Hello, I am investigating a problem on z/OS where the socket accept loop in TcpTransportServer can indefinitelly spin the CPU at 100% and churn out large amounts of messages. I have opened the issue AMQ-6937 for this. While working on this issue, I have discovered that the problem can happen on Linux as well, when you run out of file descriptors you will get the same tight loop. I have put together a patch (attached to AMQ-6937) that slows down exception handling so that these situations do not consume as much CPU. Can somebody help me understand what do I need to do to make this patch added to an ActiveMQ release?
Thanks, Tomas -- Sent from: http://activemq.2283324.n4.nabble.com/ActiveMQ-Dev-f2368404.html
