post a thread dump of the broker and consumers when it ocurrs again, it may be possible to infer what is going on from that. Or if you can code up a junit test case that can reproduce, fantastic. Also, is there anything relevant in the broker logs, enabling full trace level logging will give loads of output, but it may have a relevant nugget that will set you in the right direction.
On 27 September 2012 18:45, Michael Black <[email protected]> wrote: > I've run into what seems like a simple enough situation that is locking up > 5.6.0 activemq. > > > > I have several queues being written to by single processes. > > > > I have a consumer reading those queues. Everything has been running fine > for months. > > > > Recently another developer on the project started working on the consumer > part. > > > > What happens though is that we can end up with 2 consumers running at once. > > > > Sometime during this process activemq stops responding. The producers are > sending data but the data doesn't show up in the queues. > > > > Restarting activemq solves the problem. > > > > Can anybody enlighten me as to what might be happening and how to prevent > this? > -- http://fusesource.com http://blog.garytully.com
