| Author |
Message
|
| FBTO |
 Posted: Wed Nov 08, 2006 12:56 pm Post subject: Need help understanding Expiry Time?? |
 |
|
Newbie
Joined: 08 Nov 2006
Posts: 5
|
I have a question about the MQ Expiry Time.
I have a senario where I send a stream of messages over a 5 minute period, with an expiry time of 15 seconds. If I don't get a response from any of these messages in 15 seconds, I get a host down message. During my test, I processed around 394 cases in a 5 minute window without any host down messages. I ran the exact same test, but the only difference was that I doubled the MQ Expiry time to 30 seconds. Strangely enough, I got a better throughput, and was able to process over 455 cases in the 5 minute window.
So my question is, How does MQ expiry time work? I increased my expiry, and got a better throughput. I repeated these tests multiple times, and always got better results with higher expiry time. Does the MQ know that the expiry time is so big and runs some process that polices the queues less often, therefore allowing more messages to be processed?
I look forward to any thoughts on this matter.
Thanks!! |
|
| Back to top |
|
 |
| kevinf2349 |
| Posted: Wed Nov 08, 2006 1:35 pm Post subject: |
|
|
Grand Master
Joined: 28 Feb 2003
Posts: 1311
Location: USA
|
My take would be that MQ is not calling the cleanup process as often therefore more work can be pushed through.
Just a SWAG  |
|
| Back to top |
|
|
| jefflowrey |
| Posted: Wed Nov 08, 2006 1:42 pm Post subject: |
|
|
Grand Poobah
Joined: 16 Oct 2002
Posts: 19981
|
My take would be that the GET is returning quicker by not having to deal with expired messages.
But, again, a SWAG.
_________________
I am *not* the model of the modern major general. |
|
| Back to top |
|
|
| FBTO |
| Posted: Wed Nov 08, 2006 2:27 pm Post subject: |
|
|
Newbie
Joined: 08 Nov 2006
Posts: 5
|
In both tests, there were zero expired messages.
Can someone explain how the queuing works in regards to matching with the correlation id and timeouts? |
|
| Back to top |
|
|
| kevinf2349 |
| Posted: Wed Nov 08, 2006 3:53 pm Post subject: |
|
|
Grand Master
Joined: 28 Feb 2003
Posts: 1311
Location: USA
|
| Quote: |
| there were zero expired messages. |
...why would that matter?
I was under the impression that the process of expiring messages (not deleteing them) goes on regradless of the MQGET process but it is only the MQGET the causes then to actually be deleted.
Of course I could be wrong....but that would certainly explain the throughput too. 
| Quote: |
| Can someone explain how the queuing works in regards to matching with the correlation id and timeouts? |
What is to explain? Either a message matches the CORRELID or it doesn't. Timeout (expiration) has nothing to do with it. If a message is expired on the queue any MQGET (correlid matching or not) will result in the message being discarded |
|
| Back to top |
|
|
| Vitor |
| Posted: Thu Nov 09, 2006 12:37 am Post subject: |
|
|
Grand High Poobah
Joined: 11 Nov 2005
Posts: 26093
Location: Texas, USA
|
| kevinf2349 wrote: |
I was under the impression that the process of expiring messages (not deleteing them) goes on regradless of the MQGET process but it is only the MQGET the causes then to actually be deleted.
|
I think this changed - in v5.3 and below the MQGET determines if the message is expiried and deletes it, in v6 there's a separate process that cleans them up in addition to the MQGET. I swear I've read that some place but can't find the reference. Any kind posters who know where it's written down and share would prevent me needing my medication checked at the weekend...... 
FBTO - what level of MQ was your test run on?
_________________
Honesty is the best policy.
Insanity is the best defence. |
|
| Back to top |
|
|
| exerk |
| Posted: Thu Nov 09, 2006 2:38 am Post subject: |
|
|
Jedi Council
Joined: 02 Nov 2006
Posts: 6339
|
| Vitor wrote: |
| ...in v6 there's a separate process that cleans them up in addition to the MQGET. I swear I've read that some place but can't find the reference... |
I can't find it either but thought that was a feature of WMQ on z/OS, unless they've ported the feature cross-platform. More than happy to be corrected!
