| |
|
RSS Feed - WebSphere MQ Support
|
RSS Feed - Message Broker Support
|
  |
|
| Compute Node "Compute Mode" as related to performa |
« View previous topic :: View next topic » |
| Author |
Message
|
| BergerC |
 Posted: Wed Dec 13, 2006 1:33 pm Post subject: Compute Node "Compute Mode" as related to performa |
 |
|
Novice
Joined: 08 Apr 2002
Posts: 15
|
Hi All -
I get the feeling I've read this somewhere, but I can't locate it.
Env:
MQSI v2.1 (I know)
MQ v5.3
AIX 5.2
We have some very large messages we're processing through a message flow. Fairly simple flow - essentially two compute nodes, the first to perform a DB insert and validation, the second for the majority of processing. For the sake of this discussion, let's just agree that we can't combine the two nodes...and I can't split the message into smaller chunks...
If I set the Compute Node on the Advance tab of the first Compute Node to "LocalEnvironment" (since I'm not changing the message content), do I avoid parsing the message at all?
My primary concern - we run on AIX, and we have issues with the Execution Group running out of memory with very large messages. We're reviewing all the workarounds with IBM, but anything we can do to reduce memory requirements is helpful.
An answer on the Mode, and perhaps any other lessons learned for saving memory (REFERENCE, use of LocalEnvironment, etc), would be appreciated.
Thanks All -
_________________
_____________
Chris |
|
| Back to top |
|
 |
| jefflowrey |
| Posted: Wed Dec 13, 2006 2:00 pm Post subject: |
|
|
Grand Poobah
Joined: 16 Oct 2002
Posts: 19981
|
You only avoid parsing the message by avoiding writing code that accesses fields in the message. Calling 'Set OutputRoot = InputRoot;" should not add additional parsing.
Setting ComputeNode to LocalEnvironment will mean that your next node downstream might not receive your message tree.
But if your first node is performing VALIDATION, then it needs to parse the entire message no matter what - because otherwise you don't know that it's validated.
If you're only doing a partial validation, then you can try and control how much parsing is done to get at the fields you need to validate. But that gets into stuff like changing the message definition so the important fields are at the front - which you might not be willing or able to do.
_________________
I am *not* the model of the modern major general. |
|
| Back to top |
|
|
| BergerC |
| Posted: Wed Dec 13, 2006 2:38 pm Post subject: |
|
|
Novice
Joined: 08 Apr 2002
Posts: 15
|
Ah yes - I should have been careful in saying "parsing at all". And your point on attempting to oragnize the message is well taken. I was hoping to avoid unneccessay parsing, and it sounds like Output = Input is the best way to go.
However, when you say that setting the Mode to LocalEnvironment may result in the Message content not being passed through the node, I was under the impression that this Mode setting only controls what areas may be changed in the node (not what should be passed.) In other words, if I set Mode to LocalEnvironment, I can code the Output message all I want, but it will not be changed - the input message is propogated to the output regardless of what the ESQL asks for. Am I off base here? |
|
| Back to top |
|
|
|
|
|
| |
|
Page 1 of 1 |
|
You cannot post new topics in this forum
You cannot reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum
You cannot vote in polls in this forum
|
|
|
|
