| |
|
RSS Feed - WebSphere MQ Support
|
RSS Feed - Message Broker Support
|
  |
|
| Scripted deployment performance problem |
« View previous topic :: View next topic » |
| Author |
Message
|
| jonesn |
 Posted: Fri Jul 24, 2009 5:29 am Post subject: Scripted deployment performance problem |
 |
|
Apprentice
Joined: 09 Jan 2002
Posts: 47
|
Chaps,
I am running MQSI 6.1.0.2, MQ 6.0.2.4 on AIX 5.3 and using a shell script to create execution groups and deploy flows. The problem is that the configmgr slows over time.
To investigate the problem I created a new broker and configmgr & repeatedly ran the script against them. It created 6 execution groups, deployed 20 flows, deleted all EGs and flows & then started again. The script ran overnight and the time taken to create everything started at 8 minutes & increased to 150 minutes after 15 iterations.
I also noted that the CPU time for bipconfigmgr, as reported by the UNIX ps command, rose significantly during this test & reached 800 minutes over the course of the test (about 14 hours).
Before running this test we increased the number of primary/secondary logs on the underlying queue manager as the normal scripted deployment failed and we noted errors concerning long running transactions in the queue manager error logs.
A few questions arise from this behaviour...
Why does the performance of the config manager degrade as it is being used?
Is the config manager really doing 800 minutes work in 14 hour period? This seems a very high figure & is actually the highest CPU use on the box!
Why would the configmgr have long running transactions? The errors in the MQ error logs along with the increasing CPU usage could point to long running transactions. And increasing the number of logs has stopped that problem happening (or postponed it at least). But surely the configmgr is getting dicreet requests so it should have many short transactions rather than long running.
I have searched this forum & found some posts that hint at bipconfigmgr performance but nothing that gets to a resolution.
There is no dynamic resource allocation happening on the box overnight so it has constant RAM, CPU etc throughout the test.
I plan to try to recreate the problem on a vanilla Linux box.
Any ideas?
Thanks
_________________
---
Nick Jones
IBM Certified Solutions Expert (WebSphere MQ Integrator) |
|
| Back to top |
|
 |
| fjb_saper |
| Posted: Sat Jul 25, 2009 1:22 am Post subject: |
|
|
Grand High Poobah
Joined: 18 Nov 2003
Posts: 20756
Location: LI,NY
|
Search the forum again... There is an APAR for this config mgr problem and you need to run a compress command...
Have fun. 
_________________
MQ & Broker admin |
|
| Back to top |
|
|
| PeterPotkay |
| Posted: Sat Jul 25, 2009 2:08 am Post subject: |
|
|
Poobah
Joined: 15 May 2001
Posts: 7722
|
And apply Fix Pack 6.1.0.4.
_________________
Peter Potkay
Keep Calm and MQ On |
|
| 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
|
|
|
|
