| |
|
RSS Feed - WebSphere MQ Support
|
RSS Feed - Message Broker Support
|
  |
|
| MSCS Service Groups |
« View previous topic :: View next topic » |
| Author |
Message
|
| exerk |
 Posted: Wed Oct 12, 2011 12:24 am Post subject: MSCS Service Groups |
 |
|
Jedi Council
Joined: 02 Nov 2006
Posts: 6339
|
In the past few days I've been having a 'philosophical' discussion with a bunch of people in regard to Service Groups and WMQ. My view is that if there are multiple queue managers with an HA requirement, each queue manager should sit within its own Service Group, and with its own VIP address, and not use that of the cluster. Other views are that a single Service Group is good enough, and using the cluster VIP address is OK.
Some people have said yea, others nay, the arguments ranging from ease of management and less complexity (some networks, storage and OS admins) to not having eggs in one basket, and it makes it less of a chore if a particular queue manager needs decommissioning or rebuilding (me and the other half of the networks, storage and OS admins).
I know there is never a one-size-fits-all solution to any implementation, and to forestall the inevitable "...what are the requirements...", and the argument of "...but you wouldn't put everything in one Resource Group if it was UNIX...", I'm just curious to find out whether there is a consensus for one side of the argument or the other, and whether my personal preference is clouding my judgement and objectivity.
The pebble is now in the pond, and I shall observe what ripples it makes 
Also, if this generates enough interest I'll turn it into a poll.
_________________
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 |
|
 |
| mqjeff |
| Posted: Wed Oct 12, 2011 3:02 am Post subject: |
|
|
Grand Master
Joined: 25 Jun 2008
Posts: 17447
|
I'm not as up on MSCS as I am on HA on other platforms.
Does using a single service group require that all qmgrs fail over or not? |
|
| Back to top |
|
|
| exerk |
| Posted: Wed Oct 12, 2011 3:21 am Post subject: |
|
|
Jedi Council
Joined: 02 Nov 2006
Posts: 6339
|
| mqjeff wrote: |
I'm not as up on MSCS as I am on HA on other platforms.
Does using a single service group require that all qmgrs fail over or not? |
If the Resource Group (RG) needs to fail over then all queue managers will go with it, with the subsequent temporary total loss of service. What the good book is not very clear about is whether if the cluster manager decides there is a problem with one queue manager that it will stop/start that queue manager, or the whole RG, or fail over the RG.
The manual states, , "...You choose which queue managers are placed under MSCS control by using the MSCS Cluster Administrator to create a resource instance for each such queue manager. This process presents you with a list of resources from which to select the queue manager that you want that instance to manage...", which implies that each queue manager can be independently stop/started within the RG.
_________________
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 |
|
|
| mqjeff |
| Posted: Wed Oct 12, 2011 3:23 am Post subject: |
|
|
Grand Master
Joined: 25 Jun 2008
Posts: 17447
|
| If you kill one qmgr, do the others fail over? |
|
| Back to top |
|
|
| exerk |
| Posted: Wed Oct 12, 2011 3:37 am Post subject: |
|
|
Jedi Council
Joined: 02 Nov 2006
Posts: 6339
|
| mqjeff wrote: |
| If you kill one qmgr, do the others fail over? |
Don't know. I've only ever had either one queue manager in the only Rervice Group on a cluster, or multiple RGs where I've had multiple queue managers (read two) where one was added at a later time and it seemed the sensible way to configure it without impacting the other one, e.g. fail-over testing etc.
_________________
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 |
|
|
| Vitor |
| Posted: Wed Oct 12, 2011 5:33 am Post subject: |
|
|
Grand High Poobah
Joined: 11 Nov 2005
Posts: 26093
Location: Texas, USA
|
I've always seen one queue manager per service group. Typically rationalisations include increased control, ease of WMQ maintenance and clarity on PowerPoint slides.
Interestingly (at least to me), I've just realised I've never actually set up WMQ under MCSC. Used queue managers under MCSC obviously, but never actually been involved in the set up. After all these years, you'd think I'd have done one. HACMP, Veritas, SecureGuard, been there, done that but not MCSC.
Weird. 
_________________
Honesty is the best policy.
Insanity is the best defence. |
|
| Back to top |
|
|
| PeterPotkay |
| Posted: Wed Oct 12, 2011 11:32 am Post subject: |
|
|
Poobah
Joined: 15 May 2001
Posts: 7722
|
| mqjeff wrote: |
| If you kill one qmgr, do the others fail over? |
I suspect you are asking a leading question.
Yes, they do! Why anyone would want to put multiple QMs into one cluster resource group is beyond me. Why have 2 QMs in one group if a problem with one can bring down the other? Just put all the queues on one QM! Or if you need 2 QMs, then you need two separate groups and the isolation and flexibility they provide.
As for using the overall cluster VIP for your QM, I'd advise against it. The overall health, stability and control of the cluster should be separate where possible from the individual cluster group(s). Its trivial to ask for one more IP address when setting the cluster and the cluster group up.
| Quote: |
| The manual states, , "...You choose which queue managers are placed under MSCS control by using the MSCS Cluster Administrator to create a resource instance for each such queue manager. This process presents you with a list of resources from which to select the queue manager that you want that instance to manage...", which implies that each queue manager can be independently stop/started within the RG. |
That is correct, you can offline and online them independently within the same group, just like the other resources that may be in the group, unless you added dependencies between them. If QM1 in GroupA is dependent on QM2 in Group A, and you offline QM2, down goes QM1. But the overall group would not fail over in this case. If QM1 or QM2 are set up as critical resources in the group, and either has an unplanned fail, the whole group swings over, bringing the other QM along for the ride.
_________________
Peter Potkay
Keep Calm and MQ On |
|
| Back to top |
|
|
| exerk |
| Posted: Thu Oct 13, 2011 12:27 am Post subject: |
|
|
Jedi Council
Joined: 02 Nov 2006
Posts: 6339
|
Peter, thank you! I'm going to print what you said, frame it, and place it in a very prominent place on the wall. 
_________________
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 |
|
|
|
|
|
| |
|
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
|
|
|
|
