| Author |
Message
|
| dp111443 |
 Posted: Mon May 24, 2004 1:07 am Post subject: Monitoring MQ channels when in retry status |
 |
|
Voyager
Joined: 25 Feb 2004
Posts: 82
|
Hi All,
We are currently using BMC Patrol for Websphere MQ to monitor MQ on all systems. What we have done is to monitor all Queue Managers and objects centrally using Queue Manager and Node Manager concept that BMC provides.
We have also written unix scripts to take action on particular events (e.g. QMgr down, queue full, etc) to raise tickets on helpdesk system and produce web page which issues the current status of alerts using mcmmqrc (run runmqsc command remotely to all nodes that are managed) when a particular event is raised.
anyway, with Patrol for MQ we can automatically act upon events (queue full - 2053, Qmgr down - 2223, etc.) based on the corresponding MQ event number. However am struggling to do the same on channels. we get events on channel stopped, channel stopped by user, channel started, etc. however we want to pick up on channels that are in "retrying" status, but MQ events arn't raised for this. Any idea why and what I need to do to get round this problem to raise the issue of a retrying channel (on any node manager) ???
I have noticed that when a channel is in retrying status the qmgr logs are updated with this information saying the host/port are invalid. I guess you could grep for instances once in a while there. However, this monitoring has to be done locally and we could potentially have 200 of these servers with MQ on which we'd prefer to monitor centrally (ideally using BMC Patrol) like we monitor all other instances with too.
If you have any questions please ask as i understand my question might not be clear and thanks for your help in advance.
_________________
Integration Design/Developer
IBM Certified System Administrator -
WebSphere MQ V5.3 |
|
| Back to top |
|
 |
| JasonE |
| Posted: Mon May 24, 2004 1:53 am Post subject: |
|
|
Grand Master
Joined: 03 Nov 2003
Posts: 1220
Location: Hursley
|
Try this...
| Quote: |
The behaviour has changed since the introduction of IC29815 so that
"Channel Stopped" events are only generated when the channel program itself generates a stopped event. Customers will therefore find that if
a channel goes straight into a retry mode after it is started, they will
not get a MQRC_CHANNEL_STOPPED event with MQRQ_CHANNEL_STOPPED_RETRY.
.
Therefore the customer will need to add the following
to the CHANNEL stanza of the qm.ini StopEvent=ALWAYS if they want the previous behaviour |
|
|
| Back to top |
|
|
| dp111443 |
| Posted: Mon May 24, 2004 4:22 am Post subject: |
|
|
Voyager
Joined: 25 Feb 2004
Posts: 82
|
Hi Jason,
Thanks for your reply. Unfortunately, I don't quiet understand what the entry to the qm.ini file will do as I didn't quite understand the quote previous to that.
Could you explain to me what this change will do and is this something that will have to be done to all MQ Server nodes that are installed?
is there any links that can help my lack of understanding?
Many thanks for your help.
Regards,
dp111443
_________________
Integration Design/Developer
IBM Certified System Administrator -
WebSphere MQ V5.3 |
|
| Back to top |
|
|
| JasonE |
| Posted: Mon May 24, 2004 4:42 am Post subject: |
|
|
Grand Master
Joined: 03 Nov 2003
Posts: 1220
Location: Hursley
|
You used to get a stopped event when the channel pgm ran and ended, but an APAR changed it so you only got the stopped event IF (and only IF) the channel pgm ran and issued a started event first.
You need a stopped event without a started, otherwise you cannot see retry (where it never gets started). Setting this value will give you the events you said you wanted to trigger off to monitor retry events. You would need to set it on all qmgrs you wish to monitor the retry events from.
No further docs, sorry.... |
|
| Back to top |
|
|
| dp111443 |
| Posted: Mon May 24, 2004 6:16 am Post subject: |
|
|
Voyager
Joined: 25 Feb 2004
Posts: 82
|
Hi,
I have added the new stanza to a test queue manager. Have I done it correctly below:
Log:
LogPrimaryFiles=3
LogSecondaryFiles=2
LogFilePages=1024
LogType=CIRCULAR
LogBufferPages=0
LogPath=/var/mqm/log/TEST/
LogWriteIntegrity=TripleWrite
Service:
Name=AuthorizationService
EntryPoints=10
ServiceComponent:
Service=AuthorizationService
Name=MQSeries.UNIX.auth.service
Module=/opt/mqm/lib/amqzfu
ComponentDataSize=0
CHANNELS:
StopEvent=ALWAYS
now that I have added it, do i need to restart the qmgr to refresh the settings?
thanks,
dp111443
_________________
Integration Design/Developer
IBM Certified System Administrator -
WebSphere MQ V5.3 |
|
| Back to top |
|
|
| dp111443 |
| Posted: Mon May 24, 2004 6:30 am Post subject: |
|
|
Voyager
Joined: 25 Feb 2004
Posts: 82
|
Jason,
that has worked as I have tested it and am now getting RETRYING for the stopped event reason code.
many thanks,
dp111443
p.s do you see any draw back of doing this and why this was taken out of the original functionality??
_________________
Integration Design/Developer
IBM Certified System Administrator -
WebSphere MQ V5.3 |
|
| Back to top |
|
|
| JasonE |
| Posted: Mon May 24, 2004 8:27 am Post subject: |
|
|
Grand Master
Joined: 03 Nov 2003
Posts: 1220
Location: Hursley
|
| Quote: |
| p.s do you see any draw back of doing this and why this was taken out of the original functionality?? |
No real drawback. It was taken out to resolve another problem where starts/stops didnt match up, you got a stopped but not started event and that also caused problems. Trouble is you cant win sometimes, so we leave both available  |
|
| Back to top |
|
|
| xxx |
| Posted: Mon May 24, 2004 6:33 pm Post subject: |
|
|
Centurion
Joined: 13 Oct 2003
Posts: 137
|
I guess the only problem is when the messages are not transfered ,
so if you monitor the sdr channel indirectly , i.e to monitor the xmitq depth you will have no issues , |
|
| Back to top |
|
|
|
|
