[prev in list] [next in list] [prev in thread] [next in thread]
List: veritas-ha
Subject: QRE: [Veritas-ha] MQCluster question
From: "Cronin, John S" <John.Cronin () bellsouth ! com>
Date: 2005-05-05 12:43:01
Message-ID: AC8DA4BA69EE3445ADD3C6DDBE1BD7EE016C2E56 () 01al10015010147
[Download RAW message or body]
I am not sure what your expected configuration is. If you EVER plan to
have more than one queue manager running in your cluster, it is better
to fail things over at the queue manager level rather than for MQ as a
whole.
Since MQ is poorly designed from a system install and management
perspective, you CANNOT do things like specify where logs and data will
go, or install the binaries into a location other than the default; this
last requirement prevents running more than one version of MQ in a
single cluster, which is very annoying.
Perhaps in the near future IBM will rewrite MQ so that it is better
behaved and more configurable, or maybe system virtualization techniques
will allow each MQ instance to have it's own virtual machine, but in the
interim, we have to work around MQ's limitations using symlinks.
Hopefully you are using VCS 3.5 or later on Unix, in which case you
should be using the VCS Agent 3.5 for WebSphere MQ. On page 13 of the
Installation and Configuration Guide for this agent, there is a section
about "Active-Active Configuration". Basically, you move the
queuemanager specific data to filesystems on shared disks, and then
create the appropriate symbolic links to those filesystems from the
default locations in /var/mqm that MQ expects to use. Here is a link to
the VCS Agent 3.5 for WebSphere MQ Installation and Configuration Guide
on the VERITAS support site:
http://ftp.support.veritas.com/pub/support/products/ClusterServer_UNIX/2
73084.pdf
IBM also provides similar instructions as part of the MC6A SupportPac.
I recommend using the VERITAS MQ agent instead of the IBM SupportPac,
because the IBM SupportPac is only tested at all on Solaris, it is old
(Solaris 7 and VCS 1.3.1) and provided "as-is" without any promise of
support. That said, there is useful information, and it doesn't hurt to
look at the scripts and agents supplied, and it does not look like it
would be hard to port this to other flavors of Unix (eg Linux, AIX and
HP/UX). Here is a link to SupportPac MC6A and the documentation for it:
http://www-1.ibm.com/support/docview.wss?rs=171&uid=swg24000950&loc=en_U
S&cs=utf-8&lang=en
ftp://ftp.software.ibm.com/software/integration/support/supportpacs/indi
vidual/mc6a.pdf
NOTE: The first link above also includes IBM SupportPac IC63, which is
used for clustering WebSphere MQ Business Integration Brokers using VCS
on Solaris. Each broker requires an MQ queue manager, which is why
these are grouped together.
I just used the Application agent to cluster the Broker components
rather than using the IBM SupportPac, but again the IBM SupportPac can
be useful to look at when figuring out the best way to do things.
There are also some SupportPacs for doing this on AIX using IBM HACMP
that were less useful to me, but I mention them in case others want to
look at everything that is available. The specific SupportPacs are IC61
and MC63:
http://www-1.ibm.com/support/docview.wss?rs=171&uid=swg24000105&loc=en_U
S&cs=utf-8&lang=en
http://www-1.ibm.com/support/docview.wss?rs=171&uid=swg24006416&loc=en_U
S&cs=utf-8&lang=en
--
John Cronin
ipager: johncronin
Desk Phone: 404-986-3992
Cell Phone: 678-480-6266
-----Original Message-----
From: veritas-ha-admin@mailman.eng.auburn.edu
[mailto:veritas-ha-admin@mailman.eng.auburn.edu] On Behalf Of Brown,
Rodrick
Sent: Wednesday, May 04, 2005 3:37 PM
To: veritas-ha@mailman.eng.auburn.edu
Subject: [Veritas-ha] MQCluster question
Just some design considerations with MQ Clustering and file system
layout If I have the following partition layout
/opt/mqm 200MB
/var/mqm 2GB
/var/mqm/errors 4GB
/var/mqm/log 4GB
Does it make sense to put the /opt/mqm partition on a shared san volume?
So I can import it to another system when I need to failover? Or should
I leave the /opt/mqm on dedicated internal drives? So basically each
node will have an independent installation/configuration and basically I
will only failover the /var/mqm /var/mqm/errors and /var/mqm/log
directories.
I'm not really familiar with how MQ works so I would like to eliminate
as much trial and error as possible can anyone advise, thanks.
--
Rodrick R. Brown
The City of New York (DoITT)
Unix Shared Services
_______________________________________________
Veritas-ha maillist - Veritas-ha@mailman.eng.auburn.edu
http://mailman.eng.auburn.edu/mailman/listinfo/veritas-ha
*****
The information transmitted is intended only for the person or entity to which it is \
addressed and may contain confidential, proprietary, and/or privileged material. Any \
review, retransmission, dissemination or other use of, or taking of any action in \
reliance upon this information by persons or entities other than the intended \
recipient is prohibited. If you received this in error, please contact the sender and \
delete the material from all computers. 162
_______________________________________________
Veritas-ha maillist - Veritas-ha@mailman.eng.auburn.edu
http://mailman.eng.auburn.edu/mailman/listinfo/veritas-ha
[prev in list] [next in list] [prev in thread] [next in thread]
Configure |
About |
News |
Add a list |
Sponsored by KoreLogic