Consumer Dispatch Timeout on Failed Connection Question

[rarciere <ro...@peersoftware.com>]

I am sending a message to a unique single consumer and after some timeout
period expires my application attempts to reclaim the message via a message
selector utilizing a unique message property.  This generally works very
well.  However, I noticed unwanted behavior when the Consumers network
connection to the broker is severed (I am utilizing a single host
FailoverTransport to retry indefinitely), any new message sent to the
consumer will be stuck in that consumer’s dispatch queue until that consumer
is purged/removed by the Broker, which usually happens after a timeout of
around 20 seconds.  During the time that the message is in the primary
consumers dispatch queue, the message cannot be reclaimed/consumed by
another consumer since it is already dispatched to the disconnected
consumer, however, the message can be consumed once the stale disconnected
consumer reference is removed by the Broker.  My primary question is what
timeout value controls the approximate 20 second delay that I am observing,
e.g. some kind of ACK timeout, or is it something related to a failed
consumer connection auto closing? 

Thanks,

Rob Arciere



--
View this message in context: http://activemq.2283324.n4.nabble.com/Consumer-Dispatch-Timeout-on-Failed-Connection-Question-tp4685385.html
Sent from the ActiveMQ - User mailing list archive at Nabble.com.