| Author |
Message
|
| fredmoore |
 Posted: Wed Nov 07, 2018 6:40 am Post subject: Is async MQPUT (MQPMO_ASYNC_RESPONSE) exploitable in IIB? |
 |
|
Novice
Joined: 23 Mar 2009
Posts: 24
|
Hi,
is there a way to achieve asynch MQPUT behavior in IIB (i.e. MQPMO_ASYNC_RESPONSE)? How?
I don't see anything related to this in MQOutput node...
Cheers,
F. |
|
| Back to top |
|
 |
| Vitor |
| Posted: Wed Nov 07, 2018 6:58 am Post subject: Re: Is async MQPUT (MQPMO_ASYNC_RESPONSE) exploitable in IIB |
|
|
Grand High Poobah
Joined: 11 Nov 2005
Posts: 26093
Location: Texas, USA
|
| fredmoore wrote: |
| is there a way to achieve asynch MQPUT behavior in IIB (i.e. MQPMO_ASYNC_RESPONSE)? How? |
Use the MQOutput node outside of the transaction the flow starts (or not use a transaction at all)
_________________
Honesty is the best policy.
Insanity is the best defence. |
|
| Back to top |
|
|
| fredmoore |
| Posted: Wed Nov 07, 2018 7:48 am Post subject: |
|
|
Novice
Joined: 23 Mar 2009
Posts: 24
|
| Quote: |
| Use the MQOutput node outside of the transaction the flow starts (or not use a transaction at all) |
This means that any non transactional MQPUT originating from an MQOutput node will then be executed with MQPMO_ASYNC_RESPONSE?
I could not find any trace of this in IIB docs... |
|
| Back to top |
|
|
| Vitor |
| Posted: Wed Nov 07, 2018 8:43 am Post subject: |
|
|
Grand High Poobah
Joined: 11 Nov 2005
Posts: 26093
Location: Texas, USA
|
My bad - saw "async", read "syncpoint".
IIB doesn't (yet) support this.
_________________
Honesty is the best policy.
Insanity is the best defence. |
|
| Back to top |
|
|
| fredmoore |
| Posted: Wed Nov 07, 2018 2:52 pm Post subject: |
|
|
Novice
Joined: 23 Mar 2009
Posts: 24
|
| Quote: |
| IIB doesn't (yet) support this. |
I did some digging, assuming that the broker is not passing any option related to sync/async put, the put will behave as MQPMO_RESPONSE_AS_Q_DEF, so setting DEFPRESP=ASYNC on the target QLOCAL should do the trick.
Does it make sense?
Cheers,
F. |
|
| Back to top |
|
|
| hughson |
| Posted: Wed Nov 07, 2018 9:04 pm Post subject: |
|
|
Padawan
Joined: 09 May 2013
Posts: 1959
Location: Bay of Plenty, New Zealand
|
| fredmoore wrote: |
| Quote: |
| IIB doesn't (yet) support this. |
I did some digging, assuming that the broker is not passing any option related to sync/async put, the put will behave as MQPMO_RESPONSE_AS_Q_DEF, so setting DEFPRESP=ASYNC on the target QLOCAL should do the trick.
Does it make sense?
Cheers,
F. |
_________________
Morag Hughson @MoragHughson
IBM MQ Technical Education Specialist
Get your IBM MQ training here!
MQGem Software |
|
| Back to top |
|
|
| fredmoore |
| Posted: Thu Nov 08, 2018 12:10 am Post subject: |
|
|
Novice
Joined: 23 Mar 2009
Posts: 24
|
| Great! Thanks Vitor & Morag! |
|
| Back to top |
|
|
| rekarm01 |
| Posted: Thu Nov 08, 2018 6:51 pm Post subject: |
|
|
Grand Master
Joined: 25 Jun 2008
Posts: 1415
|
| fredmoore wrote: |
| the put will behave as MQPMO_RESPONSE_AS_Q_DEF, so setting DEFPRESP=ASYNC on the target QLOCAL should do the trick. |
Yes, but the MQOutput node doesn't call MQSTAT, so:
- For messages not put under syncpoint, the message flow has no way to check whether the put ultimately succeeds.
- For messages put under syncpoint, if the put fails, then the commit fails too, but the message flow won't know why.
|
|
| Back to top |
|
|
| Vitor |
| Posted: Fri Nov 09, 2018 6:00 am Post subject: |
|
|
Grand High Poobah
Joined: 11 Nov 2005
Posts: 26093
Location: Texas, USA
|
| rekarm01 wrote: |
| Yes, but the MQOutput node doesn't call MQSTAT |
Hence my original comment.
_________________
Honesty is the best policy.
Insanity is the best defence. |
|
| Back to top |
|
|
| fredmoore |
| Posted: Fri Nov 09, 2018 6:11 am Post subject: |
|
|
Novice
Joined: 23 Mar 2009
Posts: 24
|
In my use case messages are non persistent and outside of UOW, and it is tolerated to miss a message now and then if this gives a gain in throughput.
So DEFPRESP=ASYNC sounds promising.
Cheers,
F. |
|
| Back to top |
|
|
| Vitor |
| Posted: Fri Nov 09, 2018 6:14 am Post subject: |
|
|
Grand High Poobah
Joined: 11 Nov 2005
Posts: 26093
Location: Texas, USA
|
| fredmoore wrote: |
In my use case messages are non persistent and outside of UOW, and it is tolerated to miss a message now and then if this gives a gain in throughput.
So DEFPRESP=ASYNC sounds promising. |
And I wish you joy of it.
But given we never delete threads from this forum, it's important not to give the impression to future readers it's fully supported at present.
I'm sure this will change in time.
_________________
Honesty is the best policy.
Insanity is the best defence. |
|
| Back to top |
|
|
|
|
