Please post a test case + JIRA and someone will investigate
Also, please specify what version you are using? Have you verified if the behaviour occurs on trunk?
today I downloaded hornetq svn trunk && new version of netty 3.1.5GA
And it seems, that the problem has disappeared.
At least now I'm not able to reproduce that issue again using new compiled hornetQ.
But still I have one question:
Is there some way to enlist (for example through jmx)
which connections are 'listening' (consuming messages) to particular queue?
Thank you very much.