| Author |
Message
|
| feng11 |
 Posted: Wed Apr 23, 2003 12:13 am Post subject: MQ crashes and unable to start up |
 |
|
Newbie
Joined: 22 Apr 2003
Posts: 3
|
I have 3 queue managers running in one machine under the same account mqm.
2 weeks ago, one of the queue manager crash and unable to start up. Then we shutdown the other 2, and restart one by one. We tried different sequences, but when one of the 3 queue mgrs is started, the other 2 will crash ( these 2 came up ok). is there a shared memeory problem as indicated by the error thrown out as below ?
Has anyone hit the below problem before ?
Probe Id :- XY167002 |
| Application Name :- MQM |
| Component :- xstAllocBlockInSharedMemSet |
| Build Date :- Jan 5 2001 |
| UserID :- 00004003 (mqm) |
Program Name :- amqcrsta_nd |
| Process :- 00020459 |
| Thread :- 00000001 |
| Major Errorcode :- xecF_E_UNEXPECTED_RC |
| Minor Errorcode :- OK |
| Probe Type :- MSGAMQ6118 |
| Probe Severity :- 2 |
| Probe Description :- AMQ6118: An internal MQSeries error has occurred (0) |
MQM Trace History
<-- xstReleaseSerialisationOnExtent rc=OK
<-- xstAllocBlockInExtent rc=xecS_E_NO_MEM
--> xstAllocBlockInExtent
--> xstSerialiseExtent
--> xllSpinLockRequest
<-- xllSpinLockRequest rc=OK
<-- xstSerialiseExtent rc=OK |
|
| Back to top |
|
 |
| oz1ccg |
| Posted: Wed Apr 23, 2003 12:42 am Post subject: |
|
|
Yatiri
Joined: 10 Feb 2002
Posts: 628
Location: Denmark
|
What platform are you running on ?
Some Windows free platform, right ?
I've faced this problem prev. on HP-UX, where operations "just" KILLed processes this left the shared storage in use, and forced us to boot the box again to free up storage....
After this the operations team was WAKED up, and after that no problems.
Just my $0.02 
_________________
Regards, Jørgen
Home of BlockIP2, the last free MQ Security exit ver. 3.00
Cert. on WMQ, WBIMB, SWIFT. |
|
| Back to top |
|
|
| feng11 |
| Posted: Wed Apr 23, 2003 12:46 am Post subject: |
|
|
Newbie
Joined: 22 Apr 2003
Posts: 3
|
Machine is running Solaris Sun OS 2.8.
I tried to simulate this on devlopment, but it just doesn't crash. The problem happens maybe in 6months, maybe becos machine was not reboot at all.
Does that mean if i clear all the shared memory and semaphores, i can restart my queue managers properly? BTW, my mq is 5.1 |
|
| Back to top |
|
|
| oz1ccg |
| Posted: Wed Apr 23, 2003 1:01 am Post subject: |
|
|
Yatiri
Joined: 10 Feb 2002
Posts: 628
Location: Denmark
|
The only thing I know about Solaris is: it's a Sun box and it's able to run MQ... 
I'm not shure if you can release the memory, because it might be flagged as in use by the orphaned processes.
Normally it's applications that have the responsibility to free up shared the shared memory. There might be some commands on Solaris that allows you to release the storage, but anyway it might be dangerous.
Just my $0.02
_________________
Regards, Jørgen
Home of BlockIP2, the last free MQ Security exit ver. 3.00
Cert. on WMQ, WBIMB, SWIFT. |
|
| Back to top |
|
|
| feng11 |
| Posted: Wed Apr 23, 2003 1:09 am Post subject: |
|
|
Newbie
Joined: 22 Apr 2003
Posts: 3
|
Thanks for the advise. SO can i say based on the errors flag out, is due to the abnormal termination of client that holds on to the memeory resources and not free it up?
FYI,
in solaris OS, u can use the command
# ipcs -a
to show every memory and semaphoers that is assigned to your mq account. ANd you can use the command "ipcrm" to remove all the allocated memory. Of course the queue managers must all be shutdown. |
|
| Back to top |
|
|
| oz1ccg |
| Posted: Wed Apr 23, 2003 3:01 am Post subject: |
|
|
Yatiri
Joined: 10 Feb 2002
Posts: 628
Location: Denmark
|
Yes MQ uses "some" shared memory, to allow comminication between the different processes.
Tnx for the ipcs clue.
Just my $0.02
_________________
Regards, Jørgen
Home of BlockIP2, the last free MQ Security exit ver. 3.00
Cert. on WMQ, WBIMB, SWIFT. |
|
| Back to top |
|
|
| leongor |
| Posted: Thu Apr 24, 2003 5:03 am Post subject: |
|
|
Master
Joined: 13 May 2002
Posts: 264
Location: Israel
|
Have you checked kernel definitions for Sun ?
You can find them in MQSeries Quick Beginnings for Sun Solaris.
It looks like you don't have enough shared memory allocated.
_________________
Regards.
Leonid.
IBM Certified MQSeries Specialist. |
|
| Back to top |
|
|
| jagan |
| Posted: Thu May 01, 2003 5:58 am Post subject: |
|
|
Acolyte
Joined: 26 Jun 2001
Posts: 55
|
| I tried to remove "ipcrm" command for mqm shared memory but failed. Pl give me that exact command for this. I used as mqm user but I logged into it as switch user (su). I dont have root previlages for that box. |
|
| Back to top |
|
|
| vennela |
| Posted: Thu May 01, 2003 6:21 am Post subject: |
|
|
Jedi Knight
Joined: 11 Aug 2002
Posts: 4055
Location: Hyderabad, India
|
You can also use 'amqiclen' command. But I am not sure if this one will solve your problem. If the shared mems are owned by root, then mqm will not have the right to remove them.
-------
Venny |
|
| Back to top |
|
|
|
|
