QueueBrowser only sees local queue when connected to cluster?
I think I've come across a problem with 3.5 SP1. One of our applications uses QueueBrowser to view the contents of a queue before deciding what to do with the messages.
Assume we have a two-machine cluster where A is the master and B is a regular node. Queue X has messages L and M on server A, and messages N and P on B.
If I attach a consumer to either server, I'll see messages L,M, N, and P as expected for a cluster. However if I attach a QueueBrowser to server A, then I only see mesages L and M. Attaching to server B only shows messages N and P.
Is this behavior by design? Our applications are designed to expect that the contents of each node's queue is combined to create a "meta-queue" for entire cluster when consuming, producing or browsing.
Please clarify if this behavior is a bug or by design. Thanks.
[860 byte] By [
guy_davis] at [2007-11-25 17:09:01]

# 1
Can u please confirm whether the destinations which u have created are administratively created or autocreated.
Secondly please also check the attribute isLocalOnly for the destination. If it is set to true in that case your local consumers will only receive the messages and the detinations will not be replicated in a cluster.
You can use the following command to query a particular broker.
imqcmd query dst -t q -n QueueX
This attribute cannot be updated once your destination is created. Hence if you have set it to true
you'll have to delete the destination and create it again with the correct isLocalOnly attribute.
If it is an autocreted destination in that case check out table 2.10 in the admin guide and you'll see that in case u autocreate destinations there are properties in the broker config.properties that needs to be set. One of the properties is as follows
imq.autocreate.destination.isLocalOnly
thanks
Vibhor
# 2
This is the answer that was posted by one of the MQ Engineers please consider this to be the appropriate
answer to your query.
It's by design. This is the same behavior some of our competitors have. The JMS spec does not specify this in case of cluster. We may
have not documented this in our docs.