Author |
Message
|
ms_mq |
Posted: Tue Sep 18, 2007 1:36 am Post subject: MQ triggering problem |
|
|
Acolyte
Joined: 23 Feb 2007 Posts: 69
|
Hi,
Please help me out on this mq triggering problem.
This is on a windows 2000 server and MQ version is 6.0.2.0 .When messages comes to a local queue, it triggers an application exe which processes the messages from the queue. Trigger type is FIRST.
On this instance, the MQ triggering was showing as ON, but the EXE was not running and hence the messages were not getting processed, but they were getting piled up.
There were no error messages either at MQ end or at application side.Also,the triggering was not OFF, it was showing ON only. The issue was finally resovled when we trigger OFF and made it ON again.SO, iam confused what would have caused the problem.Any thoughts??
Please advise. _________________ Regards,
MS |
|
Back to top |
|
 |
Vitor |
Posted: Tue Sep 18, 2007 1:52 am Post subject: |
|
|
 Grand High Poobah
Joined: 11 Nov 2005 Posts: 26093 Location: Texas, USA
|
At the risk of being obvious, not all the conditions for triggering were met.
Was there a problem with the trigger monitor at the time in question? Can you duplicate the problem by reading the queue to destruction, stopping the app then adding another message? _________________ Honesty is the best policy.
Insanity is the best defence. |
|
Back to top |
|
 |
ms_mq |
Posted: Tue Sep 18, 2007 2:03 am Post subject: |
|
|
Acolyte
Joined: 23 Feb 2007 Posts: 69
|
Hi Vitor,
Thanks for your reply.
I checked the AMQ error logs and i could not see any errors related to triggering/trigger monitor . so i assume there were no problems with the TMC. Also, i will not be able to reproduce this problem as it is running in production.
Please advise. _________________ Regards,
MS |
|
Back to top |
|
 |
Vitor |
Posted: Tue Sep 18, 2007 2:23 am Post subject: |
|
|
 Grand High Poobah
Joined: 11 Nov 2005 Posts: 26093 Location: Texas, USA
|
Trigger monitor runs as a separate process & doesn't log to the queue manager queue. It sounds like it wasn't running for a period; if this was production is it possible it crashed & some monitoring software / process restarted it? _________________ Honesty is the best policy.
Insanity is the best defence. |
|
Back to top |
|
 |
ms_mq |
Posted: Tue Sep 18, 2007 2:52 am Post subject: |
|
|
Acolyte
Joined: 23 Feb 2007 Posts: 69
|
There is no possibility for some other program to restart the TM as we do not use any tools. In that case , only we should have started it manually , which did not happen.Also, there was not message in the Initiation queue or DLQ.
I guess there might have been a scheduling problem between the Application exe and the trigger monitor. Is that possible? _________________ Regards,
MS |
|
Back to top |
|
 |
Vitor |
Posted: Tue Sep 18, 2007 2:56 am Post subject: |
|
|
 Grand High Poobah
Joined: 11 Nov 2005 Posts: 26093 Location: Texas, USA
|
ms_mq wrote: |
I guess there might have been a scheduling problem between the Application exe and the trigger monitor. Is that possible? |
It's possible that the trigger monitor tried to start the application & it failed for 1-n reasons. You'd know more about how or if that would be logged in your system than me; obviously if the application never actually started then the application log would be clear as you indicated.
There's certainly no scheduling in terms of the message or the monitor. Trigger conditions are met -> trigger message sent -> trigger monitor receives message -> process definition invoked. _________________ Honesty is the best policy.
Insanity is the best defence. |
|
Back to top |
|
 |
Satsuma |
Posted: Wed Sep 19, 2007 2:11 am Post subject: |
|
|
Apprentice
Joined: 30 Apr 2007 Posts: 25 Location: Finland
|
Hi,
I've had lots of same kind of problems with TM. It's always got something to do with missing permissions (eg. cmd.exe, folders..).
