| Author |
Message
|
| MarcoGruppo |
 Posted: Tue Nov 16, 2004 5:28 am Post subject: Shared queue as Initiation queue |
 |
|
Newbie
Joined: 16 Nov 2004
Posts: 1
|
Hi,
I have any doubt using shared queue as initiation queue with CICS and Trigger Monitor Batch. I have a simple environment with two Qmgr in a Queue Sharing Group; Qmgr QMA resides on server SYSA and QMB on SYSB. I have also three CICS in the same CISCPLEX, CICSA1 and CICSA2 reside on SYSA and are connected with QMA, CICSB is defined on SYSB and connected with QMB.
TEST 1:
I have defined a shared queue as initq with TRIGGERTYPE=FIRST and a process on QMA. Every time a message is putted on, it will start only a transaction alternatively on the three CICS.
TEST 2:
I have defined a process also on QMB. Every time a message is putted on, it will start a transaction alternatively on CICSA1 and CICSA2, and every time on CICSB.
TEST 3:
I have deleted the old two process and I've defined a new one with QSGDISP=GROUP.
Now the situation is worst than above, because every message which is putted on, it may be happen that two transaction could be started on the same CICS.
Please help to manage the situation, I'd like to start only one transaction in only one CICS, but I want to define the process on both Qmgr, so I could have the availability also if one of them will be go down.
Thanks in advance for your support and have a Good day, Marco |
|
| Back to top |
|
 |
| oz1ccg |
| Posted: Tue Nov 16, 2004 2:32 pm Post subject: |
|
|
Yatiri
Joined: 10 Feb 2002
Posts: 628
Location: Denmark
|
You should try using local init.queues, one on each QMGR with the same name... (If I remember right).
The hint is that MQ generates the trigger message on one of the connected qmgrs, depending on ? There was a session on the T&M conf. 2004 on Shared Queues, where we discussed this special topic.
So give it a try.
Just my $0.02 
_________________
Regards, Jørgen
Home of BlockIP2, the last free MQ Security exit ver. 3.00
Cert. on WMQ, WBIMB, SWIFT. |
|
| Back to top |
|
|
| techno |
| Posted: Fri Sep 30, 2005 9:00 am Post subject: |
|
|
Chevalier
Joined: 22 Jan 2003
Posts: 429
|
| Triggering on a queue being accessed by number of applications? Manual says: triggering does not happen if queue is already opened. Why is it so? That means, if an application takes more time to close (say, it happens after the next message arrives and triggering is expected), no triggering takes place and messages do not get processed. Is this correct? |
|
| Back to top |
|
|
| wschutz |
| Posted: Fri Sep 30, 2005 9:12 am Post subject: |
|
|
Jedi Knight
Joined: 02 Jun 2005
Posts: 3316
Location: IBM (retired)
|
|
| Back to top |
|
|
| javagate |
| Posted: Fri Sep 30, 2005 9:25 am Post subject: |
|
|
Disciple
Joined: 15 Nov 2004
Posts: 159
|
Are you using the same initq for more then one CICS region and batch trig monitor?
Makes no sence... Each CICS region and batch trig monitor should have their own initq.
_________________
WebSphere Application Server 7.0 z/OS &
MQ 6.0. I work with WebSphere in the real world not in some IBM lab. |
|
| Back to top |
|
|
| javagate |
| Posted: Fri Sep 30, 2005 9:30 am Post subject: |
|
|
Disciple
Joined: 15 Nov 2004
Posts: 159
|
By the way... I assume when you say 'shared queues' you are on z/os and using DB2 data sharing to manage the MQ objects and using CF structures.
_________________
WebSphere Application Server 7.0 z/OS &
MQ 6.0. I work with WebSphere in the real world not in some IBM lab. |
|
| Back to top |
|
|
| techno |
| Posted: Fri Sep 30, 2005 9:38 am Post subject: |
|
|
Chevalier
Joined: 22 Jan 2003
Posts: 429
|
Regular queue on dist platform. Not the z/OS ones.
If there is any delay in processing the arrived messages, say there are 10 msgs and triggering happens. application keeps processing those messages. Meanwhile, I have another 10 messages, where I expect triggering again. In reality, triggering may not happen, as the application is still proceessing first 10 messages. |
|
| Back to top |
|
|
| jefflowrey |
| Posted: Fri Sep 30, 2005 9:46 am Post subject: |
|
|
Grand Poobah
Joined: 16 Oct 2002
Posts: 19981
|
| techno wrote: |
| Regular queue on dist platform. Not the z/OS ones. |
I've never really considered it good practice to share the same queue between different applications.
| techno wrote: |
| If there is any delay in processing the arrived messages, say there are 10 msgs and triggering happens. application keeps processing those messages. Meanwhile, I have another 10 messages, where I expect triggering again. In reality, triggering may not happen, as the application is still proceessing first 10 messages. |
This is why everyone gets told to always read the queue until it is empty, and only then shut the application down.
If this is done, then your application will keep running until all 20 messages are processed, as well as any more that have come in. Then the queue will be empty, and the app can close the queue, and triggering will occur again.
_________________
I am *not* the model of the modern major general. |
|
| Back to top |
|
|
| wschutz |
| Posted: Fri Sep 30, 2005 10:08 am Post subject: |
|
|
Jedi Knight
Joined: 02 Jun 2005
Posts: 3316
Location: IBM (retired)
|
| Quote: |
Then the queue will be empty, and the app can close the queue, and triggering will occur again.
|
And.... even if the queue isn't empty when you close it, triggering will occur again at MQCLOSE time. I assume we're talking about TRIGTYPE+FIRST) 
_________________
-wayne |
|
| Back to top |
|
|
| javagate |
| Posted: Fri Sep 30, 2005 10:13 am Post subject: |
|
|
Disciple
Joined: 15 Nov 2004
Posts: 159
|
I still think the original post is talking about the mf since they mention CICS, batch trigger monitor and a CICS Plex.
_________________
WebSphere Application Server 7.0 z/OS &
MQ 6.0. I work with WebSphere in the real world not in some IBM lab. |
|
| Back to top |
|
|
| JT |
| Posted: Sun Oct 02, 2005 9:47 am Post subject: |
|
|
Padawan
Joined: 27 Mar 2003
Posts: 1564
Location: Hartford, CT.
|
javagate,
You have to look more closely at the dates of the posts. I learned this some time ago.
This thread has been hijacked. The creator of the thread opened it almost a year ago. |
|
| Back to top |
|
|
|
|
