Author |
Message
|
sharonvk |
Posted: Wed Sep 21, 2016 4:03 am Post subject: QMGR not releasing logs properly |
|
|
Newbie
Joined: 21 Sep 2016 Posts: 8
|
HI
I'm running MQ 7506 on Linux REDHUT machine , which I have migrated to it recently , it is a test environment.
recently I have got calls from the application team that mq is not working
after searching I noticed the log disk is full.
We are using a linear logs.
I run the rcdmqimg and noticed that it stated that the oldest log required is very old ( for example I got like 100 logs from 1-100 and it requested log number 1, the oldest )
I then took down the qmgr and tried again , it still returned the old log file
I started the qmgr , and checked the command again now it stated a different log file quite a new one and I could delete most of the logs
I wonder if anyone has encountered such behavior and if there is a solution for this . ( I don't want to take down the qmgr in production ...) |
|
Back to top |
|
|
mqjeff |
Posted: Wed Sep 21, 2016 4:13 am Post subject: |
|
|
Grand Master
Joined: 25 Jun 2008 Posts: 17447
|
What you've seen is normal behavior, at least in terms of the log files.
I can't speak to the behavior of rcdmqimg, as I've not run it in quite a while.
If the problem is that the logs are filling up the disk, then either the logs are too big or the disk is too small... _________________ chmod -R ugo-wx / |
|
Back to top |
|
|
sharonvk |
Posted: Wed Sep 21, 2016 4:32 am Post subject: |
|
|
Newbie
Joined: 21 Sep 2016 Posts: 8
|
Hi
Why do you consider this as normal behavior , should the qmgr hold the old log and only after a restart release it.
as I stated the old log was like 100 and more logs and took up a lot of space in the disk .
if it is not needed why it is held , is this not a bug .... |
|
Back to top |
|
|
mqjeff |
Posted: Wed Sep 21, 2016 4:39 am Post subject: |
|
|
Grand Master
Joined: 25 Jun 2008 Posts: 17447
|
The logs should be rolled over at some point. It's not clear that you reached that point.
Was the size of the 100th log file different than the size of the others?
Also, if it ran out of disk space, it might not have been able to roll over the log files. _________________ chmod -R ugo-wx / |
|
Back to top |
|
|
fjb_saper |
Posted: Wed Sep 21, 2016 7:53 pm Post subject: |
|
|
Grand High Poobah
Joined: 18 Nov 2003 Posts: 20729 Location: LI,NY
|
What are you doing to automate the maintenance of the log files in linear logging?
What support pack are you using for this task? (MS0L?) _________________ MQ & Broker admin |
|
Back to top |
|
|
Andyh |
Posted: Wed Sep 21, 2016 11:01 pm Post subject: |
|
|
Master
Joined: 29 Jul 2010 Posts: 239
|
When you run rcdmqimg, is it successfully recording a media image for all MQ objects ?
For example, do you have any damaged queues, as you have to recover (or delete) any damaged queues before the qmgr medialog can be advanced beyond the medialog for that specific queue.
If you run "DIS QSTATUS(*) MEDIALOG" does it show some queue as having an 'old' medialog ?
Are there any FDC's ? |
|
Back to top |
|
|
sharonvk |
Posted: Tue Sep 27, 2016 1:35 am Post subject: |
|
|
Newbie
Joined: 21 Sep 2016 Posts: 8
|
Hi
Thanks for the replies
1) regarding maintaining the linear logs , I have a script that run a the rcdmqimg and then delete all the old logs ,
this scripts runs twice a day . midday and midnight and it working fine in several application in production .
Andyh
an FDC is created stating that the resource for log space is out. I can record it to you if you think it has any additional information .
as I run the rcdmqimg in a script I haven't noticed if there were errors.
I will check the next time this event occurs .
mqjeff has I deleted the logs I haven't had the chance to check what you asked , I will do it the next time
can you explain why after shutting down the qmgr , the rdmqimg still didn't free the logs , only after starting it up the logs were free |
|
Back to top |
|
|
umatharani |
Posted: Thu Sep 29, 2016 5:31 pm Post subject: |
|
|
Apprentice
Joined: 23 Oct 2008 Posts: 39
|
Hi,
As Andy H indicated, if there is a failure in recording media image then starting log extent required for the media recovery will not change. Could you please provide the output of ffstsummary command executed from /var/mqm/errors directory.
Thanks, mahesh |
|
Back to top |
|
|
|