| Author |
Message
|
| visionR32 |
 Posted: Tue Feb 03, 2009 5:31 am Post subject: migration 2 qmgrs 5.3 to 6.0.2.x - VCS with SAN H/W HA |
 |
|
Novice
Joined: 29 Jan 2009
Posts: 11
|
Hi guys,
I have 2 qmgrs on 2 separate boxes (QM1 "active qmgr (running)" on BOX_A and QM1 "stand-by (not running)" on BOX_B) but they pointing to the 'same' LOGS and QMGRS directorys which resides on SAN storage. the qmgrs are in a Veritas Cluster Server HA.
I'm planning to upgrade the current 5.3 code to 6.0.2 -
Question: won't i corrupt the qmgr LOGS if i upgrade "stand-by" QM2 to v6.0.2 then switch the SAN LOGS (from v5.3 QM1) to the upgraded QM2
what is the safe step to go about upgrading the mq code? - does the upgrade update any file in the QMGRS directory?
kind regards |
|
| Back to top |
|
 |
| exerk |
| Posted: Tue Feb 03, 2009 5:56 am Post subject: |
|
|
Jedi Council
Joined: 02 Nov 2006
Posts: 6339
|
This sounds like a half-way house HA queue manager solution 
Were the queue managers(?) created using the appropriate SupportPac? Am I perhaps just misinterpreting your post? More information please...
_________________
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 |
|
|
| visionR32 |
| Posted: Tue Feb 03, 2009 6:15 am Post subject: |
|
|
Novice
Joined: 29 Jan 2009
Posts: 11
|
Ok lets see if i can be clear... 
The 2 qmgrs are part of a veritas clustering solution - where if QM1 fails (BOX_A goes down or something) - then a switch will be made to BOX_B and QM2 will be started - and point to the SAME location were failed QM1 was pointing on SAN -
| Quote: |
| This sounds like a half-way house HA queue manager solution |
Y u say that?
my question is that - when i do an upgrade on QM2 whilst QM1 is still running won't the upgrade try to access the LOGS / OR QMGRS directory which will still be used by QM1 at that time? - (which might corrupt it?) |
|
| Back to top |
|
|
| mqjeff |
| Posted: Tue Feb 03, 2009 6:17 am Post subject: |
|
|
Grand Master
Joined: 25 Jun 2008
Posts: 17447
|
You need to take appropriate steps to ensure that while the two nodes are at different versions, that you do not fail over to the older code.
As you say, this will cause the qmgr to become corrupt.
One possible way to do this is to stop the qmgr entirely on both nodes, and upgrade them.
I'm sure you're clever enough to work out other solutions. |
|
| Back to top |
|
|
| exerk |
| Posted: Tue Feb 03, 2009 6:24 am Post subject: |
|
|
Jedi Council
Joined: 02 Nov 2006
Posts: 6339
|
| Quote: |
| This sounds like a half-way house HA queue manager solution |
I said this because your description of the set-up implied that each queue manager is a separate entity, that had been 'manually' finagled to point to the same set of logs - I've seen it done before.
If it is a 'genuine' HA queue manager solution, ensure that the stand-by node is de-selected as a fail-over node, upgrade that node, then reinstate it as a fail-over node after you have upgraded the 'active' node..
_________________
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 |
|
|
|
|
