ASG
IBM
Zystems
Cressida
Icon
Netflexity
 
  MQSeries.net
Search  Search       Tech Exchange      Education      Certifications      Library      Info Center      SupportPacs      LinkedIn  Search  Search                                                                   FAQ  FAQ   Usergroups  Usergroups
 
Register  ::  Log in Log in to check your private messages
 
RSS Feed - WebSphere MQ Support RSS Feed - Message Broker Support

MQSeries.net Forum Index » Clustering » What happens to PRs if a FR is recreated?

Post new topic  Reply to topic Goto page 1, 2  Next
 What happens to PRs if a FR is recreated? « View previous topic :: View next topic » 
Author Message
Sam Uppu
PostPosted: Mon Nov 17, 2008 9:11 am    Post subject: What happens to PRs if a FR is recreated? Reply with quote

Yatiri

Joined: 11 Nov 2008
Posts: 610

Hi Guys,
We wre on SunOS and MQ version 6 fixpack 2.

We have 2 Full repositories and 20 partial repositories in the same cluster.

These 22 QMgrs(2 FRs+20 PRs) in two different datacenters.

1. One FR resides in one datacenter and 10 PRs also resides in the same datacenter.
The partial repositories cluster sender channels are pointing to Full repository in the same data center where the full repository resides.

2. The other Fullrepository resides in a different datacenter along with other 10 partial repositories.

The partial repositories cluster sender channels are pointing to Full repository in the same data center where the full repository resides.

The cluster sender channels are defined between both Full repositories.

We are planning to recreate one of the full repository. Is there any affect to the partial repositories which are connected to it with the recreation of FR?. When we recreate the FR, it will have a new UUID(QMID). Do you have guys have ever faced such kind of issue?.

Thanks for your thoughts.
Back to top
View user's profile Send private message
exerk
PostPosted: Mon Nov 17, 2008 9:40 am    Post subject: Reply with quote

Jedi Council

Joined: 02 Nov 2006
Posts: 6339

There is a post somewhere on the site that illustrates what can happen if FR's are 'blitzed' and rebuilt; in that case it was both FR's.

I would expect that a RESET CLUSTER QMID will be necessary, to flush the old references.

It might be better to migrate the FR to another queue manager, or derate the FR to a PR and run on one FR until the rebuild is complete.
_________________
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
View user's profile Send private message
xhaxk
PostPosted: Mon Nov 17, 2008 9:42 am    Post subject: Reply with quote

Apprentice

Joined: 30 Oct 2008
Posts: 31

don't do this, if by recreate you mean delete the qmgr and create it again

all hell will break loose in the cluster when the name of the FR in the rest of the qmgrs in the cluster no longer matches its UUID

if you must do this, remove the old FR from the cluster, create a new one with a different name and point all the manual cluster senders to the new FR
Back to top
View user's profile Send private message
PeterPotkay
PostPosted: Mon Nov 17, 2008 10:18 am    Post subject: Reply with quote

Poobah

Joined: 15 May 2001
Posts: 7722

Make the FR a PR.
Uncluster the PR.
Delete the PR QM.
Recreate the QM.
Make it a FR again.

The cluster will be fine. Follow the detailed instructions for each of the above steps in the Clusters manual. Don't do all 5 steps via script and expect it to work in 10 seconds. You need some time for each step to propogate thru the cluster.

If done correctly, there will be no need for RESET CLUSTER or REFRESH CLUSTER. Follow the steps in the manual!


Ideally, you would recreate the QM with a new name ahead of time, repoint all the PRs to the new FR, then de-cluster and delete the old FR.
_________________
Peter Potkay
Keep Calm and MQ On
Back to top
View user's profile Send private message
Sam Uppu
PostPosted: Mon Nov 17, 2008 10:49 am    Post subject: Reply with quote

Yatiri

Joined: 11 Nov 2008
Posts: 610

PeterPotkay wrote:
Make the FR a PR.
Uncluster the PR.
Delete the PR QM.
Recreate the QM.
Make it a FR again.

The cluster will be fine. Follow the detailed instructions for each of the above steps in the Clusters manual. Don't do all 5 steps via script and expect it to work in 10 seconds. You need some time for each step to propogate thru the cluster.

If done correctly, there will be no need for RESET CLUSTER or REFRESH CLUSTER. Follow the steps in the manual!


Ideally, you would recreate the QM with a new name ahead of time, repoint all the PRs to the new FR, then de-cluster and delete the old FR.


