| Author |
Message
|
| thecamel |
 Posted: Wed May 08, 2002 4:15 am Post subject: |
 |
|
Newbie
Joined: 07 May 2002
Posts: 8
Location: UK
|
Help!
I have a problem whereby MQ processes are building up on my box without being closed. I have a .disconnect() method in the finally {} statement whereever MQ in called.
Any help much appreciated!! |
|
| Back to top |
|
 |
| PeterPotkay |
| Posted: Wed May 08, 2002 4:31 am Post subject: |
|
|
Poobah
Joined: 15 May 2001
Posts: 7716
|
2 things:
Before the disconnect, you should specifically close each queue when you are done using it. Do not rely on the disconnect to cleanly close objects.
Make sure you always do the close/disconnect. i.e. if you get an error, make sure the path you get thrown down closes/disconnects.
_________________
Peter Potkay
Keep Calm and MQ On |
|
| Back to top |
|
|
| thecamel |
| Posted: Wed May 08, 2002 4:37 am Post subject: |
|
|
Newbie
Joined: 07 May 2002
Posts: 8
Location: UK
|
Okay,
Yes. I always close the queue first.
Yes. The queue close() and MQManager close() is always done in the finally statement.
As this runs on UNIX, is there any way the local MQ Manager itself could have been configured incorrectly?? We have had this problem since it was restarted. |
|
| Back to top |
|
|
| thecamel |
| Posted: Fri May 10, 2002 12:42 am Post subject: |
|
|
Newbie
Joined: 07 May 2002
Posts: 8
Location: UK
|
| This is still a problem for me. Can anyone help??? |
|
| Back to top |
|
|
| mqonnet |
| Posted: Fri May 10, 2002 3:41 am Post subject: |
|
|
Grand Master
Joined: 18 Feb 2002
Posts: 1114
Location: Boston, Ma, Usa.
|
Also would be useful to check the return codes from MQCLOSE and MQDISC calls. Not quite sure if you are handling the errors for these calls. And as suggested by Peter, check to see if the path followed is appropriate down under after the close call.
Cheers.
Kumar
_________________
IBM Certified WebSphere MQ V5.3 Developer
IBM Certified WebSphere MQ V5.3 Solution Designer
IBM Certified WebSphere MQ V5.3 System Administrator |
|
| Back to top |
|
|
| PeterPotkay |
| Posted: Fri May 10, 2002 4:32 am Post subject: |
|
|
Poobah
Joined: 15 May 2001
Posts: 7716
|
Are these MQ proccesses you refer to Client connections?
_________________
Peter Potkay
Keep Calm and MQ On |
|
| Back to top |
|
|
| thecamel |
| Posted: Fri May 10, 2002 6:38 am Post subject: |
|
|
Newbie
Joined: 07 May 2002
Posts: 8
Location: UK
|
| Yes, they are client connections. |
|
| Back to top |
|
|
| PeterPotkay |
| Posted: Fri May 10, 2002 7:34 am Post subject: |
|
|
Poobah
Joined: 15 May 2001
Posts: 7716
|
What is the heartbeat interval on the SVRCONN channel? What might be happening is that the client apps end abnormally, before they have a chance to close the queue. The queue manager does not realize this and thus the channel (connection) stays up. If you have a heartbeat flowing, say every 30 seconds back and forth, then the first heartbeat after the client app goes away will let the QM side know that the other side is gone.
_________________
Peter Potkay
Keep Calm and MQ On |
|
| Back to top |
|
|
| thecamel |
| Posted: Mon May 13, 2002 1:09 am Post subject: |
|
|
Newbie
Joined: 07 May 2002
Posts: 8
Location: UK
|
| The heartbeat interval is currently set to 300 secs (the default). |
|
| Back to top |
|
|
| PeterPotkay |
| Posted: Mon May 13, 2002 8:22 am Post subject: |
|
|
Poobah
Joined: 15 May 2001
Posts: 7716
|
Well, I'm outta ideas. 
Do these proccesses build up and never go away (i.e. the # just keeps rising)? Or do they eventually go away?
Does restarting the QM get rid of them?
_________________
Peter Potkay
Keep Calm and MQ On |
|
| Back to top |
|
|
| GregJ |
| Posted: Mon May 13, 2002 9:04 am Post subject: |
|
|
Acolyte
Joined: 24 Oct 2001
Posts: 69
Location: Markham, On. Canada
|
I recently had of a problem very similar to this - the svrconn channels kept piling up even though the DISC and CLOSE calls were made.
The problem was due to IIS on the client machine holding the connections. Installing the latest (5.0) patch seemed to help, but the problem is still there.
I have a script to restart the listener every night - that takes care of that. |
|
| Back to top |
|
|
| bduncan |
| Posted: Mon May 13, 2002 10:06 am Post subject: |
|
|
Padawan
Joined: 11 Apr 2001
Posts: 1554
Location: Silicon Valley
|
Did you try comparing the queue handles to MQHO_UNUSABLE_HOBJ after you issued the MQCLOSE? This will tell you without a doubt whether the close succeeded. There is an equivalent constant for disconnects, but I can't remember the name off the top of my head...
_________________
Brandon Duncan
IBM Certified MQSeries Specialist
MQSeries.net forum moderator |
|
| Back to top |
|
|
| thecamel |
| Posted: Tue May 14, 2002 12:01 am Post subject: |
|
|
Newbie
Joined: 07 May 2002
Posts: 8
Location: UK
|
| Would changing the kernel parameters on the unix box which is hosting the local queue manager cause amqcrsta connections to pile up? |
|
| Back to top |
|
|
| mrlinux |
| Posted: Tue May 14, 2002 3:32 am Post subject: |
|
|
Grand Master
Joined: 14 Feb 2002
Posts: 1261
Location: Detroit,MI USA
|
No it would not, it sounds more like application issue with connections.
Have there been recent application changes ????
_________________
Jeff
IBM Certified Developer MQSeries
IBM Certified Specialist MQSeries
IBM Certified Solutions Expert MQSeries |
|
| Back to top |
|
|
| thecamel |
| Posted: Tue May 14, 2002 4:05 am Post subject: |
|
|
Newbie
Joined: 07 May 2002
Posts: 8
Location: UK
|
| Yes, but I can't see where the change would have caused this to occur. I guess I will just have to keep looking. Trial and error and all that... |
|
| Back to top |
|
|
|
|
