Author |
Message
|
vinumon |
Posted: Wed May 07, 2014 5:07 am Post subject: Perform switchover from PROD to resilience site for MQ |
|
|
Acolyte
Joined: 06 May 2014 Posts: 59
|
What are the considerations need to take after PROD server goes down
and resilience site is up. Will message gets lost?
Actually we are having data and log mount points shared betwen PROD and resilience servers. So what happens to messages on PROD server when it goes down. |
|
Back to top |
|
 |
shashivarungupta |
Posted: Wed May 07, 2014 5:17 am Post subject: |
|
|
 Grand Master
Joined: 24 Feb 2009 Posts: 1343 Location: Floating in space on a round rock.
|
The Question is , messages are persistent ?
If no, you can not recover them. _________________ *Life will beat you down, you need to decide to fight back or leave it. |
|
Back to top |
|
 |
Vitor |
Posted: Wed May 07, 2014 5:24 am Post subject: |
|
|
 Grand High Poobah
Joined: 11 Nov 2005 Posts: 26093 Location: Texas, USA
|
Another consideration is the size of your units of work. Uncommitted messages (persistent or not) may be lost.
@shashivarungupta: To slightly clarify, committed persistent messages will be present on restart at the resilience site. Non-persistent messages may not be and should not be assumed to be.
If the data really is a shared SAN across 2 sites the OP may find some non-persistent messages survive the cut. If it's replicated data less so.
But as you correctly point out, any non-persistent message cannot be assumed to be present on the resilience site and any such messages not present are unrecoverable.
 _________________ Honesty is the best policy.
Insanity is the best defence. |
|
Back to top |
|
 |
shashivarungupta |
Posted: Wed May 07, 2014 5:30 am Post subject: |
|
|
 Grand Master
Joined: 24 Feb 2009 Posts: 1343 Location: Floating in space on a round rock.
|
Vitor wrote: |
To slightly clarify, committed persistent messages will be present on restart at the resilience site. Non-persistent messages may not be and should not be assumed to be. |
 _________________ *Life will beat you down, you need to decide to fight back or leave it. |
|
Back to top |
|
 |
vinumon |
Posted: Wed May 07, 2014 6:13 am Post subject: |
|
|
Acolyte
Joined: 06 May 2014 Posts: 59
|
Thanks for the reply, But one question, will there be duplicate messages on resilience site once PROD goes down? |
|
Back to top |
|
 |
JosephGramig |
Posted: Wed May 07, 2014 6:48 am Post subject: |
|
|
 Grand Master
Joined: 09 Feb 2006 Posts: 1244 Location: Gold Coast of Florida, USA
|
No. You say there is only one data and logs file system and it is shared between the sites and that is why there will be no duplicates. |
|
Back to top |
|
 |
PeterPotkay |
Posted: Wed May 07, 2014 10:13 am Post subject: |
|
|
 Poobah
Joined: 15 May 2001 Posts: 7722
|
Assuming the storage in both sites is being updated synchronously, so that both copies are EXACTLY the same at all times, no issue.
If you are replicating the data asynchronously from one site to the other, all bets are off. You can have missing messages. You can have already deleted messages back again. _________________ Peter Potkay
Keep Calm and MQ On |
|
Back to top |
|
 |
Vitor |
Posted: Wed May 07, 2014 10:30 am Post subject: |
|
|
 Grand High Poobah
Joined: 11 Nov 2005 Posts: 26093 Location: Texas, USA
|
PeterPotkay wrote: |
If you are replicating the data asynchronously from one site to the other, all bets are off. You can have missing messages. You can have already deleted messages back again. |
The OP did say "shared". We must have faith.  _________________ Honesty is the best policy.
Insanity is the best defence. |
|
Back to top |
|
 |
Andyh |
Posted: Wed May 07, 2014 11:29 am Post subject: |
|
|
Master
Joined: 29 Jul 2010 Posts: 239
|
"Uncommitted messages (persistent or not) may be lost"
The idea that an uncommitted message could be lost is an oxymoron to me.
Any indoubt transaction would be restored to its correct indoubt state, any active transaction would be (correctly) rolled back following a failover. |
|
Back to top |
|
 |
Andyh |
Posted: Wed May 07, 2014 11:36 am Post subject: |
|
|
Master
Joined: 29 Jul 2010 Posts: 239
|
"If the data really is a shared SAN across 2 sites the OP may find some non-persistent messages survive the cut."
If the customer is using NPMCLASS(HIGH) for some queues and a controlled failover occurs (where the active instance is ended in a controlled fashion) then NPM messages for those queues will typically be recovered on the backup.
However failover to a disaster recovery site rarely occurs in a controlled fashion and in the event that the active does not end normally, or NPMCLASS(HIGH) is not enabled, then all non-persistent messages will be lost across the failover. |
|
Back to top |
|
 |
|