I believe the cause of the EOFException and the blocking behaviour and slowdown is caused by creating and closing connections for each read/write. In the standalone j2se application I attached in an earlier post, each thread creates a new connection for each request before closing down the resources once it has written to or consumed from the queue. This produces lots of EOFExceptions and the blocking/slowdown behaviour I outlined earlier. If I modify the class to ensure that a single connection is used for all requests from the producer thread and a single connection is shared for all the requests from the consumer thread, this stops the slowdown and the EOFExceptions are no longer thrown. Unfortunately, this is not a viable approach for us, as our application uses stateless session beans, which may be persisted by the container, therefore a new connection to the broker is created and closed once used for each request.
My question now is, what could be causing this delay/problem with opening and closing lots of connections to the broker? Does ActiveMQ maintain a connection pool internally - could it be maintenance of this pool causing the delay? Is there some configuration that can be made to the broker which may impact the behaviour of the connection handling? -- View this message in context: http://activemq.2283324.n4.nabble.com/Slow-throughput-after-several-hundred-messages-tp3082431p3089010.html Sent from the ActiveMQ - User mailing list archive at Nabble.com.
