| |
|
RSS Feed - WebSphere MQ Support
|
RSS Feed - Message Broker Support
|
| MQ High Availability |
« View previous topic :: View next topic » |
| Author |
Message
|
| jcv |
 Posted: Mon Dec 15, 2008 8:52 am Post subject: |
 |
|
Chevalier
Joined: 07 May 2007
Posts: 411
Location: Zagreb
|
| Mark Hiscock & Simon Gormley wrote: |
In the standby machine solution, a second machine is used to host a second queue manager that is activated when the original machine or queue manager fails. |
| manicminer wrote: |
Assuming that your reason for failure isn't the 'shared disk' failing  (Yes I know this should be a SAN with RAID etc. etc. but it is still a valid point) |
| KramJ wrote: |
| Quote: |
| What if Shared data /logs are corrupted and Qmgr failed on one node.Guess we should not able to start it on another node till data /logs are repaired. |
-Yes, that is correct because you'll be accessing shared data, just from a different node. |
Is there a chance for a queue manager to fail if original machine is not failed, as well as shared disk, and Shared data /logs are not corrupted. If qmgr is dedicated to a certain local application and no other local or remote application connects to it, is there a point to configure HA clustering only for that qmgr, and not for that application too, as a separate unit of failover. |
|
| Back to top |
|
 |
| exerk |
| Posted: Mon Dec 15, 2008 9:23 am Post subject: |
|
|
Jedi Council
Joined: 02 Nov 2006
Posts: 6339
|
| jcv wrote: |
| Is there a chance for a queue manager to fail if original machine is not failed, as well as shared disk, and Shared data /logs are not corrupted... |
Yes, depending on how the fail-over for the queue manager is configured; usual practice seems to be attempt restart of the queue manager on the same node, then fail across to the other node if the same-node restart fails. The queue manager may fail, or be perceived as failing by the HA software agent, but the node is quite 'healthy'.
| jcv wrote: |
| If qmgr is dedicated to a certain local application and no other local or remote application connects to it, is there a point to configure HA clustering only for that qmgr, and not for that application too, as a separate unit of failover. |
In my opinion, no. If the application is the only one using the queue manager, and the application is not highly-available, there is no point in making the queue manager HA - unless the application is database-backed and a separate instance can run on the stand-by server and only needs access to the messages. Whether it's a good idea to put the application within a separate resource group is a bit more difficult to answer - it depends on the 'solidity' of the application - but certainly a dependency can be placed on the queue manager resource group, i.e. if that fails over, so does the application, or vice-versa. Horses for courses, as the saying goes - any solution is dependent on the requirements.
_________________
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 |
|
|
| jcv |
| Posted: Mon Dec 15, 2008 1:45 pm Post subject: |
|
|
Chevalier
Joined: 07 May 2007
Posts: 411
Location: Zagreb
|
Thanks exerk, I hope I understand all your points.
| exerk wrote: |
| Yes, depending on how the fail-over for the queue manager is configured; usual practice seems to be attempt restart of the queue manager on the same node, then fail across to the other node if the same-node restart fails. The queue manager may fail, or be perceived as failing by the HA software agent, but the node is quite 'healthy'. |
I suppose its purpose is for circumstances when the system is temporarily low on resources but not to that extent that the operating system is threatened? Being complete beginner about HA things, I have missed that fact about the usual practice. And there is the old habit, to start to discuss things, before studying them first.
| exerk wrote: |
| In my opinion, no. If the application is the only one using the queue manager, and the application is not highly-available, there is no point in making the queue manager HA - unless the application is database-backed and a separate instance can run on the stand-by server and only needs access to the messages. |
Yes, that's what I meant, there is no redundant, separate, remote instance of the database-backed application, at the moment. |
|
| Back to top |
|
|
| exerk |
| Posted: Mon Dec 15, 2008 2:42 pm Post subject: |
|
|
Jedi Council
Joined: 02 Nov 2006
Posts: 6339
|
| jcv wrote: |
| I suppose its purpose is for circumstances when the system is temporarily low on resources but not to that extent that the operating system is threatened? |
Not quite...the server should be powerful enough to never be affected at the OS level as the server may have a number of resource groups attached, the queue manager being just one of them. Occasionally, one of these groups may be perceived by the monitoring agent to be failing, and the resource group may be restarted automatically by the HA software, and that will have an impact on any other dependent resources. It is generally quicker to attempt a restart on the 'current' node than to mount the disk on the other node and restart there, and if there are a number of dependent resource groups, moving one resource group may require the moving of them all.
_________________
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 |
|
|
|
|
|
|
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
|
|
|
|
