| |
|
RSS Feed - WebSphere MQ Support
|
RSS Feed - Message Broker Support
|
  |
|
| MQ-triggered Transaction with backout queue???????? |
« View previous topic :: View next topic » |
| Author |
Message
|
| sumansengupta07 |
 Posted: Thu May 22, 2014 9:02 am Post subject: MQ-triggered Transaction with backout queue???????? |
 |
|
Novice
Joined: 14 Nov 2013
Posts: 23
|
If any MQ-triggered Transaction got abended 3000 times for same reason and suppose there is a corresponding backout queue with backout threshold 3, then after abending 3000 times the depth of the backout queue should be 1000.
I need all of yours valuable input on this and of course correct me if I my logic is wrong.
 |
|
| Back to top |
|
 |
| fjb_saper |
| Posted: Thu May 22, 2014 9:06 am Post subject: Re: MQ-triggered Transaction with backout queue???????? |
|
|
Grand High Poobah
Joined: 18 Nov 2003
Posts: 20756
Location: LI,NY
|
| sumansengupta07 wrote: |
If any MQ-triggered Transaction got abended 3000 times for same reason and suppose there is a corresponding backout queue with backout threshold 3, then after abending 3000 times the depth of the backout queue should be 1000.
I need all of yours valuable input on this and of course correct me if I my logic is wrong.
|
Only if your process / programs actually puts the messages exceeding the threshold to the backout queue.
JMS / XMS does that for you. Any other program,you'd have to program this behavior... 
_________________
MQ & Broker admin |
|
| Back to top |
|
|
| PeterPotkay |
| Posted: Thu May 22, 2014 9:31 am Post subject: |
|
|
Poobah
Joined: 15 May 2001
Posts: 7722
|
And only if you got the original messages under syncpoint.
And only if you put the messages successfully to the backout queue.
Some, none, or all of the messages could have expired off the backout queue.
Some, none, or all of the messages could have been consumed off the backout queue.
There could have been 0,1 or more messages on the backout queue at the start of it.
In other words, the depth of that backout queue could be anything, regardless of how many times the app abended.
_________________
Peter Potkay
Keep Calm and MQ On |
|
| Back to top |
|
|
| exerk |
| Posted: Thu May 22, 2014 2:33 pm Post subject: Re: MQ-triggered Transaction with backout queue???????? |
|
|
Jedi Council
Joined: 02 Nov 2006
Posts: 6339
|
| sumansengupta07 wrote: |
| If any MQ-triggered Transaction got abended 3000 times for same reason and suppose there is a corresponding backout queue with backout threshold 3, then after abending 3000 times the depth of the backout queue should be 1000. |
I'm going to break this down...
"...If any MQ-triggered Transaction got abended 3000 times for same reason and suppose there is a corresponding backout queue with backout threshold 3..."
The back-out queue will not have a threshold of 3, that will be set on the triggered queue.
"...then after abending 3000 times the depth of the backout queue should be 1000...."
As PeterPotkay has pointed out, the queue depth could be zero to n in depth. Your question assumes the "Transaction" got as far as connecting to the queue manager,opened the queue, and got a message - if it didn't get that far then the back-out queue (assuming it was at zero depth to begin with) will have zero depth.
I would further suggest that if your application abends 3000 times there is a serious problem with the application, e.g. it cannot handle poison messages and the developer should therefore suffer a long, lingering, and painful death, or it should never have been deployed in the first place, in which case all involved should suffer the same fate as the aforementioned developer...
_________________
It's puzzling, I don't think I've ever seen anything quite like this before...and it's hard to soar like an eagle when you're surrounded by turkeys. |
|
| Back to top |
|
|
| fjb_saper |
| Posted: Thu May 22, 2014 3:46 pm Post subject: |
|
|
Grand High Poobah
Joined: 18 Nov 2003
Posts: 20756
Location: LI,NY
|
Or your transaction should get inhibited because a needed resource (DB) is not available...
_________________
MQ & Broker admin |
|
| Back to top |
|
|
| MQsysprog |
| Posted: Thu May 22, 2014 11:54 pm Post subject: |
|
|
Centurion
Joined: 24 Feb 2014
Posts: 116
|
Well, firstly the type of question sound sligtly like an examination one,but
the things to consider carefully are several, I would put at the beginning of the abend process a trigger stop to avoid the large numbers ...
The abend processing is a heavy function ,in transaction processing
so is better to avoid this type of behavior, expecially with TM like Cics or Ims .
Then you have a second chance using the bridging functions available in IMS and Cics ,here i am assuming you are on Z/os enviroment of course ...
The advantage in this case ,could be a direct start of the transactions and possibly a fast materialization of some type of abend in the TM side ,determined by the payload of the messages or other conditions for example resources unavailable ,as said previously by my proficent colleagues. |
|
| Back to top |
|
|
| bob_buxton |
| Posted: Fri May 23, 2014 2:31 am Post subject: |
|
|
Master
Joined: 23 Aug 2001
Posts: 266
Location: England
|
MQGMO_MARK_SKIP_BACKOUT can be a very useful attribute in this type of scenario allowing you to put the 'poison' message to the Backout queue in your transactions abend handler routine.
If you don't have an abend handler you can create an infinite loop where a transaction is triggered, abends, message is rolled back, transaction is triggered again abends ....
_________________
Bob Buxton
Ex-Websphere MQ Development |
|
| 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
|
|
|
|