We need to reboot the FR QMgr server as there is a patch needs to be installed. As part of server reboot, all the MQ objects are defined from scratch which includes deleteing the QMgr and recreating it(we will modify our scripts not to delete the QMgr in the future but we need to do it for now).

Any necessary steps I need to follow?.

Thanks.
Back to top
View user's profile Send private message
Vitor
PostPosted: Mon Nov 17, 2008 10:53 am    Post subject: Reply with quote

Grand High Poobah

Joined: 11 Nov 2005
Posts: 26093
Location: Texas, USA

Sam Uppu wrote:
As part of server reboot, all the MQ objects are defined from scratch which includes deleteing the QMgr and recreating it


Who's bright idea was this?? What is it supposed to achieve?

Sam Uppu wrote:
Any necessary steps I need to follow?.


Yes - demote and uncluster the queue manager as Peter (and the documentation!) suggests.
_________________
Honesty is the best policy.
Insanity is the best defence.
Back to top
View user's profile Send private message
bruce2359
PostPosted: Mon Nov 17, 2008 11:40 am    Post subject: Reply with quote

Poobah

Joined: 05 Jan 2008
Posts: 9469
Location: US: west coast, almost. Otherwise, enroute.

Quote:
We need to reboot the FR QMgr server as there is a patch needs to be installed.

Do NOT, do NOT, perform both the o/s fix and MQ stuff at the same time.

Do the o/s fix first; then, if MQ and everyting else works OK, then do the MQ cluster stuff. Why? If you make two major changes, and something doesn't work, which is the culprit?
_________________
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
View user's profile Send private message
Sam Uppu
PostPosted: Mon Nov 17, 2008 1:30 pm    Post subject: Reply with quote

Yatiri

Joined: 11 Nov 2008
Posts: 610

PeterPotkay wrote:
Make the FR a PR.
Uncluster the PR.
Delete the PR QM.
Recreate the QM.
Make it a FR again.

The cluster will be fine. Follow the detailed instructions for each of the above steps in the Clusters manual. Don't do all 5 steps via script and expect it to work in 10 seconds. You need some time for each step to propogate thru the cluster.

If done correctly, there will be no need for RESET CLUSTER or REFRESH CLUSTER. Follow the steps in the manual!


Ideally, you would recreate the QM with a new name ahead of time, repoint all the PRs to the new FR, then de-cluster and delete the old FR.


Hi Guys,
I commented out the QMgr deletion part for Full repositories in our MQ shutdown/ startup script. When the server rebooted, the QMgr started(older Qmgr) and the cluster sender/receiver channels have been defined pointing to other full repository. Now the cluster look good.

As we have many partial repository servers need to be rebooted for patching, is that ok if we delete and recreate the partial repository?.

One thing I wanted to share with you guys:
1. the full repository cluster sender/receiver channels were altered(changed the cluster attribute to Null).
2. Stopped the cluster sender/receiver channels.
3. Deleted the cluster sender/receiver channels.
Note: I was able to delete the cluster sender channel but while deleteing the cluster receiver channel pointing to itself was complaining. It was saying...the channel is in use.

I needed to stop the QMgr and restarted..then needed to delete the receiver channel. What could be causing the receiver channel in use?. any thoughts.

Thanks for your valuable info.
Back to top
View user's profile Send private message
fjb_saper
PostPosted: Mon Nov 17, 2008 1:38 pm    Post subject: Reply with quote

Grand High Poobah

Joined: 18 Nov 2003
Posts: 20756
Location: LI,NY

PeterPotkay wrote:
You need some time for each step to propogate thru the cluster.

Peter said it all
_________________
MQ & Broker admin
Back to top
View user's profile Send private message Send e-mail
PeterPotkay
PostPosted: Mon Nov 17, 2008 2:56 pm    Post subject: Reply with quote

Poobah

Joined: 15 May 2001
Posts: 7722

Sam Uppu wrote:

As we have many partial repository servers need to be rebooted for patching, is that ok if we delete and recreate the partial repository?.

No, its not OK. No offense, but deleting and recreating a QM just to reboot a box is one of the strangest ideas put forth on this site. Definitely one for a Top Ten list.

It makes zero sense and can only cause a whole bunch of problems.[/list]
_________________
Peter Potkay
Keep Calm and MQ On
Back to top
View user's profile Send private message
bruce2359
PostPosted: Mon Nov 17, 2008 4:20 pm    Post subject: Reply with quote

Poobah

Joined: 05 Jan 2008
Posts: 9469
Location: US: west coast, almost. Otherwise, enroute.

Quote:
is that ok if we delete and recreate the partial repository?.

