| Author |
Message
|
| ashishsuvarna |
 Posted: Sun Dec 25, 2005 1:56 am Post subject: FDC's being created continously |
 |
|
Apprentice
Joined: 25 Dec 2005
Posts: 30
Location: INDIA
|
Hi,
We are facing problems in one of our servers with FDC's being created continously .The server has MQSERIES 5.3 CSD 05 on it. It has about 750 queues (mostly transmit queues) and about a 1000 channels. The FDC'S are mainly created by processes amqrmppa and amqzlaa0 and I can see them doing a lot of CPU and memory utilization. Over 700 queues are using the process amqrmppa ( checked queue usage on MO71). Almost all FDC'S indicate
AMQ6119: An internal WebSphere MQ error has occurred (Failed to attach shared
memory map: Handle(000004C8) [rc=536895769 errorCode=8])
We had this problem earlier, but then recycling MQSERIES services fixed the problem. Can you please suggest some problem area we might need to look into.
Regards,
Ashish A Suvarna
_________________
Young techie |
|
| Back to top |
|
 |
| gunter |
| Posted: Tue Dec 27, 2005 6:20 am Post subject: |
|
|
Partisan
Joined: 21 Jan 2004
Posts: 307
Location: Germany, Frankfurt
|
|
| Back to top |
|
|
| ashishsuvarna |
| Posted: Tue Dec 27, 2005 6:32 am Post subject: |
|
|
Apprentice
Joined: 25 Dec 2005
Posts: 30
Location: INDIA
|
Yeah, I came across that topic. I don't think my issue is the same as that. The problem I face is mainly due to FDC's created by amq processes. The FDC's throw out
AMQ6119: An internal WebSphere MQ error has occurred (Failed to attach shared memory map
Just yesterday I tried to start one channel on the same server and it threw me the same FDC and the channel went into retrying mode. After about 2 hours I was able to start the channel.
I deffinitely think it is an issue with memory but need to know which areas I can check for memory leak or something which is leading to this problem
_________________
Young techie |
|
| Back to top |
|
|
| jefflowrey |
| Posted: Tue Dec 27, 2005 6:44 am Post subject: |
|
|
Grand Poobah
Joined: 16 Oct 2002
Posts: 19981
|
What platform?
If a unix platform, did you check your kernel parameters against the values listed in the quick beginnings guide for your Unix?
_________________
I am *not* the model of the modern major general. |
|
| Back to top |
|
|
| ashishsuvarna |
| Posted: Tue Dec 27, 2005 6:49 am Post subject: |
|
|
Apprentice
Joined: 25 Dec 2005
Posts: 30
Location: INDIA
|
Thanx for a quick response.
The environment is WIN 2000 and it is a Queue Manager in a Microsoft clustered environment.
_________________
Young techie |
|
| Back to top |
|
|
| hguapluas |
| Posted: Tue Dec 27, 2005 2:37 pm Post subject: |
|
|
Centurion
Joined: 05 Aug 2004
Posts: 105
Location: San Diego
|
Check your available free memory in the system(s) as well as check your free disk space where you are storing your messages. Are you using circular or linear logs?
Is there a pattern to your errors. Does it occur randomly, when users TS in remotely, login locally, at a particular time of day/night?
Also check your permissions for "Create Global Objects" (primarily associated with Terminal Services sessions). Have you or your network folks instituted a new set of NW policies recently? If so, they may have done something to the permissions needed for MQ to work properly. (Thought it is not likely to be this particular problem)
There must be other messages in your error logs that may be able to provide clues. Check carefully.
Cheers, |
|
| Back to top |
|
|
| ashishsuvarna |
| Posted: Fri Dec 30, 2005 2:04 am Post subject: |
|
|
Apprentice
Joined: 25 Dec 2005
Posts: 30
Location: INDIA
|
We are using cirular logging.
The only pattern that I am able to see is FDC's are created in a bunch. They are mainly created by amqzla0 process. Two days back 8 FDC's were generated with AMQ6118 : internal websphere MQ error by processs amqzla0 and 1 FDC with AMQ6119 : An internal WebSphere MQ error has occurred (Failed to attach shared memory map) by the process aqrmppa.
Free disk space is more than 18 GB so I do not think this should be a problem.
The virtual memory for the server is 7049188KB. I strongly believe that increasing virtual memory could help fix this problem What do you suggest ? Also if possible please suggest an appropriate way to increase the virtual memory.
_________________
Young techie |
|
| Back to top |
|
|
| fjb_saper |
| Posted: Fri Dec 30, 2005 9:30 am Post subject: |
|
|
Grand High Poobah
Joined: 18 Nov 2003
Posts: 20756
Location: LI,NY
|
| ashishsuvarna wrote: |
We are using cirular logging.
The only pattern that I am able to see is FDC's are created in a bunch. They are mainly created by amqzla0 process. Two days back 8 FDC's were generated with AMQ6118 : internal websphere MQ error by processs amqzla0 and 1 FDC with AMQ6119 : An internal WebSphere MQ error has occurred (Failed to attach shared memory map) by the process aqrmppa.
Free disk space is more than 18 GB so I do not think this should be a problem.
The virtual memory for the server is 7049188KB. I strongly believe that increasing virtual memory could help fix this problem What do you suggest ? Also if possible please suggest an appropriate way to increase the virtual memory. |
Did you check whether there was an APAR for AMQ6118/6119 ? |
|
| Back to top |
|
|
| ashishsuvarna |
| Posted: Fri Dec 30, 2005 8:29 pm Post subject: |
|
|
Apprentice
Joined: 25 Dec 2005
Posts: 30
Location: INDIA
|
I do not understand what an APAR is and where will it be generated. 
_________________
Young techie |
|
| Back to top |
|
|
| fjb_saper |
| Posted: Fri Dec 30, 2005 8:49 pm Post subject: |
|
|
Grand High Poobah
Joined: 18 Nov 2003
Posts: 20756
Location: LI,NY
|
An APAR is a fix that has been released between Fix packs (CSD). It will fix the problem and get incorporated into future fixpacks..
Enjoy  |
|
| Back to top |
|
|
| ashishsuvarna |
| Posted: Fri Dec 30, 2005 9:15 pm Post subject: |
|
|
Apprentice
Joined: 25 Dec 2005
Posts: 30
Location: INDIA
|
Thanks for such a quick response.
I do not find any APAR' s released for AMQ 6118/6119. Can someone suggest me any other solution ?
_________________
Young techie |
|
| Back to top |
|
|
| fjb_saper |
| Posted: Sat Dec 31, 2005 10:06 am Post subject: |
|
|
Grand High Poobah
Joined: 18 Nov 2003
Posts: 20756
Location: LI,NY
|
Check the IBM website with the probe id in the FDC. you will most likely find that this has been solved in a subsequent CSD.
Can you move to CSD 11 ?
Enjoy  |
|
| Back to top |
|
|
|
|
