| Author |
Message
|
| tkieran |
 Posted: Fri May 13, 2005 3:35 am Post subject: Transactions rolled back but then succeed once resubmitted |
 |
|
Newbie
Joined: 13 May 2005
Posts: 2
|
Hi Folks,
We have a nightly job that runs at 22:30 and trys to push approx 9,000 messages through MQ with a final commit at the end (not a good idea).
This job fails each night at 23:00 with the message
AMQ7469: Transactions rolled back to release log space.
This is the only job that is going through MQ at this time.
However if we re-submit this job or put the job on hold until after 23:00 it will succeed.
I have checked all our cron and at jobs on this system and nothing is happening at this time. When I check the active log directory the first active log is created roughly at the same time of the job failure 23:02.
This is our log params from mq.ini
Log:
LogPrimaryFiles=30
LogSecondaryFiles=2
LogFilePages=1024
LogType=LINEAR
LogBufferPages=17
LogPath=/MQHA/QM_TIBMQS2/log/QM_TIBMQS2/
I think the log params is okay as the job will succeed once run after 23:00. Is there anyting within MQ I can check to see what is running at this time? 
_________________
TonyK |
|
| Back to top |
|
 |
| kevinf2349 |
| Posted: Fri May 13, 2005 5:34 am Post subject: |
|
|
Grand Master
Joined: 28 Feb 2003
Posts: 1311
Location: USA
|
Isn't that what schedulers are for? 
What is starting the job off? |
|
| Back to top |
|
|
| tkieran |
| Posted: Fri May 13, 2005 5:52 am Post subject: |
|
|
Newbie
Joined: 13 May 2005
Posts: 2
|
The job is kicked off within a VMS scheduler on a remote VMS server.
The job first extracts data from an Oracle db and use OPG to get the messages over to MQ where the data is feed into a remote MIS database. The job has dependencies which are also held back when this job is held back thus we are trying to get it to go before 23:00.
_________________
TonyK |
|
| Back to top |
|
|
| fjb_saper |
| Posted: Fri May 13, 2005 10:52 am Post subject: |
|
|
Grand High Poobah
Joined: 18 Nov 2003
Posts: 20756
Location: LI,NY
|
Are you by any chance running rcdmqimg between 22:30 and 24:00 ?.
This process frees up old logs. I would probably just add a bunch more of secondaries to your setup. 2 is an awfully short number.
We run with 2 primaries and 50 secondaries...
 |
|
| Back to top |
|
|
| seunwoby |
| Posted: Wed Jun 01, 2005 1:33 am Post subject: |
|
|
Newbie
Joined: 22 Apr 2003
Posts: 1
Location: England
|
| check that your Maximum Uncommited Message Threshold has not been exceeded. |
|
| Back to top |
|
|
| fjb_saper |
| Posted: Wed Jun 01, 2005 3:29 am Post subject: |
|
|
Grand High Poobah
Joined: 18 Nov 2003
Posts: 20756
Location: LI,NY
|
| seunwoby wrote: |
| check that your Maximum Uncommited Message Threshold has not been exceeded. |
Seuwoby might be on to something here. The default value is 10,000. You say that your batch process is loading 9,000 messages. That might hit your max threshold at the first go round the qmgr being otherwise busy.
Until the second go round the qmgr had the time to process or commit the outstanding 1,000 uncommitted messages and now the load will go through.
Try upping the limit to 20,000.
Enjoy  |
|
| Back to top |
|
|
| CShep |
| Posted: Mon Nov 27, 2006 7:24 am Post subject: MAX UNCOMMITED MESSAGES |
|
|
Novice
Joined: 20 Oct 2004
Posts: 22
Location: London, England
|
Hi,
Can anyone confirm if the MAXIMUM UNCOMMITED MESSAGES value for the Queue Manager is a total figure, across all connections to the Queue Manager? or per MQCONN? or per MQOPEN?
Sorry if this has been answered somewhere else before...
Thanks,
_________________
A friend in need is a pain in the @ss |
|
| Back to top |
|
|
| fjb_saper |
| Posted: Mon Nov 27, 2006 2:42 pm Post subject: |
|
|
Grand High Poobah
Joined: 18 Nov 2003
Posts: 20756
Location: LI,NY
|
As far as I know it is the total figure for the qmgr, evidenced by the fact that it can be changed through alter qmgr.... You will probably need to recycle the qmgr before the new number takes effect.
_________________
MQ & Broker admin |
|
| Back to top |
|
|
| exerk |
| Posted: Mon Nov 27, 2006 2:58 pm Post subject: |
|
|
Jedi Council
Joined: 02 Nov 2006
Posts: 6339
|
| fjb_saper wrote: |
Are you by any chance running rcdmqimg between 22:30 and 24:00 ?.
This process frees up old logs. I would probably just add a bunch more of secondaries to your setup. 2 is an awfully short number.
We run with 2 primaries and 50 secondaries...
|
According to his qm.ini the log type is LINEAR...surely the number of P & S logs is irrelevant?
_________________
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 |
|
|
| CShep |
| Posted: Tue Nov 28, 2006 12:01 am Post subject: I guessed as much ... |
|
|
Novice
Joined: 20 Oct 2004
Posts: 22
Location: London, England
|
Thanks fjb_saper
_________________
A friend in need is a pain in the @ss |
|
| Back to top |
|
|
| fjb_saper |
| Posted: Tue Nov 28, 2006 11:49 am Post subject: |
|
|
Grand High Poobah
Joined: 18 Nov 2003
Posts: 20756
Location: LI,NY
|
| exerk wrote: |
According to his qm.ini the log type is LINEAR...surely the number of P & S logs is irrelevant? |
Reread the chapter about linear logging and the max size of all transactions....
_________________
MQ & Broker admin |
|
| Back to top |
|
|
| exerk |
| Posted: Wed Nov 29, 2006 8:15 am Post subject: |
|
|
Jedi Council
Joined: 02 Nov 2006
Posts: 6339
|
| fjb_saper wrote: |
| exerk wrote: |
According to his qm.ini the log type is LINEAR...surely the number of P & S logs is irrelevant? |
Reread the chapter about linear logging and the max size of all transactions.... |
Have done so and believe you mean this...
"Although secondary log files are defined for linear logging, they are not used in normal operation. If a situation arises when, probably due to long-lived transactions, it is not possible to free a file from the active pool because it might still be required for a restart, secondary files are formatted and added to the active log file pool."
I'm still a bit confused as the preceding paragraph states that if a transaction spans a log another log will be formatted, and so on until disk space runs out; this implies that the active pool will be as big as disk allows.
Am I having a 'senior' moment here where the obvious is staring me in the face but I just can't see it?
_________________
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 Nov 29, 2006 8:22 am Post subject: |
|
|
Grand High Poobah
Joined: 11 Nov 2005
Posts: 26093
Location: Texas, USA
|
Another log will be formatted until the total number of log files in the active pool = number of primary logs defined + number of secondary logs defined.
In the example quoted by fjb_saper a maximum of 52 log files could exist in the active pool. But he would seldom if ever have more than 2.
_________________
Honesty is the best policy.
Insanity is the best defence. |
|
| Back to top |
|
|
|
|