I cannot imagine why you would want to delete and redefine your queue manager.

Why do you want/need to delete and recreate your queue manager?

What is your purpose? What end-result do you believe you will attain by doing so?

If this is someone else's idea, please ask them to provide an explanation.
_________________
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
View user's profile Send private message
Vitor
PostPosted: Mon Nov 17, 2008 4:27 pm    Post subject: Reply with quote

Grand High Poobah

Joined: 11 Nov 2005
Posts: 26093
Location: Texas, USA

Sam Uppu wrote:
As we have many partial repository servers need to be rebooted for patching, is that ok if we delete and recreate the partial repository?.


No, no, no, no no!

How many times, and by how many people, must you be told that if you're going to delete a clustered queue manager you must remove it from the cluster first? If you don't believe us, follow the other suggestion you've been given and read the cluster manual!!!

Doing what you suggest will break the cluster into little, tiny pieces. Called separate queue managers.

I repeat my earlier question - what is the deletion and recreation of the queue manager supposed to achieve? Someone has clearly decided to do this, but why? Do you delete and redefine any databases on the boxes as well?

It's a bizarre administrative decision. No matter how I weird I get (and forum regulars will know I can get really weird if I try) I can't make this sound good. Please give us some insight into the justification of this, so I can sleep again!
_________________
Honesty is the best policy.
Insanity is the best defence.
Back to top
View user's profile Send private message
Vitor
PostPosted: Mon Nov 17, 2008 4:31 pm    Post subject: Reply with quote

Grand High Poobah

Joined: 11 Nov 2005
Posts: 26093
Location: Texas, USA

Sam Uppu wrote:
One thing I wanted to share with you guys:
1. the full repository cluster sender/receiver channels were altered(changed the cluster attribute to Null).
2. Stopped the cluster sender/receiver channels.
3. Deleted the cluster sender/receiver channels.
Note: I was able to delete the cluster sender channel but while deleteing the cluster receiver channel pointing to itself was complaining. It was saying...the channel is in use.

I needed to stop the QMgr and restarted..then needed to delete the receiver channel. What could be causing the receiver channel in use?. any thoughts.


One thought - don't make up procedures yourself. Use the documented one. That way you won't get these problems.
_________________
Honesty is the best policy.
Insanity is the best defence.
Back to top
View user's profile Send private message
Sam Uppu
PostPosted: Mon Nov 17, 2008 7:32 pm    Post subject: Reply with quote

Yatiri

Joined: 11 Nov 2008
Posts: 610

Vitor wrote:
Sam Uppu wrote:
One thing I wanted to share with you guys:
1. the full repository cluster sender/receiver channels were altered(changed the cluster attribute to Null).
2. Stopped the cluster sender/receiver channels.
3. Deleted the cluster sender/receiver channels.
Note: I was able to delete the cluster sender channel but while deleteing the cluster receiver channel pointing to itself was complaining. It was saying...the channel is in use.

I needed to stop the QMgr and restarted..then needed to delete the receiver channel. What could be causing the receiver channel in use?. any thoughts.


One thought - don't make up procedures yourself. Use the documented one. That way you won't get these problems.


Somebody who was working in this position earlier to me made the decision that way. Not sure what caused him to opt this way. We need to change from deleting and recreating the objects as you guys suggested.

As the scripts are built that way and they are in PROD, I cant change them without a change request. Thatswhy I needed to run that way which will cause the problem.

Thatswhy I was asking you guys whether it is OK to delete and recreate the PRs.
Back to top
View user's profile Send private message
Vitor
PostPosted: Tue Nov 18, 2008 2:32 am    Post subject: Reply with quote

Grand High Poobah

Joined: 11 Nov 2005
Posts: 26093
Location: Texas, USA

Sam Uppu wrote:
Somebody who was working in this position earlier to me made the decision that way. Not sure what caused him to opt this way.


He forgot to take his meds one day?

Sam Uppu wrote:

Thatswhy I was asking you guys whether it is OK to delete and recreate the PRs.


It's not. It really, really isn't ok.
_________________
Honesty is the best policy.
Insanity is the best defence.
Back to top
View user's profile Send private message
Display posts from previous:   
Post new topic  Reply to topic Goto page 1, 2  Next Page 1 of 2

MQSeries.net Forum Index » Clustering » What happens to PRs if a FR is recreated?
Jump to:  



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
Protected by Anti-Spam ACP
 
 


Theme by Dustin Baccetti
Powered by phpBB © 2001, 2002 phpBB Group

Copyright © MQSeries.net. All rights reserved.