| Author |
Message
|
| Hari |
 Posted: Wed Dec 03, 2008 5:45 pm Post subject: Msg flow perf issue after moving to new server[Solved] |
 |
|
Centurion
Joined: 21 Nov 2002
Posts: 117
Location: USA
|
[b]The setup:[/b]
Broker: 6.0.0.6
MQ: 6.0.2.3
DB2: 8.2.9
Env: Windows 2003 SP1(Cluster)
[b]The problem:[/b]
We recently moved to a new server and since then noticed a drastic reduction in performance for SOME (not all) message flows which cater to high volume of messages.
Please note the message volume has not changed much. And of course, we haven't modified the code.
The following has been done, but to no avail:
1.) Runstats on the DB2 table and rebind packages. Also did a rebind on packages which were marked as Invalid.
2.) Exclude virus scans for trace folders etc.
3.) Our storage group has also confirmed that there are no disk related issues, and the i/o response time is low.
4.) Our administrators have taken broker statistics, which is ok.
Appreciate your help in resolving this issue.
Cheers,
Hari
Last edited by Hari on Tue Feb 17, 2009 10:06 am; edited 1 time in total |
|
| Back to top |
|
 |
| fjb_saper |
| Posted: Wed Dec 03, 2008 8:42 pm Post subject: |
|
|
Grand High Poobah
Joined: 18 Nov 2003
Posts: 20756
Location: LI,NY
|
Do your flows involve DB access?
Was the time spent on the DB ever measured?
What is that time today?
_________________
MQ & Broker admin |
|
| Back to top |
|
|
| Hari |
| Posted: Thu Dec 04, 2008 12:22 pm Post subject: |
|
|
Centurion
Joined: 21 Nov 2002
Posts: 117
Location: USA
|
Yes, it involves DB access.
Yes, the DB times are good, within milliseconds...
Thanks. |
|
| Back to top |
|
|
| fjb_saper |
| Posted: Thu Dec 04, 2008 2:49 pm Post subject: |
|
|
Grand High Poobah
Joined: 18 Nov 2003
Posts: 20756
Location: LI,NY
|
| Hari wrote: |
Yes, it involves DB access.
Yes, the DB times are good, within milliseconds...
Thanks. |
How do you know? What is the time spent on the DB, the time spent on the roundtrip (DB + Network)?
_________________
MQ & Broker admin |
|
| Back to top |
|
|
| Esa |
| Posted: Fri Dec 05, 2008 4:03 am Post subject: |
|
|
Grand Master
Joined: 22 May 2008
Posts: 1387
Location: Finland
|
If only some of the flows have performance problems, it could be an indication of a too small queue cache size. Especially if it is not always the same flows. You or somebody else may have tuned open handle caches on the old broker, or you have migrated it from version 5 that had a larger default cache size. But execution groups on the new broker will have the default queue cache size, which is 10 for version 6. If you have several flows in an execution group it is quite possible that some flows may have to wait for a free open queue handle.
Try increasing queue cache sizes of your execution groups like this:
mqsichangeproperties YourBroker -e YourExecutionGroup -o ComIbmMQConnectionManager -n queueCacheMaxSize -v YourNewCacheSize |
|
| Back to top |
|
|
| zpat |
| Posted: Fri Dec 05, 2008 6:04 am Post subject: |
|
|
Jedi Council
Joined: 19 May 2001
Posts: 5866
Location: UK
|
|
| Back to top |
|
|
| Hari |
| Posted: Thu Feb 19, 2009 5:11 pm Post subject: |
|
|
Centurion
Joined: 21 Nov 2002
Posts: 117
Location: USA
|
thanks [sic] to our original DBA who said that everything look good at the database end....
so finally we were able to resolve the problem and the culprit:
DB2 parameter DB2NTNOCACHE was turned ON. |
|
| Back to top |
|
|
| fjb_saper |
| Posted: Thu Feb 19, 2009 6:16 pm Post subject: |
|
|
Grand High Poobah
Joined: 18 Nov 2003
Posts: 20756
Location: LI,NY
|
| Hari wrote: |
thanks [sic] to our original DBA who said that everything look good at the database end....
so finally we were able to resolve the problem and the culprit:
DB2 parameter DB2NTNOCACHE was turned ON. |
Thanks for the update Hari. This shows that sometimes a little doubt can go a long way, or it always pays to doublecheck...
_________________
MQ & Broker admin |
|
| Back to top |
|
|
|
|
