| Author |
Message
|
| James_O_K |
 Posted: Wed Nov 29, 2006 2:11 am Post subject: SVRCONN channel has irregular amount of messages |
 |
|
Apprentice
Joined: 09 Mar 2006
Posts: 39
Location: UK
|
Hello all,
I've just been told that a SVRCONN channel on one of our queue managers has a ridiculously high value for the MSGS param:
CHANNEL(HCSCHANNEL) XMITQ( )
CONNAME(10.160.203.81) CURRENT
CHLTYPE(SVRCONN) STATUS(RUNNING)
LSTMSGTI(09.47.37) LSTMSGDA(2006-11-29)
MSGS(76500747) BYTSSENT(3278627650)
BYTSRCVD(3278628316) HBINT(300)
CHSTATI(05.00.1 CHSTADA(2006-11-29)
BUFSSENT(76500748) BUFSRCVD(76500750)
JOBNAME(000DC01400000001) MCASTAT(RUNNING)
STOPREQ(NO) LOCLADDR(10.160.203.81(4414))
SSLPEER() RQMNAME( )
This has incremented from 0 last night, and I'm told that we dont have 76 million messages going through all the MQ infrastructure in the organisation in one day!
Ive looked in the documentation without much success as to what this parameter is. Can someone enlighten me?
There is a java program running against the queue which we suspect may be causing the problem (as we have an identical queue manager with identical setup on another machine that is running fine).
Does anyone know of any common causes/symptoms of this problem? |
|
| Back to top |
|
 |
| James_O_K |
| Posted: Wed Nov 29, 2006 2:15 am Post subject: |
|
|
Apprentice
Joined: 09 Mar 2006
Posts: 39
Location: UK
|
| I should add its MQ 5.3 running on AIX |
|
| Back to top |
|
|
| Vitor |
| Posted: Wed Nov 29, 2006 2:16 am Post subject: |
|
|
Grand High Poobah
Joined: 11 Nov 2005
Posts: 26093
Location: Texas, USA
|
| James_O_K wrote: |
| I should add its MQ 5.3 running on AIX |
Which CSD? Java and low levels of 5.3 CSD don't mix too well.
_________________
Honesty is the best policy.
Insanity is the best defence. |
|
| Back to top |
|
|
| James_O_K |
| Posted: Wed Nov 29, 2006 2:47 am Post subject: |
|
|
Apprentice
Joined: 09 Mar 2006
Posts: 39
Location: UK
|
| Vitor wrote: |
| James_O_K wrote: |
| I should add its MQ 5.3 running on AIX |
Which CSD? Java and low levels of 5.3 CSD don't mix too well. |
Heres the info:
$ mqver
Name: WebSphere MQ
Version: 530.10 CSD10
CMVC level: p530-10-L050504
BuildType: IKAP - (Production) |
|
| Back to top |
|
|
| jefflowrey |
| Posted: Wed Nov 29, 2006 2:48 am Post subject: |
|
|
Grand Poobah
Joined: 16 Oct 2002
Posts: 19981
|
Could it be that the java program is in a get/rollback loop due to a poison message?
Also, the MSGS parameter, if I remember other discussions I've seen here, is not 1 to 1 with the number of actual MQ messages that have passed through the channel. It's something like the number of MQI commands that have passed over the svrconn.
_________________
I am *not* the model of the modern major general. |
|
| Back to top |
|
|
| James_O_K |
| Posted: Wed Nov 29, 2006 3:12 am Post subject: |
|
|
Apprentice
Joined: 09 Mar 2006
Posts: 39
Location: UK
|
| jefflowrey wrote: |
Could it be that the java program is in a get/rollback loop due to a poison message?
Also, the MSGS parameter, if I remember other discussions I've seen here, is not 1 to 1 with the number of actual MQ messages that have passed through the channel. It's something like the number of MQI commands that have passed over the svrconn. |
Jeff,
Thanks for prompt reply.
I dont think its in get/rollback as its straight forward mq messaging rather than JMS.
Your second point is an interesting one, we are aware of some performance problems with the current code (which is fixed in next version). Currently the program loops and does queue.get() every time (the new code does a "get wait"). This figure could potentially be because of the amount of loops in the program. Bad bad bad code I know.
OK, so another question then. Will it affect anything having this value so high (from an MQ series point of view). |
|
| Back to top |
|
|
| jefflowrey |
| Posted: Wed Nov 29, 2006 3:16 am Post subject: |
|
|
Grand Poobah
Joined: 16 Oct 2002
Posts: 19981
|
You can do syncpoint with base MQ.
It shouldn't affect anything having this number be large.
Yes, running get repeatedly in a loop without any delaying in it will cause this number to get large, and not in a 1-1 with the number of GETS.
_________________
I am *not* the model of the modern major general. |
|
| Back to top |
|
|
| James_O_K |
| Posted: Wed Nov 29, 2006 4:00 am Post subject: |
|
|
Apprentice
Joined: 09 Mar 2006
Posts: 39
Location: UK
|
| jefflowrey wrote: |
You can do syncpoint with base MQ.
It shouldn't affect anything having this number be large.
Yes, running get repeatedly in a loop without any delaying in it will cause this number to get large, and not in a 1-1 with the number of GETS. |
Thanks Jeff |
|
| Back to top |
|
|
|
|
