| Author |
Message
|
| gmark1 |
 Posted: Fri Oct 12, 2007 10:51 am Post subject: Change in timeout. Impact in Performance |
 |
|
Novice
Joined: 10 Jul 2007
Posts: 11
|
Currently I am working in a WebSphere Message Broker project which I've developed using the IBM toolkit 6.0.2. I have several message flows that include AggregateControl nodes set to a timeout of 15 minutes. However, due to some unexpected changes the timeout needs to be increase to 3 days. As such, I consider that timeout to be excessive.
I would like to know if changing the timeout would have a impact on the performance of the brokers. Would the resources needed by the brokers increase in a very large amount? Which strategy would be the most efficient (meaning consuming less resources form the machine is running in) setting the time out to a long period of time or eliminating the timeout completely (by setting the timeout value to 0)?
English is not my mother language. I probably made several mistakes during my post. If my doubts become too confusing, please let me know and I rephrase them.
Thanks in advance. |
|
| Back to top |
|
 |
| jefflowrey |
| Posted: Fri Oct 12, 2007 11:12 am Post subject: |
|
|
Grand Poobah
Joined: 16 Oct 2002
Posts: 19981
|
I would only disable the timeout if you are absolutely positively guaranteed to always receive all reply messages EVENTUALLY.
It looks like in v6, the Timeout is controlled using an internal expiry message - making a guess based on comments about setting EXPRYINT values on zOS. That's going to be about as efficient as anything else you could do.
So set it to a long period of time.
_________________
I am *not* the model of the modern major general. |
|
| Back to top |
|
|
| gmark1 |
| Posted: Fri Oct 12, 2007 11:39 am Post subject: |
|
|
Novice
Joined: 10 Jul 2007
Posts: 11
|
| jefflowrey wrote: |
I would only disable the timeout if you are absolutely positively guaranteed to always receive all reply messages EVENTUALLY.
It looks like in v6, the Timeout is controlled using an internal expiry message - making a guess based on comments about setting EXPRYINT values on zOS. That's going to be about as efficient as anything else you could do.
So set it to a long period of time. |
Thanks for your response. If I undertood you correctly, the impact on the perfomance is the same wheter the timeout is set to a few minutes or several days. Am I right? |
|
| Back to top |
|
|
| jefflowrey |
| Posted: Fri Oct 12, 2007 12:03 pm Post subject: |
|
|
Grand Poobah
Joined: 16 Oct 2002
Posts: 19981
|
That is what I am guessing, based on an obscure comment in the documentation and my own intuition of how I would write the code.
If you really absolutely have to know down to a fine level of detail, you can try opening a PMR and talking to IBM support. You may have to spend some time waiting for a response, and may not get anything more solid.
_________________
I am *not* the model of the modern major general. |
|
| Back to top |
|
|
| gmark1 |
| Posted: Fri Oct 12, 2007 12:06 pm Post subject: |
|
|
Novice
Joined: 10 Jul 2007
Posts: 11
|
| jefflowrey wrote: |
That is what I am guessing, based on an obscure comment in the documentation and my own intuition of how I would write the code.
If you really absolutely have to know down to a fine level of detail, you can try opening a PMR and talking to IBM support. You may have to spend some time waiting for a response, and may not get anything more solid. |
I understand. Thanks again for your help. |
|
| Back to top |
|
|
|
|
