Author |
Message
|
guest468 |
Posted: Thu Oct 11, 2007 8:09 pm Post subject: MQ message acumulation is SCTQ |
|
|
Centurion
Joined: 30 May 2006 Posts: 146 Location: NY
|
Hi
We are hitting performance problem with MQ. We have a cluster having 3 qmgrs (2 servers, first server has a gateway qmgr). Client applications connect to gateway and put a total of around 400 messages per second on 4 cluster queues (these queues are hosted in non-gateway qmgrs). only 200 of them are actually being sent to destination qmgrs and rest 200 are getting acumlated in the SCTQ of gateway . Would appreciate what could be causing this. OS resources seem good (cpu 70% idle, filesystem is also look good)
Thanks |
|
Back to top |
|
 |
Vitor |
Posted: Thu Oct 11, 2007 11:46 pm Post subject: |
|
|
 Grand High Poobah
Joined: 11 Nov 2005 Posts: 26093 Location: Texas, USA
|
If they're sitting in the SCTQ and never moving, it's not a performance problem but a routing one. Check all the cluster queue managers are working properly, channels running, etc.
Another line of enquiry is what the held messages have in common. Check the xmit headers and see where they think they're going. Note that if there is a bad queue manager in the cluster and an application has done a bind on open then they will sit until their target comes back. _________________ Honesty is the best policy.
Insanity is the best defence. |
|
Back to top |
|
 |
guest468 |
Posted: Fri Oct 12, 2007 7:45 am Post subject: |
|
|
Centurion
Joined: 30 May 2006 Posts: 146 Location: NY
|
Hi
Messages are being transmitted over cluster channels but rate is 200/sec (to 100 each on destination qmgr). Message size is about 100KB. Could this be a performance issue?
Thanks |
|
Back to top |
|
 |
fjb_saper |
Posted: Fri Oct 12, 2007 8:57 am Post subject: |
|
|
 Grand High Poobah
Joined: 18 Nov 2003 Posts: 20756 Location: LI,NY
|
Looks more like a very slow network to me...  _________________ MQ & Broker admin |
|
Back to top |
|
 |
jefflowrey |
Posted: Fri Oct 12, 2007 9:01 am Post subject: |
|
|
Grand Poobah
Joined: 16 Oct 2002 Posts: 19981
|
Or slow disk access, or very small log file space, or improperly tuned channel batching parameters. _________________ I am *not* the model of the modern major general. |
|
Back to top |
|
 |
guest468 |
Posted: Fri Oct 12, 2007 9:05 am Post subject: |
|
|
Centurion
Joined: 30 May 2006 Posts: 146 Location: NY
|
We have the following setup
LogPrimaryFiles=20
LogSecondaryFiles=10
LogFilePages=16000
LogType=CIRCULAR
LogBufferPages=512
MQIBindType=FASTPATH
batch size is 50 as we were suggested by IBM to keep that value since they are cluster channels. |
|
Back to top |
|
 |
jefflowrey |
Posted: Fri Oct 12, 2007 9:08 am Post subject: |
|
|
Grand Poobah
Joined: 16 Oct 2002 Posts: 19981
|
Is that on BOTH sides of the slow channel? Or just one? _________________ I am *not* the model of the modern major general. |
|
Back to top |
|
 |
guest468 |
Posted: Fri Oct 12, 2007 1:16 pm Post subject: |
|
|
Centurion
Joined: 30 May 2006 Posts: 146 Location: NY
|
Jeff,
It's on all qmgrs in the cluster. |
|
Back to top |
|
 |
fjb_saper |
Posted: Fri Oct 12, 2007 1:18 pm Post subject: |
|
|
 Grand High Poobah
Joined: 18 Nov 2003 Posts: 20756 Location: LI,NY
|
Is any of the target queues full ?  _________________ MQ & Broker admin |
|
Back to top |
|
 |
guest468 |
Posted: Fri Oct 12, 2007 2:19 pm Post subject: |
|
|
Centurion
Joined: 30 May 2006 Posts: 146 Location: NY
|
Nope none of them are full. Filesystem there look good too.
But iostat %util reaches 100% every few seconds. |
|
Back to top |
|
 |
guest468 |
Posted: Fri Oct 12, 2007 2:28 pm Post subject: |
|
|
Centurion
Joined: 30 May 2006 Posts: 146 Location: NY
|
Sorry, i meant iostat %util in GateWay qmgr box reaches 100% once every few seconds |
|
Back to top |
|
 |
|