| Author |
Message
|
| rsk33 |
 Posted: Mon Jan 11, 2010 5:14 am Post subject: cluster issue |
 |
|
Centurion
Joined: 21 Aug 2006
Posts: 141
|
I have issue in production.
Front end QM is connected to two MW servers QM1(cluster A) and QM2 (cluster B) in a cluster.
The cluster queue with same name is created in QM1 and QM2 and are visible to Front end QM.
The QM2 (cluster B) is shut down forcebly due to hard disc failure.
Now the messages from Front end QM passed to QM1 for some time and failed with cluster resolution error.
My question is how can i redirect the messages to QM1 server without deleting the definitions of QM2 as i need to restore back the original setup.
reply is highly appreciated.
Regards |
|
| Back to top |
|
 |
| Vitor |
| Posted: Mon Jan 11, 2010 6:01 am Post subject: Re: cluster issue |
|
|
Grand High Poobah
Joined: 11 Nov 2005
Posts: 26093
Location: Texas, USA
|
| rsk33 wrote: |
| Front end QM is connected to two MW servers QM1(cluster A) and QM2 (cluster B) in a cluster. |
This implies they're in 2 separate clusters - A & B! 
| rsk33 wrote: |
| The cluster queue with same name is created in QM1 and QM2 and are visible to Front end QM. |
From this, I'll proceed on the assumption the above is a typo and both queue managers are in the same cluster.
| rsk33 wrote: |
| My question is how can i redirect the messages to QM1 server without deleting the definitions of QM2 as i need to restore back the original setup. |
The question is how do you plan to restore back to the original setup? If you plan to recreate QM2 from a system level backup (i.e. recreate the entire failed disc) then everything will sort itself out.
If you're building a new machine and defining a new QM2 from scratch you've got problems, as this won't be the QM2 that used to be in the cluster (queue managers are not identified by name). You'll need to forcibly remove the old QM2 using the documented procedure and re-add it.
As to the cluster resolution error, where is this appearing? If it's when the front end application is doing the put then the application is specifically addressing QM2 and there's not too much you can do except get the application fixed.
You could of course put a queue manager alias in so that QM2 resolves to QM1, but that does defeat the object of the cluster.
_________________
Honesty is the best policy.
Insanity is the best defence. |
|
| Back to top |
|
|
| rsk33 |
| Posted: Mon Jan 11, 2010 8:22 am Post subject: |
|
|
Centurion
Joined: 21 Aug 2006
Posts: 141
|
Dear vitor,
Thanks for the response.
The QM1 and QM2 are in two different clusters called A and B.
The front end QM is part of both clusters A and B.
Regards |
|
| Back to top |
|
|
| Vitor |
| Posted: Mon Jan 11, 2010 8:32 am Post subject: |
|
|
Grand High Poobah
Joined: 11 Nov 2005
Posts: 26093
Location: Texas, USA
|
| rsk33 wrote: |
| The QM1 and QM2 are in two different clusters called A and B. |
Ok I'll ask - why?
| rsk33 wrote: |
| The front end QM is part of both clusters A and B. |
How does the front end QM differentiate messages? It still sounds to be like messages are being deliberately addressed to QM2.
_________________
Honesty is the best policy.
Insanity is the best defence. |
|
| Back to top |
|
|
| bruce2359 |
| Posted: Mon Jan 11, 2010 8:39 am Post subject: |
|
|
Poobah
Joined: 05 Jan 2008
Posts: 9469
Location: US: west coast, almost. Otherwise, enroute.
|
| Quote: |
| The cluster queue with same name is created in QM1 and QM2 and are visible to Front end QM. |
I'm guessing that you mean that the front-end qmgr is a full-repos for both cluster A and B. Is that right?
How is the cluster queue defined - in both clusters A and B?
_________________
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 |
|
|
| rsk33 |
| Posted: Mon Jan 11, 2010 8:41 am Post subject: |
|
|
Centurion
Joined: 21 Aug 2006
Posts: 141
|
dear vitor,
This was designed for the cluster failover ie in case there is a problem one of the clusters.
The messages are going to two mechines as the queue is defined in both QM1 (cluster A)and QM2(cluster B).
