It's a really weird problem, and we're not able to reproduce it so far.
Couple days ago, we found that after a time point, let's say, 20:00, all http session destroy events were not logged.
It seemed many http sessions started without ended.
According to log, our custom module on wowza experienced an IO exception (Too many open files) that day, about six hours earlier than this weird problem happened.
java.io.IOException: Too many open files
at sun.nio.ch.ServerSocketChannelImpl.accept0(Native Method)
at org.apache.mina.transport.socket.nio.SocketAcceptor$Worker.processSessions(Unknown Source)
at org.apache.mina.transport.socket.nio.SocketAcceptor$Worker.run(Unknown Source)
at org.apache.mina.util.NamePreservingRunnable.run(Unknown Source)
Could the exception be the cause?
We would want to know if there's any chance that the timeout-timer for http session crashes and therefore prevents http sessions from being destroyed because of some programmatic error?
Thanks in advance,