I will try to increase the logging level and leave it over the weekend, along with the ring queue config changes and new dropped connection handling logic.
I have a script that parses bits of /proc to get the memory usage of the various processes. Nothing appeared to be misbehaving. The qpidd process was reporting about 8Mb of VmRSS when this failure was going on. The only obvious things wrong with qpidd were general network instability and the large number of bindings and subscriptions. Interestingly, if I killed the qpid daemon, it didn't clear the funny network state of the box. I had to shut down all the processes using qpid for it to clear. When this happens the syslog is clear apart from the qpid disconnection messages. If it was a kernel driver or hardware issue with the network interface then I think it would be reporting there. This is why I'm thinking the network problem is somehow related to qpid or my use of it, rather than the other way round. Again, many thanks for your help. -- View this message in context: http://qpid.2158936.n2.nabble.com/Newbie-problem-with-long-term-use-of-C-broker-client-code-tp7591679p7591718.html Sent from the Apache Qpid users mailing list archive at Nabble.com. --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