Regards. |
|
| Back to top |
|
|
| rsk33 |
| Posted: Mon Jan 11, 2010 8:44 am Post subject: |
|
|
Centurion
Joined: 21 Aug 2006
Posts: 141
|
The cluster queue defined in QM1 as Cluster A and QM2 as Cluster B and
both are visible to front end.
Regards |
|
| Back to top |
|
|
| rsk33 |
| Posted: Mon Jan 11, 2010 8:45 am Post subject: |
|
|
Centurion
Joined: 21 Aug 2006
Posts: 141
|
| Front end QM is partial repository and QM1 and QM2 are full repositories. |
|
| Back to top |
|
|
| bruce2359 |
| Posted: Mon Jan 11, 2010 8:53 am Post subject: |
|
|
Poobah
Joined: 05 Jan 2008
Posts: 9469
Location: US: west coast, almost. Otherwise, enroute.
|
| rsk33 wrote: |
The cluster queue defined in QM1 as Cluster A and QM2 as Cluster B and
both are visible to front end.
Regards |
You already said this.
How exactly is the cluster queue defined in both clusters A and B?
_________________
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 |
|
|
| Vitor |
| Posted: Mon Jan 11, 2010 8:55 am Post subject: |
|
|
Grand High Poobah
Joined: 11 Nov 2005
Posts: 26093
Location: Texas, USA
|
| rsk33 wrote: |
| This was designed for the cluster failover ie in case there is a problem one of the clusters. |
But that doesn't make sense - if you were using WMQ as this "poor man's failover" both machines would need to be in the same cluster.
And you can't have a problem with a cluster - there's no such object. A cluster is an addressing method. You can only have problems with a queue manager that happens to be in a cluster.
| rsk33 wrote: |
| The messages are going to two mechines as the queue is defined in both QM1 (cluster A)and QM2(cluster B). |
So you're using a namelist to span the clusters? Or what?
_________________
Honesty is the best policy.
Insanity is the best defence.
Last edited by Vitor on Mon Jan 11, 2010 8:58 am; edited 1 time in total |
|
| Back to top |
|
|
| Vitor |
| Posted: Mon Jan 11, 2010 8:57 am Post subject: |
|
|
Grand High Poobah
Joined: 11 Nov 2005
Posts: 26093
Location: Texas, USA
|
| rsk33 wrote: |
| Front end QM is partial repository and QM1 and QM2 are full repositories. |
As we've established these queue managers are in 2 distinct clusters where is the other full repository for clusters A & B? If you've incorrectly set up the clusters with only 1 full repository then I'm getting an idea why you're getting resolution errors......
_________________
Honesty is the best policy.
Insanity is the best defence. |
|
| Back to top |
|
|
| bruce2359 |
| Posted: Mon Jan 11, 2010 8:58 am Post subject: |
|
|
Poobah
Joined: 05 Jan 2008
Posts: 9469
Location: US: west coast, almost. Otherwise, enroute.
|
(Is this one of those certification test thingies?)
_________________
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 |
|
|
| rsk33 |
| Posted: Mon Jan 11, 2010 9:01 am Post subject: |
|
|
Centurion
Joined: 21 Aug 2006
Posts: 141
|
on QM1
DEFINE QLOCAL ('a') DEFBIND(NOTFIXED) boqname('Backout') CLUSTER('A') SHARE REPLACE;
on qm2
DEFINE QLOCAL ('a') DEFBIND(NOTFIXED) boqname('Backout') CLUSTER('B') SHARE REPLACE; |
|
| Back to top |
|
|
| Vitor |
| Posted: Mon Jan 11, 2010 9:05 am Post subject: |
|
|
Grand High Poobah
Joined: 11 Nov 2005
Posts: 26093
Location: Texas, USA
|
| bruce2359 wrote: |
| (Is this one of those certification test thingies?) |
It's starting to sound like one, except the cluster configuration is so non-standard 
_________________
Honesty is the best policy.
Insanity is the best defence. |
|
| Back to top |
|
|
| bruce2359 |
| Posted: Mon Jan 11, 2010 9:08 am Post subject: |
|
|
Poobah
Joined: 05 Jan 2008
Posts: 9469
Location: US: west coast, almost. Otherwise, enroute.
|
| Quote: |
| Front end QM is partial repository and QM1 and QM2 are full repositories. |
Front end is a partial repos in which cluster?? A or B?
_________________
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 |
|
|
|
|
