| Author |
Message
|
| rcp_mq |
 Posted: Tue Mar 27, 2012 5:56 am Post subject: Do multi-instance queue manager start automatically |
 |
|
Centurion
Joined: 13 Dec 2011
Posts: 133
|
|
| Back to top |
|
 |
| Vitor |
| Posted: Tue Mar 27, 2012 6:05 am Post subject: Re: Do multi-instance queue manager start automatically |
|
|
Grand High Poobah
Joined: 11 Nov 2005
Posts: 26093
Location: Texas, USA
|
Where? It says automatically in my version:
| Quote: |
| If the active instance fails, a standby instance running on a different computer automatically takes over |
_________________
Honesty is the best policy.
Insanity is the best defence. |
|
| Back to top |
|
|
| mqjeff |
| Posted: Tue Mar 27, 2012 6:23 am Post subject: |
|
|
Grand Master
Joined: 25 Jun 2008
Posts: 17447
|
So we have machine A, and Machine B.
We have, normally, the primary instance running on Machine A, and the secondary instance on machine B.
We must issue strmqm on both machines to get both instances started.
Then the primary instance fails, and Machine B takes over.
At this point, we do not have a running instance on Machine A. This must be started by running strmqm, if we wish for Machine A to now be a secondary instance for Machine B. Presumably we only do this after we have ensured that the cause of the failure on Machine A is resolved. |
|
| Back to top |
|
|
| rcp_mq |
| Posted: Tue Mar 27, 2012 6:40 am Post subject: |
|
|
Centurion
Joined: 13 Dec 2011
Posts: 133
|
@mqjeff
So you're saying only a single failover is possible?
@vitor
Differences between multi-instance queue managers and HA clusters |
|
| Back to top |
|
|
| Vitor |
| Posted: Tue Mar 27, 2012 6:49 am Post subject: |
|
|
Grand High Poobah
Joined: 11 Nov 2005
Posts: 26093
Location: Texas, USA
|
| rcp_mq wrote: |
| So you're saying only a single failover is possible? |
By "single failover" do you mean that Machine A fails over to Machine B but a further failure doesn't have Machine B failing over to Machine C or Machine A fails over to Machine B or Machine C based on some criteria?
| rcp_mq wrote: |
| Differences between multi-instance queue managers and HA clusters |
A multi-instance queue manager provides automatic failover of the queue manager's services in the event of a failure. An HA cluster (HACMP, MSCS, etc) provides automatic failover of 1 or more services (one of which could be a queue manager) in the event of failure.
_________________
Honesty is the best policy.
Insanity is the best defence. |
|
| Back to top |
|
|
| mqjeff |
| Posted: Tue Mar 27, 2012 6:50 am Post subject: |
|
|
Grand Master
Joined: 25 Jun 2008
Posts: 17447
|
| rcp_mq wrote: |
@mqjeff
So you're saying only a single failover is possible? |
That's not remotely what I said.
I said the same thing that the InfoCenter said.
That when failover occurs, the previous primary instance is NOT converted to a secondary instance.
Remember, the primary instance has FAILED.
IT'S DEAD. IT IS AN EX-QUEUE MANAGER.
So you must, again exactly as the info center says, manually start the failed instance, if you wish it to carry on as a secondary instance and provide failover in case the now primary instance fails as well.
But it doesn't help at all to try this if the reason the primary instance failed is because it's CPU halted and let out the magic smoke. *THINK*. |
|
| Back to top |
|
|
| rcp_mq |
| Posted: Tue Mar 27, 2012 7:06 am Post subject: |
|
|
Centurion
Joined: 13 Dec 2011
Posts: 133
|
@mqjeff
I shouldv'e been elaborate.
