While I appreciate the need to have QM naming convention reflect the underlying platform, I think it makes more sense to keep QM names the same for dev test and prod environments. This is for 2 basic reasons:
1] Many MQ admins rely on script generators such as SupportPac MS03 to save/clone an MQ config for promotion to production, and if the QM names are different, it means editing the files and whatever objects were dependent on the QM name.
2] Some MQ apps do NOT want to use the local default QM for connection, or require a specific destination QM in the MQOPEN for reasons such as "forced" loadbalancing, hence different QM names between environments means changes to code.
Although 2] could (and should) be helped by externalizing all QM names/vars in property/profile files, it's a lot easier when you don't have the worry of making QM name changes when promoting to production. OTOH, if your environment is clustered, this makes having different QM names much less of an issue.
I've found it boils down to who dominates in your organization: The hardware/infrastructure, or the architect/design/development teams.
___________________________
Steve Sacho
CSSD - Enterprise Architecture
Tel: 805-577-3983
Internal: 92-598-3983
| "Boger, Dan" <[EMAIL PROTECTED]>
Sent by: "MQSeries List" <[EMAIL PROTECTED]> 06/24/2002 08:26 AM
|
To: [EMAIL PROTECTED] cc: Subject: Re: QMGR names when using MQSI |
> I'm having a disagreement with a colleague on QMGR names when using MQSI
on
> WIN/NT machines and I'm looking for some confirmation or explanation. My
> colleague contends that IBM has a requirement/restriction for all the
QMGRs
> supporting MQSI on his Development, Test and Production to have the same
> name in order for him to move his code between these MQSI machines. I'm
not
> an MQSI person, but to me, this doesn't make any sense. I'm on the OS/390
> side of the house and I'm saying that QMGR names should all have unique
> names within the environment. Can anyone help?
For what it's worth, I agree with you. We keep all the QMGRs name unique
within our environment. However, that doesn't mean you have to edit your
flows when you move from one to the other. You can leave the QMGR name
blank in the output node of an MQSI flow, which will resolve as the local
QMGR. This means you can have remote queue definitions on the dev broker,
pointing to dev servers, which have the same queue name pointing to test
servers on a test broker.
Hope that helps!
Dan
Instructions for managing your mailing list subscription are provided in
the Listserv General Users Guide available at http://www.lsoft.com
Archive: http://vm.akh-wien.ac.at/MQSeries.archive
