| |
|
RSS Feed - WebSphere MQ Support
|
RSS Feed - Message Broker Support
|
  |
|
| Less performance 5.3 vs 6.0 |
« View previous topic :: View next topic » |
| Author |
Message
|
| joaze |
 Posted: Mon Aug 11, 2008 5:40 am Post subject: Less performance 5.3 vs 6.0 |
 |
|
Newbie
Joined: 11 Aug 2008
Posts: 5
|
Hello,
We have recently upgrating our MQ version from 5.3 to 6.0.
We have noticed that performance of integration has decreased more than 50%.
Messages are about 4ko, défined persistent, and we have tests for more than 1500 messages (we expect about 20 000 per day).
Applicatoin scenario :
- principaly getting's java call from api native (com.ibm.mq.jar)
- mode bindings connection
- insert in database oracle
- and commit final (db then mq accordintg to catch exception)
- no specific mq transaction with syncpoint are used - qmgr has been created with default definition
Does someone known about similar problem between these 2 versions ?
nb : when we look for bench evaluations, they noticed degradation "by
up to 15%" for messages of 2K...
Thanks in advance. |
|
| Back to top |
|
 |
| sumit |
| Posted: Tue Aug 12, 2008 12:06 pm Post subject: |
|
|
Partisan
Joined: 19 Jan 2006
Posts: 398
|
Was there any other application that was changed since the quue manager upgrade? How are you measuring the degrade in performance? Is it related to number of messages processed by queue manager?
_________________
Regards
Sumit |
|
| Back to top |
|
|
| fjb_saper |
| Posted: Tue Aug 12, 2008 12:30 pm Post subject: |
|
|
Grand High Poobah
Joined: 18 Nov 2003
Posts: 20756
Location: LI,NY
|
Careful, Usually a performance degradation after 3,6,9 or 12 months is a DB problem where nobody accounted for a moving target. I'd say this certainly requires more analysis before laying it at MQ's feet... 
_________________
MQ & Broker admin |
|
| Back to top |
|
|
| joaze |
| Posted: Tue Aug 12, 2008 2:20 pm Post subject: |
|
|
Newbie
Joined: 11 Aug 2008
Posts: 5
|
No other applications have been installed on the platform (AIX 52) and no applicatives changes between these 2 tests.
Concerning the upgrating, 1rst tests v53 was done 2 weeks ago before tests v60.
At this step, I have just evaluated time processed for whole messages by our application.
Before changing code and trying to separate each time elapsed (db, mq, ...), we have planned to back to MQ53 and test once again in the same context.
For information, we have the exact same platform (hardware, OS) with same version of our application , products, ... except for MQ (53) - in this environment, performance seems to be completely ok. |
|
| Back to top |
|
|
| fjb_saper |
| Posted: Wed Aug 13, 2008 4:01 am Post subject: |
|
|
Grand High Poobah
Joined: 18 Nov 2003
Posts: 20756
Location: LI,NY
|
| joaze wrote: |
No other applications have been installed on the platform (AIX 52) and no applicatives changes between these 2 tests.
Concerning the upgrating, 1rst tests v53 was done 2 weeks ago before tests v60.
At this step, I have just evaluated time processed for whole messages by our application.
Before changing code and trying to separate each time elapsed (db, mq, ...), we have planned to back to MQ53 and test once again in the same context.
For information, we have the exact same platform (hardware, OS) with same version of our application , products, ... except for MQ (53) - in this environment, performance seems to be completely ok. |
No changes? You mean to say same amount of rows in your DB, same indexes, same tcp/ip latency, same DB drivers, same DB Version, same java/c/prog language version, fix pack, compiler, etc... ??
_________________
MQ & Broker admin |
|
| Back to top |
|
|
| joaze |
| Posted: Wed Aug 13, 2008 4:40 am Post subject: |
|
|
Newbie
Joined: 11 Aug 2008
Posts: 5
|
| same amount of rows in your DB same indexes, same tcp/ip latency, same DB drivers, same DB Version, same java/c/prog language version, fix pack, compiler : normaly yes - it's an integration platform similar as our production. The only real difference concern part of user's activity on each platform. |
|
| Back to top |
|
|
| bruce2359 |
| Posted: Wed Aug 13, 2008 6:34 am Post subject: |
|
|
Poobah
Joined: 05 Jan 2008
Posts: 9469
Location: US: west coast, almost. Otherwise, enroute.
|
| Quote: |
| For information, we have the exact same platform (hardware, OS) with same version of our application , products, ... except for MQ (53) - in this environment, performance seems to be completely ok. |
You might just be the first to have achieved this exact same hardware/software configuration. Similar is the best I'd hope for.
One change: v6 has a slightly larger footprint than 5.3. It's difficult to imagine how that might have resulted in 50% MQ degradation; but you might want to add some more RAM for MQs use.
Without more details (from a capacity planning/performance tuning exercise), I'd also suspect DB tuning opportunities exist.
_________________
I like deadlines. I like to wave as they pass by.
ב''ה
Lex Orandi, Lex Credendi, Lex Vivendi. As we Worship, So we Believe, So we Live. |
|
| Back to top |
|
|
| joaze |
| Posted: Wed Aug 13, 2008 6:58 am Post subject: |
|
|
Newbie
Joined: 11 Aug 2008
Posts: 5
|
We have planed new tests on this platform with a v53 configuration, then with 60. If performance are same, we could consider that mq might not be concerned.
I'll take you inform. |
|
| Back to top |
|
|
| joaze |
| Posted: Thu Aug 14, 2008 4:07 am Post subject: |
|
|
Newbie
Joined: 11 Aug 2008
Posts: 5
|
Hello,
I confirm that problem is not MQ.
We'll see with our dba for next and check our db configuration.
Thanks for your help. |
|
| 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
|
|
|
|
