Author |
Message
|
adborja |
Posted: Wed Jun 01, 2011 12:42 pm Post subject: Help with after MQOutput Blocking message flow |
|
|
Newbie
Joined: 01 Jun 2011 Posts: 4
|
Hi.
I am facing a strange problem sending a message to MQOutput node:
I have a message flow listening on JMS Queue. When a message is put in tha queue, a JMSInput node receives it, and send it to JMS To MQ Transform node, then to compute node for making some treatment, then the message is sent to MQOutput, and finally, the MQOutput out terminal is connected to Output node for leaving the subflow and make some transformations.
The problem is, SOMETIMES the flow continues after the message is sent to MQOutput node, however SOMETIMES the execution of flow stops and the message is not sent to MQOutput out terminal, blocking the flow and getting a timeout error from consumer. It's strange, because this is happenning only in some cases. What could be the reason for this behavior?. |
|
Back to top |
|
 |
mqjeff |
Posted: Wed Jun 01, 2011 12:44 pm Post subject: |
|
|
Grand Master
Joined: 25 Jun 2008 Posts: 17447
|
My first guess would be that the Broker's queue manager is not sized appropriately and is taking a long time to handle or reorganize it's logs to ensure that it can commit the message to the queue properly. |
|
Back to top |
|
 |
lancelotlinc |
Posted: Wed Jun 01, 2011 12:46 pm Post subject: |
|
|
 Jedi Knight
Joined: 22 Mar 2010 Posts: 4941 Location: Bloomington, IL USA
|
Sounds like a Abend condition that only appears when certain criteria are met in the data.
Use interactive debugger to narrow down the place where your code Abends. If you have no ESQL or JCN code, then it probably is in the message set. _________________ http://leanpub.com/IIB_Tips_and_Tricks
Save $20: Coupon Code: MQSERIES_READER |
|
Back to top |
|
 |
mqjeff |
Posted: Wed Jun 01, 2011 12:47 pm Post subject: |
|
|
Grand Master
Joined: 25 Jun 2008 Posts: 17447
|
Hrm. Good point. I had assumed that it was proven that the message actually reached the MQOutput node, and not for example that the flow had reached a RETURN FALSE; statement in a compute node earlier. |
|
Back to top |
|
 |
adborja |
Posted: Wed Jun 01, 2011 2:06 pm Post subject: |
|
|
Newbie
Joined: 01 Jun 2011 Posts: 4
|
Thanks for your feedback!
There is not any RETURN FALSE statement in any nearly compute node. |
|
Back to top |
|
 |
adborja |
Posted: Thu Jun 02, 2011 6:56 am Post subject: |
|
|
Newbie
Joined: 01 Jun 2011 Posts: 4
|
|
Back to top |
|
 |
mqjeff |
Posted: Thu Jun 02, 2011 6:59 am Post subject: |
|
|
Grand Master
Joined: 25 Jun 2008 Posts: 17447
|
Is it proven that the flow is reaching the MQOutput node?
If so, look at the qmgr configuration.
If not, take more time to determine exactly where the flow pauses. |
|
Back to top |
|
 |
adborja |
Posted: Thu Jun 02, 2011 7:14 am Post subject: |
|
|
Newbie
Joined: 01 Jun 2011 Posts: 4
|
Thanks.
Always, the MQOutput is reached, the message is put there in all cases, the problem is after the message is sent to MQOutput, sometimes the flow continue in its output terminal and sometimes does not.
What should be the qmanager configuration?? |
|
Back to top |
|
 |
mqjeff |
Posted: Thu Jun 02, 2011 7:21 am Post subject: |
|
|
Grand Master
Joined: 25 Jun 2008 Posts: 17447
|
adborja wrote: |
Thanks.
Always, the MQOutput is reached, the message is put there in all cases, the problem is after the message is sent to MQOutput, sometimes the flow continue in its output terminal and sometimes does not.
What should be the qmanager configuration?? |
It should be one that does not affect how long it takes messages to be put to queues.
You should *look* at what it is. And *look* at *what* it is doing WHEN the message is stalled.
To see if it *might* be the problem. |
|
Back to top |
|
 |
|