| Author |
Message
|
| praveenchhangani |
 Posted: Thu Aug 28, 2003 8:34 am Post subject: SYSTEM.DEAD.HOLD QUEUE |
 |
|
Disciple
Joined: 20 Sep 2002
Posts: 192
Location: Chicago, IL
|
Folks,
I have recently started noticing that every so often we have a few messages on the SYSTEM.DEAD.HOLD Queue. Any idea what this queue is used for and what it means to have messages appear in it/what can be done to fix it?
Workflow environment:
AIX 5.1
Workflow version 3.3.2
MQ5.3
Thanks,
Praveen
_________________
Praveen K. Chhangani,
IBM Certified Solutions Designer -
MQ Workflow 3.4. |
|
| Back to top |
|
 |
| vennela |
| Posted: Thu Aug 28, 2003 9:20 am Post subject: |
|
|
Jedi Knight
Joined: 11 Aug 2002
Posts: 4055
Location: Hyderabad, India
|
You can use admin utility (fmcautil) to browse those messages.
Select the execution server menu from the main menu and from there it will take you to the hold queue menu.
Check the admin guide for more details on what sort of messages this queue has.
-------
Venny |
|
| Back to top |
|
|
| amittalekar |
| Posted: Thu Aug 28, 2003 9:24 am Post subject: |
|
|
Disciple
Joined: 03 Apr 2002
Posts: 166
Location: VA, USA
|
I have faced this problem... If for some reason database dead lock happens then WF server puts the messages in hold queue and they will be processed after some time ....
I am using UPES implemention... in my case I used to query db2 externally for audit trail records... At the same time if WF tries to write some data in audit trail , it gets lock and that was deadlock situation..
you can check ur system and error log messages... u will definately find something if messages are going in HOLD queue.
go though "messages and codes" pdf and search for hold queue , u will find lot of conditions there.... |
|
| Back to top |
|
|
| praveenchhangani |
| Posted: Thu Aug 28, 2003 10:24 am Post subject: |
|
|
Disciple
Joined: 20 Sep 2002
Posts: 192
Location: Chicago, IL
|
I guess it helps if I actually give you the correct queue name! 
I actually meant SYSTEM.DEAD.HOLDING(That's okay though because all the people that answered this queue got the idea and what queue I was really referring to), queue now that I look through once more...
I am going to examine the message the next time this happens and take it from there, but i believe it has something to do with the dead letter handler process and it not being able to resend the message to where it belongs after a certain number of tries; afterwhich it puts it on the SYSTEM.DEAD.HOLDING queue.
I'll let you guys know if I find something...
_________________
Praveen K. Chhangani,
IBM Certified Solutions Designer -
MQ Workflow 3.4. |
|
| Back to top |
|
|
| Ratan |
| Posted: Thu Aug 28, 2003 11:55 am Post subject: |
|
|
Grand Master
Joined: 18 Jul 2002
Posts: 1245
|
Is this queue only specific to Unix environments, because I dont see it on Windows environments?
_________________
-Ratan |
|
| Back to top |
|
|
| vennela |
| Posted: Thu Aug 28, 2003 1:02 pm Post subject: |
|
|
Jedi Knight
Joined: 11 Aug 2002
Posts: 4055
Location: Hyderabad, India
|
I can't find it on any of the UNIX platforms too.
I have searched through WMQ and WMQWF manuals and I can't find where this queue is mentioned. It is definitely not one of the SYSTEM onjects created by MQ or Workflow.
Praveen:
Are you sure this queue is not one of the queues that you created (though the name is misleading).
The one thing that comes to mind is if it is a temporary dynamic queue or something of that sort.
When I initially saw the post I didn't pay attention to what the queue name was. I just remembered that there is some sort of that utility available in the admin utility. |
|
| Back to top |
|
|
| praveenchhangani |
| Posted: Fri Aug 29, 2003 5:28 am Post subject: |
|
|
Disciple
Joined: 20 Sep 2002
Posts: 192
Location: Chicago, IL
|
Ok here's the scoop. Sorry I was unware of this, until recently.
The SYSTEM.DEAD.HOLDING is actually a queue that was defined by us. In a typical scenario what happens is that if the destination queue is full or unavailable for some reason, MQSeries decides to put the message(s) on the Dead letter queue(and by the way forgets about it/them  ).
What we did a while ago, was to define a deadletter handler process, whereby the message is resent to the destination queue for a number of tries for a particular time.
This way, even if messages don't get to their destination immeadiately, we don't have to worry about where the messages went or anything of that sort.
I remember reading somewhere about a "hold- I believe it has something to do with the admin util" queue somewhere in workflow(not sure where), and for some reason, decided that the SYSTEM.DEAD.HOLDING queue was a workflow queue.
Anyways, I figured out what the problem was. There was a workflow application that couldn't be started on the AIX machine during that time range, and that is what was causing us to get those messages on the queue(in addition to the fact that the destination queue was full/unavailable - but that's another issue in itself  )
_________________
Praveen K. Chhangani,
IBM Certified Solutions Designer -
MQ Workflow 3.4. |
|
| Back to top |
|
|
|
|
