Further experimentation has ensued and here are my results.
I was using the free edition of PC-Tools Firewall Plus on Windows 2003
Server Enterprise, R2.
I figured that if I blocked the JAVA process, that would completely disable
ActiveMQ from talking to other instances. This was not the case, the
multicasts where still going out, which caused confusion because the
multicasts said the blocked ActiveMQ was alive but it could not be reached.
I created 2 advanced rules, one to block the multicast and one to block
normal connections, which created the expected disconnection scenario.
The two now seem to disconnect and re-connect nicely, although, I did notice
the transport connector cleaned up quite quickly, but the subscriber
channels took a while longer - maybe it's just my lack of understanding -
maybe the channels can have multiple transports, hence a different cleanup
cycle.
Thanks to those who have offered assistance.
...Lyall
--
View this message in context:
http://www.nabble.com/Durable-Subscriber-and-unclean-disconnection-tp23591185p23627612.html
Sent from the ActiveMQ - User mailing list archive at Nabble.com.
