Author |
Message
|
SimplyAniket |
Posted: Mon Sep 22, 2008 1:26 am Post subject: Queue manager ended preemptively |
|
|
 Novice
Joined: 13 Sep 2008 Posts: 13
|
Hi,
I observed strange condition on my queue manager. I checked the status of queue manager and found the status as "ended preemptively" in error log i found the below error
EXPLANATION:
An error has been detected, and the WebSphere MQ error recording routine has
been called. The failing process is process 5724.
ACTION:
Use the standard facilities supplied with your system to record the problem
identifier, and to save the generated output files. Contact your IBM support
center. Do not discard these files until the problem has been resolved.
----- amqxfdcp.c : 804 --------------------------------------------------------
EXPLANATION:
This message is issued when an attempt to write a log record is rejected
because the log is full. The queue manager will attempt to resolve the problem.
ACTION:
This situation may be encountered during a period of unusually high message
traffic. However, if you persistently fill the log, you may have to consider
enlarging the size of the log. You can either increase the number of log files
by changing the values in the queue manager configuration file. You will then
have to stop and restart the queue manager. Alternatively, if you need to make
the log files themselves bigger, you will have to delete and recreate the queue
manager.
----- amqxfdcp.c : 766 --------------------------------------------------------
So basically i need to know below things
1. What are the reasons for queue manager getting shut down preemptively.
2. My queue manager log (i.e. data folder) is on D drive and program files i.e. mq program files are on C drive for MQ. I got the above error but my C drive got full. I am confused that when queue manager logs were written on D drive then why C drive got issues.
Kindly suggest  |
|
Back to top |
|
 |
exerk |
Posted: Mon Sep 22, 2008 1:39 am Post subject: |
|
|
 Jedi Council
Joined: 02 Nov 2006 Posts: 6339
|
You already know the answer!
SimplyAniket wrote: |
1. What are the reasons for queue manager getting shut down preemptively. |
Quote: |
EXPLANATION:
This message is issued when an attempt to write a log record is rejected because the log is full. The queue manager will attempt to resolve the problem. |
What log method are you using? What log values are you using? If you have very long running units of work, and a few, small, circular logs, you can get this condition without filling disk. _________________ 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 |
|
 |
SimplyAniket |
Posted: Mon Sep 22, 2008 3:22 am Post subject: |
|
|
 Novice
Joined: 13 Sep 2008 Posts: 13
|
I am using Circular logging. Details of logging are
Primary log files - 5
Secondary log files - 45
Log buffer pages - 512
Log file pages - 16384
Long running units of work is one reason . But we do not have such long running unit of works.
My Data folder drive has 54 GB and C drive had less space remained. |
|
Back to top |
|
 |
fjb_saper |
Posted: Mon Sep 22, 2008 3:27 am Post subject: |
|
|
 Grand High Poobah
Joined: 18 Nov 2003 Posts: 20756 Location: LI,NY
|
Did FDCs fill up your <MQDATA>/errors directory space??  _________________ MQ & Broker admin |
|
Back to top |
|
 |
bruce2359 |
Posted: Mon Sep 22, 2008 6:18 am Post subject: |
|
|
 Poobah
Joined: 05 Jan 2008 Posts: 9469 Location: US: west coast, almost. Otherwise, enroute.
|
Quote: |
But we do not have such long running unit of works |
One of the other major influence on log file usage is message size. What
size persistent messages do your applications get/put?
Another influence: Do you record mq object imgess with rcdmqimg? _________________ 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: Mon Sep 22, 2008 6:52 am Post subject: |
|
|
 Jedi Council
Joined: 02 Nov 2006 Posts: 6339
|
SimplyAniket wrote: |
I am using Circular logging... |
_________________ 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 |
|
 |
bruce2359 |
Posted: Mon Sep 22, 2008 9:10 am Post subject: |
|
|
 Poobah
Joined: 05 Jan 2008 Posts: 9469 Location: US: west coast, almost. Otherwise, enroute.
|
Quote: |
Another influence: Do you record mq object imgess with rcdmqimg? |
Ah, yes. Disregard. This is not available with circular logging. Speed-reading failed me again. _________________ 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 |
|
 |
SimplyAniket |
Posted: Tue Sep 23, 2008 1:31 am Post subject: |
|
|
 Novice
Joined: 13 Sep 2008 Posts: 13
|
There was only one FDC generated of size 508,031 KB but it is on D drive. C drive was full with system files so it had low space left. My main concern how come quue manager logs got full when D drive had 54 GB space left. Is it the case that my both primary and secondary logs got full..  |
|
Back to top |
|
 |
Gaya3 |
Posted: Tue Sep 23, 2008 2:16 am Post subject: |
|
|
 Jedi
Joined: 12 Sep 2006 Posts: 2493 Location: Boston, US
|
open/read the FDC,
In the middle of FDC, you can see the Environment Variables details
try to find out which path is pointing to C:\ and what it is.
