| |
|
RSS Feed - WebSphere MQ Support
|
RSS Feed - Message Broker Support
|
  |
|
| WMQ High Availibility on AIX |
« View previous topic :: View next topic » |
| Author |
Message
|
| jeasterl |
 Posted: Fri Nov 08, 2002 4:15 am Post subject: WMQ High Availibility on AIX |
 |
|
Acolyte
Joined: 24 Jun 2001
Posts: 65
|
Hello All,
I am in need of some best practice information regarding WMQ high availibility on the AIX platform. Any help would be appriciated.
Thanks! |
|
| Back to top |
|
 |
| jsware |
| Posted: Fri Nov 08, 2002 2:29 pm Post subject: |
|
|
Chevalier
Joined: 17 May 2001
Posts: 455
|
|
| Back to top |
|
|
| jeasterl |
| Posted: Sat Nov 09, 2002 8:43 am Post subject: |
|
|
Acolyte
Joined: 24 Jun 2001
Posts: 65
|
Thank you for the response, however the customer is not using HACMP. In addition, I failed to mention that this is a pub/sub configuration. My initial thought was to cluster the queue managers such that, if one fails, all messages would fail over to the other queue manager. However, the pub/sub documentation clearly states that queues used in the the pub/sub environment cannot be clustered.
I understand that, in order to fully achieve HA, several diretories need to be in a shared directory, i.e /var/mqm/ . Other than that, if the customer is not using HACMP, are there other consideriations that I should take to make this environment highly available? |
|
| Back to top |
|
|
| jsware |
| Posted: Thu Nov 14, 2002 12:54 am Post subject: |
|
|
Chevalier
Joined: 17 May 2001
Posts: 455
|
Hmm, without using HACMP, how are you going to fail over the IP address so that if there is a failure, the second box can be configured as the primary box (you can do this manually, and write your own version of HACMP).
Basically you would need the following three scripts -
checkmq
startmq
stopmq
You can then have an "event management" script run checkmq every minute or so, this can check that all the processes are running and that the qm responds to a "ping qmgr".
If this fails, then the management script can run stopmq, then unmount and mount that /var/mqm filesystem onto the other machine, reconfigure the IP address etc. and bring it up. This is what HACMP does basically.
Without the /var/mqm filesystem being on shared disk (not shared via NFS because MQ logging does not support this) I don;t know how you're going to fail the filesystem over to a backup box.
You'll have to consider where the pub/sub messages are stored (in a database?) and fail those filesystems over too.
You can also use the start/stop/check scripts to attempt an automatic re-start - if the check fails issue a stop followed by a start... |
|
| Back to top |
|
|
|
|
|
| |
|
Page 1 of 1 |
|
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
|
|
|
|
