> We would just like for someone to confirm that there is a bug present
here.
Yes, I believe there is a bug here. It's impossible to see directly from
the thread dump due to the involvement of the ordered executor, but
following the flow of the source code it's relatively apparent.
Your analysis
So we're repackaging com.sun.xml classes in the ActiveMQ JARs? Are you
using activemq-all.jar by any chance? If so, please switch to using the
individual JARs and see if that solves the problem; activemq-all.jar is
intended as a quick-start for standalone use, but anytime that users find
that they
Hello Justin,
We managed to get thread dump of the service while issue is in effect,
relevant thread stacks are below.
These thread stacks were taken with artemis version *2.6.4*, we also
managed to reproduce the issue locally on latest version of artemis (
*2.10.1*).
Stacks of 2 *deadlocked *thr
Hello Tim,
What I meant wrt to 7/8 broker is as follows:
1. I have one frontend broker which only have topics. These topics are then
subscribed by the backend broker
2. Each backend broker can have many routes, which means that a single
backend broker(with n number of queues) can subscribe to a
