| Author |
Message
|
| rahuldhanpal |
 Posted: Fri Jan 07, 2011 10:23 am Post subject: question |
 |
|
Voyager
Joined: 24 Jan 2009
Posts: 84
Location: Kenosha WI
|
Guys,
We have a below scenario...
We have A(Full repository),B (Full repository),C(Partial repository), queuemanagers in a MQ cluster.
B and C have same name queues say “F” created physically on both these queue managers and both are in cluster.
Senario:
We have Application responding back for a message received from either B Broker qmgr or C broker qmgr and this application puts to a clustered queue ”F” by connecting to “A” queue manager. This message contains Reply-to-Qmgr and reply-to queue set in the header.
Question : Should this message when put on A gets to the exact queue manager in this if B requested then go to B if C requested then go to C(point is the message should end up in the queue "F" but of which qmgr is it "B" or "C")
any thoughts highly appreciated!!
Thanks
Rahul. |
|
| Back to top |
|
 |
| Vitor |
| Posted: Fri Jan 07, 2011 10:30 am Post subject: |
|
|
Grand High Poobah
Joined: 11 Nov 2005
Posts: 26093
Location: Texas, USA
|
If the message is truly a reply, then the putting application should use the ReplyToQmgr as well as the ReplyToQ to specify where the reply should be sent.
If it doesn't but just puts it to the clustered queue F "because it knows that's the reply queue" then it will go to B, C, or both according to the normal rules of clustering.
_________________
Honesty is the best policy.
Insanity is the best defence. |
|
| Back to top |
|
|
| run |
| Posted: Fri Jan 07, 2011 10:49 am Post subject: question |
|
|
Apprentice
Joined: 14 May 2007
Posts: 45
Location: Newyork, USA
|
Hi Vitor,
"If the message is truly a reply, then the putting application should use the ReplyToQmgr as well as the ReplyToQ to specify where the reply should be sent. "
Means if the message contains ReplyToQmgr as well as the ReplyToQ and Msg type as Reply if put to the clustered queue then it on goes to that ReplyToQmgr queue.
"If it doesn't but just puts it to the clustered queue F "because it knows that's the reply queue" then it will go to B, C, or both according to the normal rules of clustering"
Say using RFHUTIL if we set the "ReplyToQmgr as well as the ReplyToQ" and put to a response cluster queue then this should not work right as it follows the cluster rules? |
|
| Back to top |
|
|
| mqjeff |
| Posted: Fri Jan 07, 2011 10:51 am Post subject: |
|
|
Grand Master
Joined: 25 Jun 2008
Posts: 17447
|
The reply should go back to the queue that the application that is expecting it on.
Whether or not you specify a ReplyToQmgr or not has nothing to do with what type of channel will move a message to that queue manager. |
|
| Back to top |
|
|
| Vitor |
| Posted: Fri Jan 07, 2011 10:57 am Post subject: Re: question |
|
|
Grand High Poobah
Joined: 11 Nov 2005
Posts: 26093
Location: Texas, USA
|
| run wrote: |
Hi Vitor,
"If the message is truly a reply, then the putting application should use the ReplyToQmgr as well as the ReplyToQ to specify where the reply should be sent. "
Means if the message contains ReplyToQmgr as well as the ReplyToQ and Msg type as Reply if put to the clustered queue then it on goes to that ReplyToQmgr queue. |
No, it means that the application should use those values. If you just do a put to the queue, the software won't automagically fix it for you.
| run wrote: |
"If it doesn't but just puts it to the clustered queue F "because it knows that's the reply queue" then it will go to B, C, or both according to the normal rules of clustering"
Say using RFHUTIL if we set the "ReplyToQmgr as well as the ReplyToQ" and put to a response cluster queue then this should not work right as it follows the cluster rules? |
There's no such thing as a "response cluster queue". There's just a queue, which may or may not belong to a cluster, and the options you use on the put. If you specify a desination queue manager as well as a destination queue then this is used as part of the name resolutiion. If you just use a queue name, or a queue name and a cluster alias, these scenarios work differently according to the cluster rules.
Do you think there's a "request cluster queue" type?
_________________
Honesty is the best policy.
Insanity is the best defence. |
|
| Back to top |
|
|
| bruce2359 |
| Posted: Fri Jan 07, 2011 11:01 am Post subject: |
|
|
Poobah
Joined: 05 Jan 2008
Posts: 9469
Location: US: west coast, almost. Otherwise, enroute.
|
The app that creates the request message gets to tell in the replying app where the reply message should be sent.
However, the replying app can ignore the ReplyToQueue and ReplyToQMgr fields in the inbound MQMD, and send the reply message anywhere it chooses - to a differnt qmgr and/or a different queue.
If the replying app mqopens (or mqput1's) a cluster queue name, then the reply message will go to an instance of the replytoqueue according to the clustering algorithm as documented.
_________________
I like deadlines. I like to wave as they pass by.
ב''ה
Lex Orandi, Lex Credendi, Lex Vivendi. As we Worship, So we Believe, So we Live. |
|
| Back to top |
|
|
| run |
| Posted: Fri Jan 07, 2011 11:48 am Post subject: question |
|
|
Apprentice
Joined: 14 May 2007
Posts: 45
Location: Newyork, USA
|
Hi mqjeff,
Does it mean if the response is put by the application to a queue that is clustered which contains "ReplyToQueue and ReplyToQMgr fields set" end up in the awaiting broker msg flow(in this case the application waiting for the message) even though there are two queue names with the same name in the cluster which are physically defined on both queue managers right?
