| Author |
Message
|
| vincent |
 Posted: Fri Jun 02, 2006 9:12 am Post subject: runmqtrm error messages |
 |
|
Voyager
Joined: 21 Apr 2006
Posts: 81
|
-----------------------------------------------------------------------------+
| |
| WebSphere MQ First Failure Symptom Report |
| ========================================= |
| |
| Date/Time :- Tuesday May 16 13:43:13 EDT 2006 |
| Host Name :- nydsdev02 (SunOS 5. |
| PIDS :- 5724B4103 |
| LVLS :- 530.12 CSD12 |
| Product Long Name :- WebSphere MQ for Sun Solaris |
| Vendor :- IBM |
| Probe Id :- XC338001 |
| Application Name :- MQM |
| Component :- xehAsySignalHandler |
| Build Date :- Dec 8 2005 |
| CMVC level :- p530-12-L051208 |
| Build Type :- IKAP - (Production) |
| UserID :- 00003101 (mqm) |
| Program Name :- runmqtrm |
| Process :- 00019407 |
| Thread :- 00000002 |
| Major Errorcode :- xecE_W_UNEXPECTED_ASYNC_SIGNAL |
| Minor Errorcode :- OK |
| Probe Type :- MSGAMQ6209 |
| Probe Severity :- 3 |
| Probe Description :- AMQ6109: An internal WebSphere MQ error has occurred. |
| FDCSequenceNumber :- 0 |
| Arith1 :- 2 2 |
| Comment1 :- SIGINT |
| Comment2 :- Signal sent by pid 0 |
| Comment3 :- runmqtrm |
| |
+-----------------------------------------------------------------------------+
Why we got this report? Probably runmqtrm might have started but thrown error messages. |
|
| Back to top |
|
 |
| jefflowrey |
| Posted: Fri Jun 02, 2006 9:28 am Post subject: |
|
|
Grand Poobah
Joined: 16 Oct 2002
Posts: 19981
|
Did root issue a kill command against the PID?
Or, wait. pid 0 is init, isn't it?
_________________
I am *not* the model of the modern major general. |
|
| Back to top |
|
|
| vincent |
| Posted: Fri Jun 02, 2006 9:46 am Post subject: |
|
|
Voyager
Joined: 21 Apr 2006
Posts: 81
|
| I do not think so but system performing very badly after getting this report |
|
| Back to top |
|
|
| Nigelg |
| Posted: Fri Jun 02, 2006 12:20 pm Post subject: |
|
|
Grand Master
Joined: 02 Aug 2004
Posts: 1046
|
| Quote: |
| Why we got this report? |
Because you pressed ^C on the console where runmqtrm was running in the foreground. The signal is caught and ignored, and the FFST produced.
init is PID 1. The 0 in the FFST probably means it cannot determine the PID which sent the signal.
Any degradation in performance is nothing to do with the FFST.
_________________
MQSeries.net helps those who help themselves.. |
|
| Back to top |
|
|
| vincent |
| Posted: Fri Jun 02, 2006 1:17 pm Post subject: |
|
|
Voyager
Joined: 21 Apr 2006
Posts: 81
|
Thanks!!
Do you think any other reason for this bad performance degradation.
If it is like so, Do we require to re-install MQ5.3 CSD12? Is thery any bad report on Solaris-8 on Fijitsu system? |
|
| Back to top |
|
|
| wschutz |
| Posted: Fri Jun 02, 2006 1:32 pm Post subject: |
|
|
Jedi Knight
Joined: 02 Jun 2005
Posts: 3316
Location: IBM (retired)
|
What do you mean by "bad system performance"? Can you be much more specific? And is it only MQ that performing badly or is this a system wide problem? Have you looked at key system metrics (cpu util, disk io contention, etc)?
_________________
-wayne |
|
| Back to top |
|
|
| vincent |
| Posted: Fri Jun 02, 2006 1:36 pm Post subject: |
|
|
Voyager
Joined: 21 Apr 2006
Posts: 81
|
hi wschutz, I think you did not get the whole problem:)
The system was working fine until the log was full. Once the log was deleted, it started performing very badly. We removed all sharemed memories,semaphotes and restarted the system too. There is no luck.
Is there any problem with Fijitsu system which has solaris 8 latest fix pack? |
|
| Back to top |
|
|
| jefflowrey |
| Posted: Fri Jun 02, 2006 1:43 pm Post subject: |
|
|
Grand Poobah
Joined: 16 Oct 2002
Posts: 19981
|
Vincent.
