All - just had loads of notifications from various
company's socalled 'anti-virus' programs.
My attached called overlap.cmd was rejected.
I have, therefore, renamed it to overlap.txt.

Regards ... Ruud

-----Original Message-----
From: Ruud van Zundert [mailto:[EMAIL PROTECTED]]
Sent: 28 May 2002 14:17
To: MQSeries List
Subject: RE: Communication about cluster boundaries


Henry - it's a bit tricky to get this to work, and if you
follow this procedure, you should be ok:

I've attached an example script that you can run on a Win2000 machine.
It is re-runnable. This is what it does (run via 'overlap.cmd') :

1. define 4 qmgrs QMT1, QMT2, QMT2B and QMGW.
   QMT1 holds the full repository for cluster CL001.
   QMT2 holds the full repository for cluster CL002.
   QMT2B holds a partial repository for cluster CL002.
   QMGW is the 'gateway' linking the 2 clusters together.
2. Cluster queue QMT2.CQ is defined on both QMT2 and QMT2B.
3. A 'bridge' queue called BRIDGE is defined on the gateway
   as a QALIAS - it resides in cluster CL001, but points to
   target queue QMT2.CQ in cluster CL002.
4. To test: amqsput BRIDGE QMT1
   Result : messages 'should' ;-) arrive on QMT2.CQ spread
   across qmgrs QMT2 and QMT2B.

Let us know how you get on.

Regards ... Ruud

-----Original Message-----
From: MQSeries List [mailto:[EMAIL PROTECTED]]On Behalf Of
Springer, Henry
Sent: 28 May 2002 11:22
To: [EMAIL PROTECTED]
Subject: Communication about cluster boundaries


Hello all,

we want to know, how to communicate about MQSeries Cluster boundaries.
Because of firewall restrictions we had to build two overlapping cluster and
the need to communicate between two nodes in both cluster. We want to use
workload balancing, the queues reside on several cloned machines in both
cluster.

A small and simple scenario:

QM1  belongs to  Cluster1  and share a local Queue  Q.TEST.ORIG  in his
cluster
QM2  belongs to  Cluster2  want to put a message to Q.TEST.ORIG.
QMR  belongs to  Cluster1,Cluster2  and should act as a MQ-Gateway

We do some tests and no one work properly:
a)      Define QALIAS  Q.TEST  on QMR  shared in Cluster1, Cluster2 and
point to Q.TEST.ORIG in the Cluster1
Result:  Reason 2082 on QMR
b)      Define QREMOTE  Q.TEST.1  on QMR  shared in Cluster2, point to
Q.TEST.ORIG (without naming a remote QMGR)
Result:  Reason: 2087 on QMR

Defining a QREMOTE on QMR  shared in Cluster2 and point to Q.TEST.ORIG on
QM1, QM2 can send messages to the Q, but this is a lost of the workload
balancing feature of the cluster.



Henry Springer
MQSeries Familie
K-DOI-23

Volkswagen AG
Brieffach 1891
38436 Wolfsburg

Telefon +49-53 61-92 97 74
Telefax +49-53 61-92 28 13
eFax +49-5361-95 72 97 74
http://www.volkswagen.de

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
endmqm -i QMT1
endmqm -i QMT2
endmqm -i QMT2B
endmqm -i QMGW
dltmqm QMT1
dltmqm QMT2
dltmqm QMT2B
dltmqm QMGW
crtmqm -u SYSTEM.DEAD.LETTER.QUEUE QMT1
amqmdain crtlsr QMT1 -t TCP -p 6001
amqmdain crtchi QMT1 SYSTEM.CHANNEL.INITQ
amqmdain start QMT1
crtmqm -u SYSTEM.DEAD.LETTER.QUEUE QMT2
amqmdain crtlsr QMT2 -t TCP -p 6002
amqmdain crtchi QMT2 SYSTEM.CHANNEL.INITQ
amqmdain start QMT2
crtmqm -u SYSTEM.DEAD.LETTER.QUEUE QMT2B
amqmdain crtlsr QMT2B -t TCP -p 6012
amqmdain crtchi QMT2B SYSTEM.CHANNEL.INITQ
amqmdain start QMT2B
crtmqm -u SYSTEM.DEAD.LETTER.QUEUE QMGW
amqmdain crtlsr QMGW -t TCP -p 6099
amqmdain crtchi QMGW SYSTEM.CHANNEL.INITQ
amqmdain start QMGW
runmqsc QMT1 < qmt1.txt
runmqsc QMT2 < qmt2.txt
runmqsc QMT2B < qmt2b.txt
runmqsc QMGW < qmgw.txt

Reply via email to