From what i infer
Say, A is primary B is an Instance of A, so
A(FAILS...DEAD)------>B
B(FAILS)------->A(DEAD...strmqm...LIVE)
[B...strmqm 'cause A&B both should be LIVE]
Am i correct? |
|
| Back to top |
|
|
| rcp_mq |
| Posted: Tue Mar 27, 2012 7:11 am Post subject: |
|
|
Centurion
Joined: 13 Dec 2011
Posts: 133
|
Missed an important point
A(FAILS...DEAD...AUTOMATIC)------>B
B(FAILS...MANUAL)------->A(DEAD...strmqm...LIVE) |
|
| Back to top |
|
|
| mqjeff |
| Posted: Tue Mar 27, 2012 7:13 am Post subject: |
|
|
Grand Master
Joined: 25 Jun 2008
Posts: 17447
|
Stop infering.
Read, think, try. |
|
| Back to top |
|
|
| bruce2359 |
| Posted: Tue Mar 27, 2012 7:14 am Post subject: |
|
|
Poobah
Joined: 05 Jan 2008
Posts: 9469
Location: US: west coast, almost. Otherwise, enroute.
|
Please re-read Jeffs post.
A failed qmgr is no longer a qmgr. It has failed for any number of possible reasons.
To become a qmgr again, you must fix what is broken.
If you want the failed-but-now-fixed qmgr to be in the MI, then you must manually restart it.
_________________
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 |
|
|
| exerk |
| Posted: Tue Mar 27, 2012 7:17 am Post subject: |
|
|
Jedi Council
Joined: 02 Nov 2006
Posts: 6339
|
| mqjeff wrote: |
| ...IT'S DEAD. IT IS AN EX-QUEUE MANAGER... |
Is it a Norwegian Blue? 
_________________
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 |
|
|
| rcp_mq |
| Posted: Tue Mar 27, 2012 7:29 am Post subject: |
|
|
Centurion
Joined: 13 Dec 2011
Posts: 133
|
| I'm not saying primary converts to secondary...i'm saying secondary takes over primary and does this happen automatically?. |
|
| Back to top |
|
|
| Vitor |
| Posted: Tue Mar 27, 2012 7:29 am Post subject: |
|
|
Grand High Poobah
Joined: 11 Nov 2005
Posts: 26093
Location: Texas, USA
|
| rcp_mq wrote: |
Missed an important point
A(FAILS...DEAD...AUTOMATIC)------>B
B(FAILS...MANUAL)------->A(DEAD...strmqm...LIVE) |
You've missed an important point. If A FAILS (to use your capitalization) how can it restart, manually or automatically, when B fails?
I join my voice to the chorus - in this scenario A is DEAD. Broken. Failed. Gone to join the Choir Invisible. Before it's going to do anything useful you'll need to fix it.
This is equally true for HA clusters. Typically failover is automatic, failback is manual because it can't be performed until the primary is available again.
If what you're trying to achieve is a fully automated A -> B -> A -> B every time there's a failure, you can't & you shouldn't want to.
_________________
Honesty is the best policy.
Insanity is the best defence. |
|
| Back to top |
|
|
| mqjeff |
| Posted: Tue Mar 27, 2012 7:31 am Post subject: |
|
|
Grand Master
Joined: 25 Jun 2008
Posts: 17447
|
And, to clarify the other point that you're struggling to make.
A working secondary instance is always available to receive a failed primary instance. |
|
| Back to top |
|
|
| Vitor |
| Posted: Tue Mar 27, 2012 7:33 am Post subject: |
|
|
Grand High Poobah
Joined: 11 Nov 2005
Posts: 26093
Location: Texas, USA
|
| rcp_mq wrote: |
| I'm not saying primary converts to secondary...i'm saying secondary takes over primary and does this happen automatically?. |
No it doesn't. The services running on the primary (a queue manager in the case of MI, any number in the case of HA) are failed over to the secondary. Many people have told you this happens automatically.
The secondary does not take over the primary, become the primary or turn into an elephant. The primary does not become the secondary, paint itself blue or suddenly become operational again.
_________________
Honesty is the best policy.
Insanity is the best defence. |
|
| Back to top |
|
|
|
|
