Tim Zielke
2014-06-23 19:50:50 UTC
Hello,
I noticed something odd when working with my Sandbox servers and validating 7.5.0.2 and 7.5.0.3. We have two sandbox servers. One is a SLES 11 server. The other is a Solaris 10 server. Each has one queue manager which is a full repository and participate in one cluster called SBOX.
I was playing around with clustering on the Linux one, and I noticed the following:
If I created a cluster queue in SBOX on the Solaris server, and then tried to PUT to the queue from the Linux server, I would get a 2085. The Solaris full repository was not communicating to the Linux full repository that the new cluster queue existed. If I then did a REFRESH CLUSTER on the Linux queue manager, the Linux queue manager would get the update and I could successfully PUT to the new queue.
When I then changed the Linux queue manager from being a full to a partial repository, this behavior did not exist. I created a cluster queue in SBOX on the Solaris queue manager, and the Linux queue manager knew right away about the new queue. The first PUT to the cluster queue on the Linux server was successful.
I then changed the Linux queue manager back to a full repository, created a new cluster queue on the Solaris queue manager, and again I get the 2085 behavior when doing the PUT on the Linux server.
This looks like a bug to me. But we would never see this in practice as we use our full repositories solely to manage clustering, and do not put any cluster queues or application work directly on them. I just happened to run across it because of our Sandbox set up. But then it made me wonder, is it ok to put cluster queues on full repositories? I am thinking yes and the manual seems to imply that is ok.
Does anyone think this is bad configuration on my part?
Or does this look like an MQ clustering bug?
Thanks,
Tim
To unsubscribe, write to LISTSERV-0lvw86wZMd9k/bWDasg6f+***@public.gmane.org and,
in the message body (not the subject), write: SIGNOFF MQSERIES
Instructions for managing your mailing list subscription are provided in
the Listserv General Users Guide available at http://www.lsoft.com
Archive: http://listserv.meduniwien.ac.at/archives/mqser-l.html
I noticed something odd when working with my Sandbox servers and validating 7.5.0.2 and 7.5.0.3. We have two sandbox servers. One is a SLES 11 server. The other is a Solaris 10 server. Each has one queue manager which is a full repository and participate in one cluster called SBOX.
I was playing around with clustering on the Linux one, and I noticed the following:
If I created a cluster queue in SBOX on the Solaris server, and then tried to PUT to the queue from the Linux server, I would get a 2085. The Solaris full repository was not communicating to the Linux full repository that the new cluster queue existed. If I then did a REFRESH CLUSTER on the Linux queue manager, the Linux queue manager would get the update and I could successfully PUT to the new queue.
When I then changed the Linux queue manager from being a full to a partial repository, this behavior did not exist. I created a cluster queue in SBOX on the Solaris queue manager, and the Linux queue manager knew right away about the new queue. The first PUT to the cluster queue on the Linux server was successful.
I then changed the Linux queue manager back to a full repository, created a new cluster queue on the Solaris queue manager, and again I get the 2085 behavior when doing the PUT on the Linux server.
This looks like a bug to me. But we would never see this in practice as we use our full repositories solely to manage clustering, and do not put any cluster queues or application work directly on them. I just happened to run across it because of our Sandbox set up. But then it made me wonder, is it ok to put cluster queues on full repositories? I am thinking yes and the manual seems to imply that is ok.
Does anyone think this is bad configuration on my part?
Or does this look like an MQ clustering bug?
Thanks,
Tim
To unsubscribe, write to LISTSERV-0lvw86wZMd9k/bWDasg6f+***@public.gmane.org and,
in the message body (not the subject), write: SIGNOFF MQSERIES
Instructions for managing your mailing list subscription are provided in
the Listserv General Users Guide available at http://www.lsoft.com
Archive: http://listserv.meduniwien.ac.at/archives/mqser-l.html