_________________
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 |
|
|
| wschutz |
| Posted: Thu Nov 09, 2006 3:33 am Post subject: |
|
|
Jedi Knight
Joined: 02 Jun 2005
Posts: 3316
Location: IBM (retired)
|
|
| Back to top |
|
|
| Vitor |
| Posted: Thu Nov 09, 2006 3:54 am Post subject: |
|
|
Grand High Poobah
Joined: 11 Nov 2005
Posts: 26093
Location: Texas, USA
|
This explains why searching the documentation was proving fruitless!
I knew I'd seen something about it somewhere.
Many thanks for the clarification. I'd hate to change my doseage again...
_________________
Honesty is the best policy.
Insanity is the best defence. |
|
| Back to top |
|
|
| FBTO |
| Posted: Thu Nov 09, 2006 7:24 am Post subject: |
|
|
Newbie
Joined: 08 Nov 2006
Posts: 5
|
| Vitor wrote: |
| kevinf2349 wrote: |
I was under the impression that the process of expiring messages (not deleteing them) goes on regradless of the MQGET process but it is only the MQGET the causes then to actually be deleted.
|
I think this changed - in v5.3 and below the MQGET determines if the message is expiried and deletes it, in v6 there's a separate process that cleans them up in addition to the MQGET. I swear I've read that some place but can't find the reference. Any kind posters who know where it's written down and share would prevent me needing my medication checked at the weekend......
FBTO - what level of MQ was your test run on? |
If by level you mean what version, it is v5.3. |
|
| Back to top |
|
|
| Vitor |
| Posted: Thu Nov 09, 2006 7:29 am Post subject: |
|
|
Grand High Poobah
Joined: 11 Nov 2005
Posts: 26093
Location: Texas, USA
|
I did indeed. Apologies for my poor terminology.
Then AFAIK the expired messages are cleared off as a part of the MQGET. This whole background process thing seems to be me mis-remembering the post wschutz found.
_________________
Honesty is the best policy.
Insanity is the best defence. |
|
| Back to top |
|
|
| bruce2359 |
| Posted: Thu Nov 09, 2006 9:33 am Post subject: |
|
|
Guest
|
| Expiry is documented in substantial detail in sc34-6596-00 MQ Application Programming Reference, Chapter 12, page 152. Included is behavior in both midrange and z/OS. |
|
| Back to top |
|
|
| FBTO |
| Posted: Thu Nov 09, 2006 11:57 am Post subject: |
|
|
Newbie
Joined: 08 Nov 2006
Posts: 5
|
| bruce2359 wrote: |
| Expiry is documented in substantial detail in sc34-6596-00 MQ Application Programming Reference, Chapter 12, page 152. Included is behavior in both midrange and z/OS. |
Thanks..
Is this document on the web anywhere? |
|
| Back to top |
|
|
| kevinf2349 |
| Posted: Thu Nov 09, 2006 12:17 pm Post subject: |
|
|
Grand Master
Joined: 28 Feb 2003
Posts: 1311
Location: USA
|
| Quote: |
| Is this document on the web anywhere? |
Look above under the DOCUMENTATION button. (AP Ref)
Or you could Goggle it, or go the the IBM Website and get it, or get it of the install CD.
btw in my manual it is in Chapter 10 under the MQMD Structure |
|
| Back to top |
|
|
| FBTO |
| Posted: Mon Nov 13, 2006 8:46 am Post subject: |
|
|
Newbie
Joined: 08 Nov 2006
Posts: 5
|
| kevinf2349 wrote: |
| Quote: |
| Is this document on the web anywhere? |
Look above under the DOCUMENTATION button. (AP Ref)
Or you could Goggle it, or go the the IBM Website and get it, or get it of the install CD.
btw in my manual it is in Chapter 10 under the MQMD Structure |
Thank you!!
And thanks to everyone for all of their help!! |
|
| Back to top |
|
|
|
|