You have to try to be more specific about "performing very badly".
You have to find out exactly which pieces of the entire system are performing very badly. Do you see slow performance on something like an "ls" command? Or only for MQ?
You say you restarted the whole system - you mean the Solaris environment - like a "shutdown -r now"? Or did you do an init to restart all the services?
_________________
I am *not* the model of the modern major general. |
|
| Back to top |
|
|
| wschutz |
| Posted: Fri Jun 02, 2006 1:48 pm Post subject: |
|
|
Jedi Knight
Joined: 02 Jun 2005
Posts: 3316
Location: IBM (retired)
|
| Quote: |
| The system was working fine until the log was full. Once the log was deleted, it started performing very badly. |
What log? You haven't mentioned any log before....
_________________
-wayne |
|
| Back to top |
|
|
| vincent |
| Posted: Fri Jun 02, 2006 2:13 pm Post subject: |
|
|
Voyager
Joined: 21 Apr 2006
Posts: 81
|
Jeff,
we have restarted the solaris system using shutdown command.
The log was set to LINEAR earleir , so the log file was full and asked UNIX admin to clean it , he physically removed the file by mistaken.
Then onwards, the system was not allowing us to issue any MQ commands.
What we did is that we killed all the MQ process and cleared all the semaphores and restarted the unix box using shutdown command. But still MQ is not performaing well.
We used to get
1 process to put the 512*20 bytes messages took around 4 minutes
1 process to get the same messages from the same Q took around 3 minutes
After restarting / log is full, it took around 12 and 10 seconds respectively to do the same task. |
|
| Back to top |
|
|
| jefflowrey |
| Posted: Fri Jun 02, 2006 2:23 pm Post subject: |
|
|
Grand Poobah
Joined: 16 Oct 2002
Posts: 19981
|
Okay, so the queue manager won't start properly if you removed the wrong linear log files.
_________________
I am *not* the model of the modern major general. |
|
| Back to top |
|
|
| vincent |
| Posted: Fri Jun 02, 2006 2:31 pm Post subject: |
|
|
Voyager
Joined: 21 Apr 2006
Posts: 81
|
jeff,
Any idea why MQ is taking around 12 seconds which was processed earlier in 4 seconds? Where should i look into to bring back the performance of the MQ? |
|
| Back to top |
|
|
| jefflowrey |
| Posted: Fri Jun 02, 2006 2:46 pm Post subject: |
|
|
Grand Poobah
Joined: 16 Oct 2002
Posts: 19981
|
If you deleted the wrong log files, I'm surprised the queue manager started up at all.
PLEASE CONFIRM THAT YOU ARE ONLY HAVING PERFORMANCE PROBLEMS WITH MQ, AND NOT WITH THE WHOLE BOX.
_________________
I am *not* the model of the modern major general. |
|
| Back to top |
|
|
| vincent |
| Posted: Fri Jun 02, 2006 2:55 pm Post subject: |
|
|
Voyager
Joined: 21 Apr 2006
Posts: 81
|
Jeff,
we had very tough time to restart the queue manager in the first place, We killed all the process and semaphores and rebooted the machine. I think the problem sould go off once the machine restarted.... But actually not..
I did not see any other problem in the system while unix commands are executed.
Even MQ process is also working fine but the same througput is not achieved after restarting the machine.... Looks like performance degradation in the MQ level |
|
| Back to top |
|
|
| jefflowrey |
| Posted: Fri Jun 02, 2006 3:00 pm Post subject: |
|
|
Grand Poobah
Joined: 16 Oct 2002
Posts: 19981
|
Stop the queue manager entirely again. Ensure that it stops on it's own, and clear the shared memory segments if they still hang around.
Ask your Solaris admin for assistance ensuring that the box itself is still healthy.
Copy all the existing ../error and ../qmgrs/<qmgrname>/error/ files to a temporary location, and then remove the originals.
Start the queue manager, wait for it to start.. And then stop the queue manager again.
Examine the new ../error and ../qmgrs/qmgrname/error files to see everything that is reported.
IF everything reports as normal, then you have a trickier job ahead of you. Hopefully this will not be the case.
What do the partitions look like? Is all of /var/mqm/ and /usr/mqm/ on one partition? Is it all on the local disk?
_________________
I am *not* the model of the modern major general. |
|
| Back to top |
|
|
|
|