Vitor,
No there is no such thing as "response cluster queue" neither "request cluster queue" what I meant was response to the cluster queue or request to the cluster queue.  |
|
| Back to top |
|
|
| Vitor |
| Posted: Fri Jan 07, 2011 11:56 am Post subject: Re: question |
|
|
Grand High Poobah
Joined: 11 Nov 2005
Posts: 26093
Location: Texas, USA
|
| run wrote: |
| Does it mean if the response is put by the application to a queue that is clustered which contains "ReplyToQueue and ReplyToQMgr fields set" end up in the awaiting broker msg flow(in this case the application waiting for the message) even though there are two queue names with the same name in the cluster which are physically defined on both queue managers right? |
When did broker enter the picture?
I repeat what I and others have said before. It doesn't matter what those fields are set to, it matters if the putting application chooses to use them when it puts the response. The response message, like all messages, will go where the putting application addresses it. If that's the values in the MQMD, it will go there. If it's not, it won't.
_________________
Honesty is the best policy.
Insanity is the best defence. |
|
| Back to top |
|
|
| Vitor |
| Posted: Fri Jan 07, 2011 11:59 am Post subject: Re: question |
|
|
Grand High Poobah
Joined: 11 Nov 2005
Posts: 26093
Location: Texas, USA
|
| run wrote: |
| which contains "ReplyToQueue and ReplyToQMgr fields set" |
Perhaps this is clearer:
These reply fields are not fields like priority or persistence. The setting of these fields will control how the software treats a message in terms of it's persistence or priority. Setting (or not setting) the reply fields has no effect on the software in that sense.
_________________
Honesty is the best policy.
Insanity is the best defence. |
|
| Back to top |
|
|
| bruce2359 |
| Posted: Fri Jan 07, 2011 12:07 pm Post subject: |
|
|
Poobah
Joined: 05 Jan 2008
Posts: 9469
Location: US: west coast, almost. Otherwise, enroute.
|
The usual request-reply model presumes that the request message dictates that the reply message should be sent (returned) to the requesting application, which is waiting on the reply-to-queue for the reply to arrive.
Most often, reply-to-queues are created dynamically by the requesting app for the reply message(s) it expects.
There is a modified request-reply model, that presumes that the reply goes to an alternate (different) queue - possibly on an alternate qmgr.
Your app design dictates which request-reply model (if any) is being used.
_________________
I like deadlines. I like to wave as they pass by.
ב''ה
Lex Orandi, Lex Credendi, Lex Vivendi. As we Worship, So we Believe, So we Live. |
|
| Back to top |
|
|
| fatherjack |
| Posted: Fri Jan 07, 2011 2:20 pm Post subject: |
|
|
Knight
Joined: 14 Apr 2010
Posts: 522
Location: Craggy Island
|
God! This is basic MQ for heavens sake. It should be simple. MQ will send the message to where the app tells it to. So if it tells it to go to a clustered queue that's where it'll go driven by the various cluster workload options. If it tells it to go to a particular queue manager thats where it will go. If it tells it to go to a queue manager that is defined as a queue manager alias the resultant destination will be dictated by the QMA definition. If MQ encounters any problems resolving the destinations it might go to a DLQ somewhere.
OK, so maybe not quite that simple 
_________________
Never let the facts get in the way of a good theory. |
|
| Back to top |
|
|
| run |
| Posted: Fri Jan 07, 2011 2:52 pm Post subject: question |
|
|
Apprentice
Joined: 14 May 2007
Posts: 45
Location: Newyork, USA
|
| Thanks guys for your responses. It's all clear |
|
| Back to top |
|
|
| exerk |
| Posted: Fri Jan 07, 2011 2:53 pm Post subject: Re: question |
|
|
Jedi Council
Joined: 02 Nov 2006
Posts: 6339
|
| Vitor wrote: |
| ...When did broker enter the picture? ... |
The first post. I think the OP's 'B' and 'C' queue managers service a WM Broker (supposition).
_________________
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 |
|
|
| fatherjack |
| Posted: Fri Jan 07, 2011 3:00 pm Post subject: Re: question |
|
|
Knight
Joined: 14 Apr 2010
Posts: 522
Location: Craggy Island
|
| exerk wrote: |
| Vitor wrote: |
| ...When did broker enter the picture? ... |
The first post. I think the OP's 'B' and 'C' queue managers service a WM Broker (supposition). |
No need for supposition. The Op states...
| rahuldhanpal wrote: |
| We have Application responding back for a message received from either B Broker qmgr or C broker qmgr |
Not that it makes any difference to the issue. Does it?
_________________
Never let the facts get in the way of a good theory. |
|
| Back to top |
|
|
| exerk |
| Posted: Fri Jan 07, 2011 3:07 pm Post subject: Re: question |
|
|
Jedi Council
Joined: 02 Nov 2006
Posts: 6339
|
| fatherjack wrote: |
...No need for supposition. The Op states...
| rahuldhanpal wrote: |
| We have Application responding back for a message received from either B Broker qmgr or C broker qmgr |
|
Could be the internal Broker in V6.0 
| fatherjack wrote: |
| ...Not that it makes any difference to the issue. Does it? |
No...
_________________
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 |
|
|
|
|
