| Author |
Message
|
| memphis |
 Posted: Tue Jan 19, 2010 2:03 am Post subject: |
 |
|
Novice
Joined: 05 Jan 2010
Posts: 21
|
That's true. I was using the same samples you mentioned earlier in the current case.
I'm currently doing some research on the possibility of the OS to kill mqm processes before powering-off totally. I'm not sure if this works.
Can anybody suggest me if there is a possibility of a VM machine (or) OS which does not end abruptly when POWERED-OFF but issues a shutdown command or something like that? Did I make myself clear with the question? Pl ask me if this is not clear.
Appreciate your suggestions...
BTW - Does anybody have sample JAVA PROGRAMS (instead of C programs as given by IBM) for amqsghac, amqsmhac and amqsphac programs?
Thank you all in advance... |
|
| Back to top |
|
 |
| mvic |
| Posted: Tue Jan 19, 2010 3:59 am Post subject: |
|
|
Jedi
Joined: 09 Mar 2004
Posts: 2080
|
The standby/multi-instance queue manager mechanism relies on the OS / filesystem successfully releasing locks from the failed system and allowing them to be obtained by the standby system.
Leaving the client out of it for the moment, have you checked on the standby system that the standby instance of the queue manager successfully recovered those locks and took over, thus becoming the active instance?
The reconnectable aspects of the client will not work if the standby instance of the queue manager has not taken over as the active instance. |
|
| Back to top |
|
|
| exerk |
| Posted: Tue Jan 19, 2010 4:05 am Post subject: |
|
|
Jedi Council
Joined: 02 Nov 2006
Posts: 6339
|
In my case the test of the file system (using amqmfsck) was successful, but as stated I did not (and still do not) have the time to follow up whether it was a VM issue that will/will not translate across to discrete hardware.
_________________
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 |
|
|
| memphis |
| Posted: Tue Jan 19, 2010 4:07 am Post subject: |
|
|
Novice
Joined: 05 Jan 2010
Posts: 21
|
Thank you very much for the clear detailed reply. I appreciate that.
The only check I did on the standby machine (if it had acquaired locks) was to check the status of the QMGR (by doing dspmq -x -o standby -o status).
The above command just shows that the QMGR in the standby machine is in standby state (it is not neither in STARTING state (nor) in any other state).
Is there any other way to determine if the file system in the standby machine has acquaired locks from active machine? Is there anything which we can do to initiate that process (if the QMGR stays indefinitely in the standby state by not acquairing the locks?)
Your reply is highly commendable as this will resolve many issues Ive been facing with it.. Thank you Sir...
NOTE:- Just to add to my above note, I also issued amqmfsck command to verify my file system and it ran succesfully.. |
|
| Back to top |
|
|
| exerk |
| Posted: Tue Jan 19, 2010 4:09 am Post subject: |
|
|
Jedi Council
Joined: 02 Nov 2006
Posts: 6339
|
| memphis wrote: |
Thank you very much for the clear detailed reply. I appreciate that.
The only check I did on the standby machine (if it had acquaired locks) was to check the status of the QMGR (by doing dspmq -x -o standby -o status).
The above command just shows that the QMGR in the standby machine is in standby state (it is not neither in STARTING state (nor) in any other state).
Is there any other way to determine if the file system in the standby machine has acquaired locks from active machine? Is there anything which we can do to initiate that process (if the QMGR stays indefinitely in the standby state by not acquairing the locks?)
Your reply is highly commendable as this will resolve many issues Ive been facing with it.. Thank you Sir... |
Did you run amqmfsck prior to creating your queue managers to ensure that the file system was Multi-Instance capable, as stated in the manual?
_________________
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 |
|
|
| memphis |
| Posted: Tue Jan 19, 2010 4:13 am Post subject: |
|
|
Novice
Joined: 05 Jan 2010
Posts: 21
|
Yes I did issued amqmfsck command prior to the creation of the queue managers and my filesystems in both the servers does support multi instance queue manager configuration.
QUICK NOTE to you:-
I succesfully tested HA with the following:-
1. endmqm -s QMGR
2. Killing the mqm processes to be able to failover to the standby machine
3. Shutdown of the QMGR to be able to failover to the standby machine
But the only issue is when I POWER-OFF the VM, the clients are not able to reconnect to the standby machine (the standby queue manager just remains in the standby state forever).
