Shame on me...
I have found the answer to this problem in the first question of the FAQ:
https://www.jboss.org/community/docs/DOC-9239
To achieve failover during a normal shutdown of the server there is an option
in the PostOffice Configuration:
xxx-persistence-service.xml:
.
| .
| .
|
Glad that works for you :)
If you want to be sure you don't receive duplicates just make sure your
consumers use a transacted session to consume - then you should get guaranteed
once and only once delivery as per spec.
Currently you're using a non transacted session so losing messages / receivi
SP2 + the patched jboss remoting gives much better, complete results. I receive
duplicates, but I'm sure the clients can handle that if they're aware that it
can happen. (For anyone lese reading this, see the Clustering Notes at
http://labs.jboss.com/file-access/default/members/jbossmessaging/fr
If you're in a managed environment you can use the JCA JMS resource adapter to
cache connections:
See http://wiki.jboss.org/wiki/Wiki.jsp?page=JBossJMSRA
View the original post :
http://www.jboss.com/index.html?module=bb&op=viewtopic&p=491#491
Reply to the post :
http://www.jboss.com/
One observation:
You're creating a connection, session, and producer for every message that you
forward!
This will be extremely slow...
Connection and sessions are heavyweight objects that you want to create once
then re-use. Really you want to re-use the producer too.
View the original post
Just noticed SP2 released - I'll give it a whirl. In particular, the fix for
http://jira.jboss.com/jira/browse/JBMESSAGING-1184 look promising
View the original post :
http://www.jboss.com/index.html?module=bb&op=viewtopic&p=476#476
Reply to the post :
http://www.jboss.com/index.html?
Hi Tim,
First of all, I have a large portion of humble pie to eat - I hadn't clustered
the SLSB that publishes. However, I'm still seeing unreliable results, so I'll
elaborate on what I am doing...
I have a pojo that connects via JNDI to queue/testQueue on the
ClusteredConnectionFactory.
I'm not sure I understand your topology either.
Can you explain it again, just to avoid ambiguity. Not sure what you meant by
"receiving node".
Is the queue clustered? Is the MDB on each node? Are you using persistent / non
persistent messages etc?
View the original post :
http://www.jboss.co
I'm wondering why durable subscription failover is, in my experience, so flaky.
An obvious reason would be that I've got something wrong, although I don't
believe I have or otherwise I'd go and fix it (!). If I posted some zipped up
code, would anyone in the JBM team be able to give it a sanity
