| |
|
RSS Feed - WebSphere MQ Support
|
RSS Feed - Message Broker Support
|
  |
|
| Performance issue in MQ linux |
« View previous topic :: View next topic » |
| Author |
Message
|
| kishi_25 |
 Posted: Wed Feb 01, 2012 11:12 am Post subject: Performance issue in MQ linux |
 |
|
Centurion
Joined: 19 Jul 2011
Posts: 100
|
I'm doing performance testing for MQ/MB in linux environment.
Following are the details of the environment.
Linux : X86-64
MQ : 7.0.015
MB 7003
Test scenario:
I have developed simple message flow with MQ input and MQ output node.
On Server A QM1 is defined and on Server B QM2 is defined.
QM1 and QM2 are multi instance QM's.
QM1 and QM2 are in cluster.
Cluser Queue Q1 on QM1 and Cluster Queue Q2 on QM2 are defined.
Q1 is mapped to mq input node and Q2 is mapped to QM2.
I have palced around 5000 messages in MQ input node.
Each message is 1MB size with persistant type.
The file system for QM1 and QM2 are defined on shared file system contact admin.
It took around 12 minutes (6 msgs/sec).
When I checked on IBM performance report..it referes as it can process around 125 messages/sec for 1MB with persistant type.
Any one can suggest inputs on, why there was that much difference in processing these messages?
Please let me know, what should I do to improve the performance? |
|
| Back to top |
|
 |
| bruce2359 |
| Posted: Wed Feb 01, 2012 11:25 am Post subject: |
|
|
Poobah
Joined: 05 Jan 2008
Posts: 9469
Location: US: west coast, almost. Otherwise, enroute.
|
The IBM Performance Report will also tell you about the hardware and o/s configuration. Is your configuration the same?
_________________
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 |
|
|
| zpat |
| Posted: Wed Feb 01, 2012 11:54 am Post subject: |
|
|
Jedi Council
Joined: 19 May 2001
Posts: 5866
Location: UK
|
Put your MQ logs and MQ queue data on different disks and away from other active files.
Use the highest performance disks that you can get, usually 10k+ rpm SAN drives with NVR write-caches.
Check the RAID settings - some LAN admins optimise their disk arrays for space and reads (e.g. 7+1) when you need to optimise for transactional throughput and writes (e.g. 5+1).
The key will be the write performance to the disks.
You can adjust log buffers and so on, but the raw I/O performance makes the most difference. |
|
| Back to top |
|
|
| mqjeff |
| Posted: Wed Feb 01, 2012 12:01 pm Post subject: |
|
|
Grand Master
Joined: 25 Jun 2008
Posts: 17447
|
| Did you run the message flow with a single instance? or more than 0 additional instances... |
|
| Back to top |
|
|
| bruce2359 |
| Posted: Wed Feb 01, 2012 12:10 pm Post subject: |
|
|
Poobah
Joined: 05 Jan 2008
Posts: 9469
Location: US: west coast, almost. Otherwise, enroute.
|
Moved to Performance.
_________________
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 |
|
|
|
|
|
| |
|
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
|
|
|
|
