| Author |
Message
|
| bkRaju |
 Posted: Mon Mar 09, 2009 9:40 pm Post subject: |
 |
|
Centurion
Joined: 19 Aug 2008
Posts: 106
|
Hi...
Can I define my own repository process?
If so how to define and start it? |
|
| Back to top |
|
 |
| Vitor |
| Posted: Tue Mar 10, 2009 12:43 am Post subject: |
|
|
Grand High Poobah
Joined: 11 Nov 2005
Posts: 26093
Location: Texas, USA
|
| bkRaju wrote: |
| Can I define my own repository process? |
In what sense? You want to run a cluster outside of the queue manager's control? 
Your question as asked makes no sense.
| bkRaju wrote: |
| If so how to define and start it? |
It's theoretically possible to manually restart a queue manager's repository process. Given the number of parameters, the undocumentated nature of these parameters and the dire consequences of getting them wrong it's much better to recycle the queue manager.
_________________
Honesty is the best policy.
Insanity is the best defence. |
|
| Back to top |
|
|
| bkRaju |
| Posted: Tue Mar 10, 2009 2:29 am Post subject: |
|
|
Centurion
Joined: 19 Aug 2008
Posts: 106
|
what I mean is creating repository process and starting in QMGR like listener.
After starting the repository process,It stays for some time and terminates.Again I need to start the repositor process.
I am not sure why it happens so frequently..
To resolve the issue I had to recycle the qmgr.But It impacts other applications...
So could you give optimal solution... |
|
| Back to top |
|
|
| zhanghz |
| Posted: Tue Mar 10, 2009 2:35 am Post subject: |
|
|
Disciple
Joined: 17 Jun 2008
Posts: 186
|
| what do you mean by "It stays for some time and terminates" and how do you "start the repositor process"? |
|
| Back to top |
|
|
| Vitor |
| Posted: Tue Mar 10, 2009 2:46 am Post subject: |
|
|
Grand High Poobah
Joined: 11 Nov 2005
Posts: 26093
Location: Texas, USA
|
| bkRaju wrote: |
| what I mean is creating repository process and starting in QMGR like listener. |
No. It's an integral queue manager process (like the logging) not an external object (like a listener).
What do the queue manager logs say about why it terminates? What version of WMQ is this? On what platform?
_________________
Honesty is the best policy.
Insanity is the best defence. |
|
| Back to top |
|
|
| exerk |
| Posted: Tue Mar 10, 2009 3:53 am Post subject: |
|
|
Jedi Council
Joined: 02 Nov 2006
Posts: 6339
|
Not in a cluster is it? 
_________________
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 |
|
|
| bkRaju |
| Posted: Tue Mar 10, 2009 5:03 am Post subject: |
|
|
Centurion
Joined: 19 Aug 2008
Posts: 106
|
The queue manager which I am talking abt is a full repository for all qmgrs in my envronment.
This problem is frequently happening twice in a week or once.
I didnt see any logs when it got terminates.
I have to start back manually.It stays one or 2 minutes. if I issue
dis qs(system.command.queue) type(handle) opentype(input) den I am able to see the repos process.
but I cant see the same process id out side the qmgr.I mean ps -ef|grep proceess id.
If I grep for process id it shows like this
ps -ef|grep 360954
mqm 254830 466310 0 23:34:49 pts/10 0:00 grep 360954
mqm 360954 468064 13 0:00 <defunct>
It mqv5.3.
Even It is Mqv5.3 I have not seen this issue before...
So I am looking for root cause and resolution for same. |
|
| Back to top |
|
|
| Vitor |
| Posted: Tue Mar 10, 2009 5:05 am Post subject: |
|
|
Grand High Poobah
Joined: 11 Nov 2005
Posts: 26093
Location: Texas, USA
|
WMQv5.3.???
| bkRaju wrote: |
| So I am looking for root cause and resolution for same. |
If it's an old version of v5.3 that may be the root cause; early cluster code did wobble a bit.
_________________
Honesty is the best policy.
Insanity is the best defence. |
|
| Back to top |
|
|
| exerk |
| Posted: Tue Mar 10, 2009 5:11 am Post subject: |
|
|
Jedi Council
Joined: 02 Nov 2006
Posts: 6339
|
Check in the Full Repository to see if there is an explicit CLUSSDR defined to one of the Partial Repositories - it's a shot in the dark, but who knows?
_________________
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 |
|
|
| vol |
| Posted: Tue Mar 10, 2009 12:59 pm Post subject: |
|
|
Acolyte
Joined: 01 Feb 2009
Posts: 69
|
When the repos mgr process amqrrmfa terminates it will certainly write at least one error msg to the qmgr error logs in /var/mqm/qmgrs/QMGR/errors, and it may write an FDC file to /var/mqm/errors. Additionally, there will be an FDC with probe ID ZX005025 written by the EC (amqzxma0).
So, when you said you did not see any logs when it terminates, you clearly did not look very hard, or perhaps at all. |
|
| Back to top |
|
|
| JosephGramig |
| Posted: Tue Mar 10, 2009 1:05 pm Post subject: |
|
|
Grand Master
Joined: 09 Feb 2006
Posts: 1244
Location: Gold Coast of Florida, USA
|
Please upgrade to V6.0.2.5 or V7.0 (latest fix).
At least put on the final CSD14. If any QMGR is in the cluster above your FR level, that is a potential source of trouble.
How many Full Repositories do you have? Do they all have explicit cluster senders pointing to one another (for any one cluster)? For instance 3 FRs in a cluster would mean each would have two explicit cluster senders to the other two FRs in the cluster.
You might ask why? Well, if you read between the lines in the Info Center, it tells you that any FR that gets a cluster update will only forward that update to other FRs to which it has explicit cluster senders defined. If you didn't do this, your Repository is corrupt already. |
|
| Back to top |
|
|
|
|
