| |
|
RSS Feed - WebSphere MQ Support
|
RSS Feed - Message Broker Support
|
  |
|
| False triggering |
« View previous topic :: View next topic » |
| Author |
Message
|
| teebark |
 Posted: Wed Mar 11, 2009 7:03 am Post subject: False triggering |
 |
|
Newbie
Joined: 11 Mar 2009
Posts: 3
|
| I'm in a shop that uses typical MQ triggering of a CICS transaction. The queue is set at FIRST, and the application program reads the queue in a loop until the standard 2003 code is reached. It then reads the queue depth, and if it's greater than a constant value (12), spawns another version of itself. This has worked successfully for years. Then, 4 weeks ago, on Friday the 13th, we started getting serious timeouts throughout the day. We normally process thousands of these transactions daily, and rarely get timeouts. Our systems people found instances where it appears MQ has generated up to 23,000 triggers from this queue, without any data being written to the queue. This causes the CICS queues to back up, and the "real" transactions to timeout. Then, it skipped a Friday, and has come back the last two Fridays. We've checked for outside interference, but we're scratching our heads over what scenario could even be dreamed up to cause this problem. I realize it's a wide open question, but I've googled the problem, and searched your FAQ's, but I haven't found anything related to how this could happen. Any ideas? |
|
| Back to top |
|
 |
| Vitor |
| Posted: Wed Mar 11, 2009 7:34 am Post subject: Re: False triggering |
|
|
Grand High Poobah
Joined: 11 Nov 2005
Posts: 26093
Location: Texas, USA
|
| teebark wrote: |
| I'm in a shop that uses typical MQ triggering of a CICS transaction. The queue is set at FIRST, and the application program reads the queue in a loop until the standard 2003 code is reached. |
You do mean 2033 don't you?
This isn't actually that typical for CICS, where the trigger is EVERY and the transactions are throttelled by CICS. It's certainly not that typical for an application to check queue depth and start more copies of itself.
It sounds a lot (especially if you really mean 2003) like there's an application issue that's causing message backout very, very quickly and the trigger conditions to be met.
_________________
Honesty is the best policy.
Insanity is the best defence. |
|
| Back to top |
|
|
| teebark |
| Posted: Wed Mar 11, 2009 7:41 am Post subject: |
|
|
Newbie
Joined: 11 Mar 2009
Posts: 3
|
| Yes, that was just a typo. Are you referring to CICS backouts? That actually occurred to me--perhaps the CICS timeouts are cause the MQ false triggering. In which case we're looking at the problem backwards. |
|
| Back to top |
|
|
| Vitor |
| Posted: Wed Mar 11, 2009 8:15 am Post subject: |
|
|
Grand High Poobah
Joined: 11 Nov 2005
Posts: 26093
Location: Texas, USA
|
| teebark wrote: |
| Yes, that was just a typo. Are you referring to CICS backouts? That actually occurred to me--perhaps the CICS timeouts are cause the MQ false triggering. In which case we're looking at the problem backwards. |
_________________
Honesty is the best policy.
Insanity is the best defence. |
|
| Back to top |
|
|
| PeterPotkay |
| Posted: Thu Mar 12, 2009 7:47 pm Post subject: |
|
|
Poobah
Joined: 15 May 2001
Posts: 7722
|
If your CICS transaction is getting triggered, opening the q, and dying before the message is 100% consumed, and that q is then closed with more than 0 messages on it, you get retriggered.
If the 1st message on the q is the cause of the CICS transaction dying, this will happen again. And again. And again. And again. And again. (Am I up to 23,000 yet?) And again. And again....
_________________
Peter Potkay
Keep Calm and MQ On |
|
| Back to top |
|
|
| teebark |
| Posted: Mon Mar 16, 2009 4:17 am Post subject: False triggers |
|
|
Newbie
Joined: 11 Mar 2009
Posts: 3
|
| That doesn't appear to be the case--no abends occurred. We changed the queue option to EVERY, and we got through last Friday okay. We may never know what caused the problem, but it it reoccurs, we'll take another look at this possibility. The graph of the day's transactions vs. logged transactions showed a pretty clear picture. The two curves were identical, except for a huge spike where the phantom transactions overlaid the logged transactions. |
|
| Back to top |
|
|
| PeterPotkay |
| Posted: Mon Mar 16, 2009 4:26 am Post subject: |
|
|
Poobah
Joined: 15 May 2001
Posts: 7722
|
If your CICS transaction is getting triggered, opening the q, and dying....
-or-
If your CICS transaction is getting triggered, opening the q, and ending for some reason without abending...
But glad you got a workaround in place....
_________________
Peter Potkay
Keep Calm and MQ On |
|
| Back to top |
|
|
|
|
|
| |
|
Page 1 of 1 |
|
You cannot post new topics in this forum
You cannot reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum
You cannot vote in polls in this forum
|
|
|
|
