Author |
Message
|
Boomn4x4 |
Posted: Tue May 01, 2012 9:16 am Post subject: Backup qmgr log files not incrimenting. |
|
|
Disciple
Joined: 28 Nov 2011 Posts: 172
|
I'm working on my back up procedures. I have a backup queue manager on a second piece of hardware. What I'm struggling to understand is why the backup queue manager does not seem to be incrementing its log file.
What I'm doing:
1. Advancing the logs - RESET QMGR TYPE(ADVANCELOG)
2. Getting the current log extent - DIS QMSTATUS CURRLOG
3. Moving the CURRLOG - 1 to the backup system
4. Replaying the logs on the backup system STRMQM -R $QMGR
The log files replay, but the response message is always The oldest log file required to start queue manager QA_TEST_QM is S0000000.LOG.
Shouldn't that number be incrementing with each replay? If the system has already been replayed from 00.LOG then why do I need to keep starting at that point? I must be doing something wrong as this would only result in an infinity large log structure. |
|
Back to top |
|
 |
Vitor |
Posted: Tue May 01, 2012 9:58 am Post subject: Re: Backup qmgr log files not incrimenting. |
|
|
 Grand High Poobah
Joined: 11 Nov 2005 Posts: 26093 Location: Texas, USA
|
Boomn4x4 wrote: |
Shouldn't that number be incrementing with each replay? |
What does the active queue manager say is the oldest log file needed for it to restart? If my suspicion is correct, that will say it needs log 0 as well.
Boomn4x4 wrote: |
I must be doing something wrong as this would only result in an infinity large log structure. |
As you would with an incorrectly managed non-MI linear log structure. You need to take new media images in the log to advance that number. _________________ Honesty is the best policy.
Insanity is the best defence. |
|
Back to top |
|
 |
Boomn4x4 |
Posted: Tue May 01, 2012 11:05 am Post subject: |
|
|
Disciple
Joined: 28 Nov 2011 Posts: 172
|
Where this leads me to then, is am I correct in that I'm misunderstanding how linear logging works with media recovery in that if I am doing linear logging, I MUST also take routine media images, at least if I want to ever be able to delete log files from the backup?
So in reading about the media image, I'm getting more confused.
I deleted the backup qmgr and recreated it... to get a clean start.
1: Advance the logs
2: Copy the previous log to the backup
3. Replay the previous log on the backup
4. The backup says the restart log is advanced, but the media recovery log is the same
5. On the primary, run 'rcdmqimg' to record a media image to the log files
6. Advance the logs again
7. Copy the previous log to the backup
8. Replay the previous log on the backup
My assumption is that once I record a media image, the new media image is written to the log files, when I replay that log file on the backup would roll forward. However, it isn't.
What am I missing? |
|
Back to top |
|
 |
mqjeff |
Posted: Tue May 01, 2012 11:09 am Post subject: |
|
|
Grand Master
Joined: 25 Jun 2008 Posts: 17447
|
what happens if you start the backup qmgr? |
|
Back to top |
|
 |
Boomn4x4 |
Posted: Tue May 01, 2012 11:12 am Post subject: |
|
|
Disciple
Joined: 28 Nov 2011 Posts: 172
|
mqjeff wrote: |
what happens if you start the backup qmgr? |
For replay, or just try to start it?
If I just try to start it, I'm told its not activated (as it should)
if I start it for replay, -r:
WebSphere MQ queue manager 'QA_POSLAB06_QM' is being started for replay.
The oldest log file required to start queue manager QA_POSLAB06_QM is S0000024.LOG.
The oldest log file required to perform media recovery of queue manager QA_POSLAB06_QM is S0000021.LOG.
WebSphere MQ queue manager 'QA_POSLAB06_QM' replay completed. |
|
Back to top |
|
 |
Boomn4x4 |
Posted: Tue May 01, 2012 11:19 am Post subject: |
|
|
Disciple
Joined: 28 Nov 2011 Posts: 172
|
Okay... I'm scratching my head even more now.
I saw in the documentation that each time the queue manager is shut down, it takes a media image. So, for poops and giggles, I bounced the QMGR, advanced the logs, copied them over.... and now magically the media image has been advanced on the backup.
The oldest log file required to start queue manager QA_POSLAB06_QM is S0000026.LOG.
The oldest log file required to perform media recovery of queue manager QA_POSLAB06_QM is S0000025.LOG.
WebSphere MQ queue manager 'QA_POSLAB06_QM' replay completed.
So, my assumptions were correct in the first place (I think) so why does the 'rcdmgimg' command not do the same thing? |
|
Back to top |
|
 |
mqjeff |
Posted: Tue May 01, 2012 11:24 am Post subject: |
|
|
Grand Master
Joined: 25 Jun 2008 Posts: 17447
|
I have doubts that the output you are seeing reflects the state of the queue manager after the log replay is complete.
I suspect it is the state of the queue manager before the log replay. |
|
Back to top |
|
 |
Boomn4x4 |
Posted: Tue May 01, 2012 11:48 am Post subject: |
|
|
Disciple
Joined: 28 Nov 2011 Posts: 172
|
mqjeff wrote: |
I have doubts that the output you are seeing reflects the state of the queue manager after the log replay is complete.
I suspect it is the state of the queue manager before the log replay. |
.... I'll be damned
The output I'm seeing is:
WebSphere MQ queue manager 'QA_POSLAB06_QM' is being started for replay.
The oldest log file required to start queue manager QA_POSLAB06_QM is S0000037.LOG.
The oldest log file required to perform media recovery of queue manager QA_POSLAB06_QM is S0000032.LOG.
WebSphere MQ queue manager 'QA_POSLAB06_QM' replay completed.
The output in var/mqm/qmgrs/QMGR/error/AMQ.LOG is:
/05/01/2012 03:45:12 PM - Process(31972.1) User(root) Program(amqzxma0)
Host(mgra8702)
AMQ7467: The oldest log file required to start queue manager QA_POSLAB06_QM is
S0000037.LOG.
EXPLANATION:
The log file S0000037.LOG contains the oldest log record required to restart
the queue manager. Log records older than this may be required for media
recovery.
ACTION:
You can move log files older than S0000037.LOG to an archive medium to release
space in the log directory. If you move any of the log files required to
recreate objects from their media images, you will have to restore them to
recreate the objects. An older log file is one with a numerically smaller log
number (but allowing for log number wrapping at 9999999).
-------------------------------------------------------------------------------
05/01/2012 03:45:12 PM - Process(31972.1) User(root) Program(amqzxma0)
Host(mgra8702)
AMQ7468: The oldest log file required to perform media recovery of queue
manager QA_POSLAB06_QM is S0000037.LOG.
EXPLANATION:
The log file S0000037.LOG contains the oldest log record required to recreate
any of the objects from their media images. Any log files prior to this will
not be accessed by media recovery operations.
ACTION:
Use this information together with the information in the most recent AMQ7467
message. Archivable log files are all those older than BOTH S0000037.LOG and
the log file mentioned in the AMQ7467 message.
Any explanation? |
|
Back to top |
|
 |
mqjeff |
Posted: Tue May 01, 2012 12:06 pm Post subject: |
|
|
Grand Master
Joined: 25 Jun 2008 Posts: 17447
|
The messages are output as part of normal queue manager start up processing, before it starts to process the log replay actions. |
|
Back to top |
|
 |
|