| |
|
RSS Feed - WebSphere MQ Support
|
RSS Feed - Message Broker Support
|
  |
|
| [urgent] MB 6.1 performance issue |
« View previous topic :: View next topic » |
| Author |
Message
|
| yaakovd |
 Posted: Fri Sep 12, 2008 8:07 am Post subject: [urgent] MB 6.1 performance issue |
 |
|
Partisan
Joined: 20 Jan 2003
Posts: 319
Location: Israel
|
Hi ALL
I have big performance problem... hope for some ideas. 
Environment: AIX 5.3.0.7
WMB 6.1.0.3. with DB2 9.1.0.3
CPU - 2, RAM - 8 GB
Message flow that format binary file (on demand parsing) to ASCII TDS comma separated with some transformations.
Read file -> parse -> map (compute node) -> write file.
Input file 1 MB with 7000 records.
Flow doesn't work with any DB.
Processing time - 10 minutes !!! 
Same processing time on another machine (half resources).
Same flow on my laptop running about 3 minutes. I assumed that processing time on strong AIX machine should be much less.
I tryed:
ldedit -b maxdata:0x60000000 DataFlowEngine
LDR_CNTRL=MAXDATA=0x160000000 mqsistart brokername
without any affect.
| Code: |
Name PID CPU% PgSp Owner
DataFlow 942308 27.4 414.6 mbadmin
|
thanks ahead!!!
_________________
Best regards.
Yaakov
SWG, IBM Commerce, Israel |
|
| Back to top |
|
 |
| kimbert |
| Posted: Fri Sep 12, 2008 1:39 pm Post subject: |
|
|
Jedi Council
Joined: 29 Jul 2003
Posts: 5542
Location: Southampton
|
|
| Back to top |
|
|
| yaakovd |
| Posted: Fri Sep 12, 2008 11:26 pm Post subject: |
|
|
Partisan
Joined: 20 Jan 2003
Posts: 319
Location: Israel
|
Thanks, I know this article. I'm use similar pattern when processing large message (10 MB and more).
But 1 MB is not so big...
It is not really helps to understand why performance on small WIN machine 5 times better than on strong AIX server... 
_________________
Best regards.
Yaakov
SWG, IBM Commerce, Israel |
|
| Back to top |
|
|
| mqpaul |
| Posted: Mon Sep 29, 2008 4:38 am Post subject: Can you give some more information on AIX and flow? |
|
|
Acolyte
Joined: 14 Jan 2008
Posts: 66
Location: Hursley, UK
|
Flow performance depends on thress factors - the platform, the broker and the flow. (OK, pretty obvious  ). So, please tell us some more.
For example, is your AIX environment partitioned into LPARs, and if so, what resources are allocated to the LPAR running your execution group? Secondly, what does the map do, and which sort of map is it (XSLT, Mapping node, Compute node, etc.) ? Third, how is the broker configured - i.e. how much memory, how large is the JVM heap, etc. ?
You may get some useful information from the broker flow stats. (See the info centre for the mqsichangeflowstats command for more details.)
_________________
Paul |
|
| 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
|
|
|
|