Check event log and and all the files etc you are using. |
|
Back to top |
|
 |
afroz_online777 |
Posted: Tue Sep 25, 2007 10:36 pm Post subject: Are sure Trigger Monitory is Runnring |
|
|
Novice
Joined: 07 May 2007 Posts: 21
|
Hi ,
Are sure that the Trigger Monitory is Runnring ,
have you started teh trigger monitory using the commond
runmqtrm -m <MQgrm Name> -q <Init_Queue>
Just confirm ,
Bcz i too face the same problem , once, when i started the trigger monitor the problem was solved
Thank
Zorfa (Afroz) |
|
Back to top |
|
 |
fjb_saper |
Posted: Wed Sep 26, 2007 2:33 am Post subject: |
|
|
 Grand High Poobah
Joined: 18 Nov 2003 Posts: 20756 Location: LI,NY
|
How long is your wait time to get the first message?
Remember that your application gets triggered when the queue depth moves from 0 to 1. This does not mean that the message is committed or in a get able state. So your application starts does a read nowait and gets a 2033 and exits. Now your qdepth is 1. Other messages come and triggering conditions will not be satisfied for 11 days (default value for trigint on the qmgr).
Working as designed ...
Solution:
- reduce the qmgr's trigint value to something you can live with (5 mins=300,000)
Remember it really takes another message to hit the queue for it to trigger again after the trigint is expired.
- Do the get with wait. Set a reasonable time to allow for the put application to finish its UOW.
Enjoy  _________________ MQ & Broker admin |
|
Back to top |
|
 |
Nigelg |
Posted: Wed Sep 26, 2007 2:43 am Post subject: |
|
|
Grand Master
Joined: 02 Aug 2004 Posts: 1046
|
In the above case a trigger msg will be generated when the app closes the queue.
See condition 12
Quote: |
The only application serving a queue issues an MQCLOSE call, for a TriggerType of MQTT_FIRST or MQTT_DEPTH, and there is at least:
One (MQTT_FIRST), or
TriggerDepth (MQTT_DEPTH)
messages on the queue of sufficient priority (condition 2), and conditions 6 through 10 are also satisfied. |
_________________ MQSeries.net helps those who help themselves.. |
|
Back to top |
|
 |
fjb_saper |
Posted: Wed Sep 26, 2007 3:09 am Post subject: |
|
|
 Grand High Poobah
Joined: 18 Nov 2003 Posts: 20756 Location: LI,NY
|
I'd have to check for conditions 6 through 10...
I remember distinctively the manual specifying you should do a wait for specifically such a condition... Can't remember though if it was for the condition happening at the beginning or the end of the queue being serviced.  _________________ MQ & Broker admin |
|
Back to top |
|
 |
bruce2359 |
Posted: Wed Sep 26, 2007 6:25 am Post subject: |
|
|
Guest
|
Check the dead-letter queue for trigger messages. |
|
Back to top |
|
 |
UglyAngelX |
Posted: Wed Sep 26, 2007 11:22 am Post subject: |
|
|
Voyager
Joined: 04 Dec 2001 Posts: 90 Location: BEARS FAN IN STEELER COUNTRY
|
6.0.2.2 is supposed to have a fix for Triggering issues. We applied it to a server that was having issues and it did seem to help the situation. |
|
Back to top |
|
 |
Nigelg |
Posted: Wed Sep 26, 2007 7:21 pm Post subject: |
|
|
Grand Master
Joined: 02 Aug 2004 Posts: 1046
|
The only triggering fix in 6.0.2.2 is IY93155, triggering a channel where the xmitq does not have TRIGDATA (the channel name) set.
99.99% of so-called triggering issues are where the trigger conditions
have not been met.
Quote: |
We applied it to a server that was having issues and it did seem to help the situation. |
Rather a wishy-washy statement. Did it fix it, or did it not?
The most likely reason that 'it seemed to help' is that the environment has now changed so that the trigger conditions are more often fulfilled. _________________ MQSeries.net helps those who help themselves.. |
|
Back to top |
|
 |
UglyAngelX |
Posted: Thu Sep 27, 2007 3:44 am Post subject: |
|
|
Voyager
Joined: 04 Dec 2001 Posts: 90 Location: BEARS FAN IN STEELER COUNTRY
|
I guess if the situation was not helped by applying the fix pack, I would not even mention it...........so you got it......fixed! |
|
Back to top |
|
 |
|