| |
|
RSS Feed - WebSphere MQ Support
|
RSS Feed - Message Broker Support
|
  |
|
| MQ QMGR Cluster on Z/os to MQ QMGR cluster on SUN Solaris |
« View previous topic :: View next topic » |
| Author |
Message
|
| loudcfla |
 Posted: Tue Jul 13, 2004 9:54 am Post subject: MQ QMGR Cluster on Z/os to MQ QMGR cluster on SUN Solaris |
 |
|
Acolyte
Joined: 05 May 2002
Posts: 58
|
Hello, we currently have a simple point to point MQ config. A QMGR on SUN talking back and forth to a QMGR on Z/os with CICS adapter in a single CICS region.
Mgmt now wants no single point of failure, so how do I set up a cluster on Z/os with the CICS adapter (can the INITQ and BRIDGE.QUEUE) be serviced by a cluster; and how do I send messages back and forth to a cluster of QMGRs on the SUN.
I would like to use single queues managed by the QMGRs in the cluster, since we process the same type of message, I really need a group of QMGR clones to service the queue, sort of the same way we do it now only I don't have to worry if I loose a QMGR, which is really the issue here. |
|
| Back to top |
|
 |
| jefflowrey |
| Posted: Tue Jul 13, 2004 10:20 am Post subject: |
|
|
Grand Poobah
Joined: 16 Oct 2002
Posts: 19981
|
If management wants no single-point of failure, then management will pay for the following:
1) redundant Sun environments, with Veritas clustering to manage failover between Sun servers
2) sufficient resources on z/OS for multiple LPARs, and any needed hardware for Queue Sharing
Anything else won't really give you high availability.
Simple WebSphere MQ clustering will not meet management's needs, at least that is what they will tell you when one important message is stuck on a failed queue manager.
WebSphere MQ Clustering is not a failover or redundancy solution. It's a workload balancing and administration solution.
_________________
I am *not* the model of the modern major general. |
|
| Back to top |
|
|
| loudcfla |
| Posted: Tue Jul 13, 2004 12:42 pm Post subject: |
|
|
Acolyte
Joined: 05 May 2002
Posts: 58
|
Excellent response...
Okay..let's assume I'm starting from the ground zero, what would constitute a basic high availability - no single point of failure MQ config based on our scenario, i.e. SUN solaris to mainframe/CICS and back.
By the way, if we lost a QMGR insied the cluster, do we loose the whole cluster? |
|
| Back to top |
|
|
| offshore |
| Posted: Wed Jul 14, 2004 9:45 am Post subject: |
|
|
Master
Joined: 20 Jun 2002
Posts: 222
|
loudcfla,
To answer you question about if you loose a MQ cluster member does the cluster go down, the answer is NO.
About setting up true no single point of failure. There are alot of details but I'll tried to provide an overview as best I can. I'll be more detailed on the z/OS side as that is what I know best.
You will need(as Jeff previously mentioned):
1.] Dual Sun servers with Veritas clustering to manage fail over.
2.] z/OS parallel sysplex; some of which includes setting up a coupling facility, VIPA IPs, VTAM generic resources, multiple LPARs (or true sysplex failover have an additional mainframe) - it's could be possible to lose the actual machine and not just an LPAR.
3.] Define MQ Series to use a Queue Sharing Group.
4.] Multiple CICS regions to handle the transaction.
5.] Does the CICS interface w/ a DB2 region? If so you need configure DB2 data sharing or depending on how you configure MQ QSG you may need DB2 data-sharing.
FYI: Using QSG you have a 63k size limit on messages due to the CF.
I'd suggest reading the IBM Redbook: WebSphere MQ in a z/OS Parallel Sysplex Environment
I can give you some more info offline (PM me) on what I did to try to setup for no point of failure. It's not perfect, but pretty close.
Offshore |
|
| Back to top |
|
|
|
|
|
| |
|
Page 1 of 1 |
|
You cannot post new topics in this forum
You cannot reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum
You cannot vote in polls in this forum
|
|
|
|
