Question

Usage: 2 systems communicating using XML-messages that can be up to 10 MiB. Integration uses Mule ESB 3.3.1 and HornetQ 2.3.0 Finale.

We have problems with HornetQ (running on the same host machine Linux). It seems that Mule as JMS client keeps losing / dropping connections without being able to reconnect, and needs to be rebooted every now and then (couple of times a day up to every 2 days max, typically). At peak times, queues receive (much) more messages than consumers can handle.

Q: If anyone has experience successfully using Mule with HornetQ sending and receiving large XML (text) messages, we would very much appreciate if You could share your “best” configuration profile. Which parameters do you need and how to configure HornetQ (JNDI) to get most reliable and stable messaging environment? Or which (not obvious) traps to avoid? Are there any defaults that HornetQ uses, which should not be used in our kind of usage.

Was it helpful?

Solution

You probably need to tune JVM Settings, and also ping intervals.

This chapter brings some information about dead connection detection and how to configure these parameters:

http://docs.jboss.org/hornetq/2.4.0.Final/docs/user-manual/html/connection-ttl.html

And as I said on the first paragraph, tune JVM and Garbage collection options by setting proper sizes and parallel garbage collection. You can probably find good information on this by googling about it.

Licensed under: CC-BY-SA with attribution
Not affiliated with StackOverflow
scroll top