Current Thread: qtp1239170060-53 Acceptor0 SelectChannelConnector@0.0.0.0:8000
2012-07-16 23:33:29 [SEVERE] PID: 53 | Alive: true | State: BLOCKED
I feel I only bug you when something is wrong....
Using latest version of CB (rb) and Dynmapp we saw this hard freeze (looks to be caused by Dynmap.)
This is the error what makes me believe is dynmap - it signifies the port 0.0.0.0:8000 (our dynmap port)
These aren't errors - its a thread dump of the currently active threads. We always have some threads around, so they will be there, of course. I don't actually see anything that looks unusual. The thread you mentioned in the last image is the listener thread for the internal web server - its blocked on accept() because that is how the thread waits for connections to the web server's socket.
Several other users (server owners) state they are getting this same exact thing. Temp removing dynmap has resolved this issue. It feels like we're somehow getting ddos'd through dynmap and locking up the server, though I cannot say for sure.
I could be wrong and if so, it may be dynmap clashing with another plugin or cb change.
Some evidence that actually supports the notion that dynmap is the cause would be nice - the dump you provided shows nothing of the sort, so I've got nothing to go on. Given the thousands of servers running this with no reported issues (including some of the biggest and busiest servers out there), I'm not inclined to spend too much time on something without supporting evidence (like an exception or sign of resource use by us). A thread dump showing that we have threads active, given that we ALWAYS have threads active, is meaningless. And, looking at the dumps, our threads are in a perfectly normal state - your hangs look to be in the CB code (CB++, if I read this right)
Do you mind posting the following:
What happens when you remove Spout (and all plugins that require it) and start the server?