Thank you and please let me know if you require any additional information from my side and I'll be glad to provide you.. |
|
| Back to top |
|
|
| mvic |
| Posted: Tue Jan 19, 2010 4:31 am Post subject: |
|
|
Jedi
Joined: 09 Mar 2004
Posts: 2080
|
| memphis wrote: |
| The above command just shows that the QMGR in the standby machine is in standby state (it is not neither in STARTING state (nor) in any other state). |
It looks like the standby queue manager was not given the file locks by the NFS server. I am not much of an expert on NFS. But out of interest, what are the mount options in use, the NFS server/client in use (are they fully patched) and what are the OS versions in use?
| Quote: |
| Is there any other way to determine if the file system in the standby machine has acquaired locks from active machine? |
If the standby queue manager gets the locks, it will take over as the active queue manager - I assume that would be the most reliable outward sign of whether the transfer of locks had been successful.
| Quote: |
| Is there anything which we can do to initiate that process |
I do not know. |
|
| Back to top |
|
|
| memphis |
| Posted: Tue Jan 19, 2010 5:13 am Post subject: |
|
|
Novice
Joined: 05 Jan 2010
Posts: 21
|
I use the following command to mount:
mount IP:/SHAREDRIVE /SHAREDRIVE -t nfs -o intr,hard
OS i'm using is openSuse 11.2
I believe that it's totally to do with the locks.
I appreciate your honest answer and I'm sure that I'll find the best answer to the solution within 24 hours (am confident of it) and will share the solution with you guys as I resolve the issue.
Thank you all for patiently answering the questions and I'm really tending to like this forum a lot... You all are awesome.. Have a great day ahead guys!!!! |
|
| Back to top |
|
|
| mqjeff |
| Posted: Tue Jan 19, 2010 5:25 am Post subject: |
|
|
Grand Master
Joined: 25 Jun 2008
Posts: 17447
|
There is a hard requirement on the version of NFS in use.
If you didn't meet that, it is the source of your problem. |
|
| Back to top |
|
|
| memphis |
| Posted: Tue Jan 19, 2010 5:29 am Post subject: |
|
|
Novice
Joined: 05 Jan 2010
Posts: 21
|
Yeah jeff.. That's right.. I did used hard for the options while mounting..
The options which I used for the mount command are, HARD, INTERRUPTIBLE mount.
mount IP:/DRIVE /DRIVE -t nfs -o hard,intr (you could see hard interruptible option for the mount command here).
I also tried the scenario with nfsvers=2,3,4 as well but with no use. (Just to cover all options I have at my disposal) |
|
| Back to top |
|
|
| bruce2359 |
| Posted: Tue Jan 19, 2010 7:20 am Post subject: |
|
|
Poobah
Joined: 05 Jan 2008
Posts: 9472
Location: US: west coast, almost. Otherwise, enroute.
|
| Quote: |
| But the only issue is when I POWER-OFF the VM, the clients are not able to reconnect to the standby machine |
Of coures. A standby instance is just that - a standby.
Before you powered off the vm, did you do a dspmq -x -o to ensure that there was a standby qmgr? Was there?
[edit]
And after you powered off the vm, did the dspmq indicate that there was a RUNNING qmgr (the one that was standby)?
_________________
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 |
|
|
| mvic |
| Posted: Tue Jan 19, 2010 8:12 am Post subject: |
|
|
Jedi
Joined: 09 Mar 2004
Posts: 2080
|
|
| Back to top |
|
|
| mqjeff |
| Posted: Tue Jan 19, 2010 10:00 am Post subject: |
|
|
Grand Master
Joined: 25 Jun 2008
Posts: 17447
|
| memphis wrote: |
| Yeah jeff.. That's right.. I did used hard for the options while mounting.. |
That is not at all what I meant.
I meant that the requirement on a specific version of NFS is a hard requirement and not something that can be taken lightly. |
|
| Back to top |
|
|
| bruce2359 |
| Posted: Tue Jan 19, 2010 10:12 am Post subject: |
|
|
Poobah
Joined: 05 Jan 2008
Posts: 9472
Location: US: west coast, almost. Otherwise, enroute.
|
NFS V4 is a requirement for multi-instance WMQ.
_________________
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 |
|
|
| memphis |
| Posted: Tue Jan 19, 2010 9:19 pm Post subject: |
|
|
Novice
Joined: 05 Jan 2010
Posts: 21
|
Just to give heads-up to all, I followed the exact instructions as given at Infocenter (Creating Multi Instance QMGRS in Linux).
I'm really quite not sure if NFS V4 is really a requirement. It doesn't state it anywhere as a hard requirement in the manual.
But still, I am going to try my case with NFS V4 in a little while and then get back to you with the results.
Thanks to you all for sharing your thoughts.. Appreciate that... |
|
| Back to top |
|
|
|
|
