The typical reason is that some persistent messages are being left in an infrequently used/consumed queue, and this holds up the cleanup of the older journal files. I would check all of your persistent queue's and make sure you don't have any old messages that need to be consumed.
As a test, it would also be worthwhile starting with an empty /data directory and then pushing a couple of hundred thousand messages through the broker. You should see that all messages sent are consumed and the journal files are deleted..
Also, post your persistence adapter configuration on the offchance there is somehting misconfigured.
Hope this helps,
Thanks for your response.
I have colleagues running the exact same messages through their topics, and they aren't seeing the same issues. We even have the same configuration files in place to make sure everything is the same.
I have checked all the topics/queues (we don't have that many) and all of them have consumed all messages, infact doing this relatively quickly after the messages are posted.
My persistence adapter config is justL
This is stumping me, as all the usual things seem to be fine, and there really don't seem to be any unconsumed messages anywhere.
I have tried this in activemq also, and the same thing occurs, so it isn't fuse specific.
When I have some spare time, I will try and get this issue occuring for me in a cut down example (if I can!), which will obviously help track it down.
With respect to a test case, some variant of the existing DataFileNotDeletedTest junit test case may help you. see: http://svn.apache.org/viewvc/activemq/trunk/activemq-core/src/test/java/org/apache/activemq/bugs/DataFileNotDeletedTest.java?view=markup