This is certainly becoming an interesting discussion! :) I guess I didn't realize
that using an MDB would really change the semantics of how the JMS consumer (MDB)
functions. Is my understanding correct that if the MDB throws an exception, the container
does not ask for the message again, but simply retries the MDB? Like a for loop with a
built-in number of iterations and hard-coded to send to the DLQ destination afterwards?
I agree that it would be really nice if the container could delegate those actions to the
JMS container somehow, or maybe read in the JMS configs. Like Weston said, it probably
doesn't matter how it gets done, but it would be nice as a customer to take advantage
of the JBM feature set.
So, how would you (tim) recommend implementing a JMS consumer without MDB's inside of
a web app? The web app is both sending and receiving the messages.
View the original post :
http://www.jboss.com/index.html?module=bb&op=viewtopic&p=4043029#...
Reply to the post :
http://www.jboss.com/index.html?module=bb&op=posting&mode=reply&a...