fetch the probe id and Description, Search and let us know too _________________ Regards
Gayathri
-----------------------------------------------
Do Something Before you Die |
|
Back to top |
|
 |
SimplyAniket |
Posted: Tue Sep 23, 2008 10:31 pm Post subject: |
|
|
 Novice
Joined: 13 Sep 2008 Posts: 13
|
AMQ6709: The log for the Queue manager is full.
EXPLANATION:
This message is issued when an attempt to write a log record is rejected
because the log is full. The queue manager will attempt to resolve the problem.
ACTION:
This situation may be encountered during a period of unusually high message
traffic. However, if you persistently fill the log, you may have to consider
enlarging the size of the log. You can either increase the number of log files
by changing the values in the queue manager configuration file. You will then
have to stop and restart the queue manager. Alternatively, if you need to make
the log files themselves bigger, you will have to delete and recreate the queue
manager.
Same situation occured on replica of the same queue manager on which same operations occur.
Moreover i checked the number of log files Primary -5 n secondary - 15 .. but in mqmdata\log\qmgr directory i saw 49 files (i.e. S000XX.LOG files) Is it possible? ideally there should be only 20 files including primary and secondary.. How is it possible? |
|
Back to top |
|
 |
Gaya3 |
Posted: Tue Sep 23, 2008 10:45 pm Post subject: |
|
|
 Jedi
Joined: 12 Sep 2006 Posts: 2493 Location: Boston, US
|
Kindly check the below website;
http://www-304.ibm.com/jct09002c/isv/tech/faq/individual.jsp?oid=1:19040
This error indicates that you have run out of space in the recovery log, most likely due to long running transactions. Here are a few suggestions for correcting this error:
• Increase the number of log files.
• Increase the size of the log file.
• Issue commits more frequently in your application.
• Reduce the batch interval on your channels.
You mentioned the queue files are more than 65 GB, I suspect there could be lot of messages waiting to complete the transactions (commit)
AMQ6709
The log for the Queue manager is full. Severity
20 : Error Explanation
This message is issued when an attempt to write a log record is rejected because the log is full. The queue manager will attempt to resolve the problem.
Response
This situation may be encountered during a period of unusually high message traffic. However, if you persistently fill the log, you may have to consider enlarging the size of the log. You can either increase the number of log files by changing the values in the queue manager configuration file. You will then have to stop and restart the queue manager. Alternatively, if you need to make the log files themselves bigger, you will have to delete and recreate the queue manager. _________________ Regards
Gayathri
-----------------------------------------------
Do Something Before you Die |
|
Back to top |
|
 |
exerk |
Posted: Tue Sep 23, 2008 10:48 pm Post subject: |
|
|
 Jedi Council
Joined: 02 Nov 2006 Posts: 6339
|
SimplyAniket wrote: |
...ideally there should be only 20 files including primary and secondary.. How is it possible? |
Yet in a previous post you stated:
SimplyAniket wrote: |
Primary log files - 5
Secondary log files - 45 |
Check that you have sufficient space within the log folder to increase the number of logs (as this is Windows, so check a quota hasn't been set on the directory). If you have full use of that 54GB of space, increase the number of Primary logs and attempt a restart and find out what is filling your logs to such an extent that they can't wrap around.
EDIT: Gaya3 wrote "You mentioned the queue files are more than 65 GB"...where did that information come from? Because I can't find it in any of the posts above. _________________ 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 |
|
 |
TRANSPORTATOR |
Posted: Wed Sep 24, 2008 12:27 pm Post subject: MQ |
|
|
Apprentice
Joined: 04 May 2006 Posts: 27
|
the log as a secondary maximum is 63  |
|
Back to top |
|
 |
exerk |
Posted: Wed Sep 24, 2008 12:59 pm Post subject: Re: MQ |
|
|
 Jedi Council
Joined: 02 Nov 2006 Posts: 6339
|
TRANSPORTATOR wrote: |
the log as a secondary maximum is 63  |
More accurately, on V5.3 was 63 logs total (minimum 2 primary) but on V6 it's 254 (minimum 2 primary) on Windows, 510 (minimum 2 primary) on Unix. _________________ 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 |
|
 |
|