
|
| sammy123 wrote: |
| Please correct me if I am wrong |
If you read back what you've posted, it says:
| Quote: |
| The full repository location will be recovered from the manually defined CLUSSDR definitions |
According to the instructions and your own post, the sequence of events is:
| Quote: |
4) Stop and delete the CLUSSDR channel
5) REFRESH CLUSTER(<ClusName>) REPOS(YES) |
So how can you think you're right when you're contradicting yourself?
In this thread, you've been told twice you don't need to issue a REFRESH command and you yourself have proved it can't work. So why are you so obsessed with using it? 
_________________
Honesty is the best policy.
Insanity is the best defence. |
|
| Back to top |
|
|
| sumit |
| Posted: Thu Apr 29, 2010 2:07 am Post subject: |
|
|
Partisan
Joined: 19 Jan 2006
Posts: 398
|
| exerk wrote: |
| sammy123 wrote: |
thanks for your reply !!!
you are right that I have missed the one step in the starting ....
I think, refresh cluster command shlud be executed from the repositiry queue manager ? |
If you mean the Full Repositories yes.. |
I am wondering.. are you sure? Refresh Cluster in a way means cold restart. I would never dare to do that on FR unless the situation is so dire. (even on small cluster setup, it's not a good practice).
I agree with you all that there is no use of executing refresh cluster command once the channel is deleted. However, as you are aware, refresh cluster on PR also flush out all the unnecessary cluster definition. Issue it on a PR would clean it up and get it ready for future use.
_________________
Regards
Sumit |
|
| Back to top |
|
|
| Vitor |
| Posted: Thu Apr 29, 2010 5:30 am Post subject: |
|
|
Grand High Poobah
Joined: 11 Nov 2005
Posts: 26093
Location: Texas, USA
|
| sumit wrote: |
| However, as you are aware, refresh cluster on PR also flush out all the unnecessary cluster definition. Issue it on a PR would clean it up and get it ready for future use. |
Or you could do what the rest of us do & leave the queue manager to clean these out itself. It's ready for "future use" as soon as it's left the cluster, even if the "future use" is to join another cluster.
More seriously, on receiveing this command the PR will try (as described above) to recover it's FR from the cluster sender. This will (in this scenario) cause the repository process to beat itself up for a while & then abend. That's not especially clean even if harmless.
_________________
Honesty is the best policy.
Insanity is the best defence. |
|
| Back to top |
|
|
| sumit |
| Posted: Thu Apr 29, 2010 9:53 am Post subject: |
|
|
Partisan
Joined: 19 Jan 2006
Posts: 398
|
| Vitor wrote: |
| More seriously, on receiveing this command the PR will try (as described above) to recover it's FR from the cluster sender. This will (in this scenario) cause the repository process to beat itself up for a while & then abend. That's not especially clean even if harmless. |
Here is a quote from quick beginnings-
| Quote: |
| Issue a refresh cluster command against the queue manager for the cluster that the queue manager is leaving. This ensures that the information about the cluster is cleared from the repository held by the queue manager. |
I know this covers a scenario wherein qmgr could be a part of multiple clusters. However, I believe it should be same even if the qmgr is in 1 cluster. In both the cases refresh command will work only for the mentioned cluster name. Is it still a malpractice?
_________________
Regards
Sumit |
|
| Back to top |
|
|
| Vitor |
| Posted: Thu Apr 29, 2010 10:43 am Post subject: |
|
|
Grand High Poobah
Joined: 11 Nov 2005
Posts: 26093
Location: Texas, USA
|
| sumit wrote: |
I know this covers a scenario wherein qmgr could be a part of multiple clusters. |
In a grumpy way I could point out that as a WMQ administrator there should be no could be about cluster membership; you should have proper control of your environment.
| sumit wrote: |
| Here is a quote from quick beginnings |
Where? Post the link, I'm having trouble spotting it.
I prefer this link, which says:
| Quote: |
| It is not normally necessary to issue a REFRESH CLUSTER command except in one of the following circumstances |
None of which circumstances are the normal removal of a queue manager from a cluster.
_________________
Honesty is the best policy.
Insanity is the best defence. |
|
| Back to top |
|
|
| bruce2359 |
| Posted: Thu Apr 29, 2010 2:51 pm Post subject: |
|
|
Poobah
Joined: 05 Jan 2008
Posts: 9471
Location: US: west coast, almost. Otherwise, enroute.
|
IMS, repository data stales out after 91 days. Thus, no action beyond deleting the cluster qmgrs cluster channels is necessary. Clustering software is self-cleaning.
_________________
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 |
|
|
| sumit |
| Posted: Fri Apr 30, 2010 9:09 am Post subject: |
|
|
Partisan
Joined: 19 Jan 2006
Posts: 398
|
| Vitor wrote: |
| sumit wrote: |
| Here is a quote from quick beginnings |
Where? Post the link, I'm having trouble spotting it. |
Please find the link http://publib.boulder.ibm.com/infocenter/wmqv6/v6r0/index.jsp?topic=/com.ibm.mq.csqzah.doc/qc12140_.htm. I took that passage from quick beginning pdf.
| Vitor wrote: |
I prefer this link, which says:
| Quote: |
| It is not normally necessary to issue a REFRESH CLUSTER command except in one of the following circumstances |
|
Which also says
| Quote: |
| Issuing a REFRESH CLUSTER command has been recommended by IBM® Service. |
And the same is mentioned in the link that I provided.
_________________
Regards
Sumit |
|
| Back to top |
|
|
|
|
