I want to refresh this thread since it was posted over the break...
My situation is that I have a server I've let sit idle for 13 days now...and it used all available memory on my system and has not changed for the entire time.
java.util.concurrent.ConcurrentLinkedQueue$Node is the object that will not clean up.
I have run jmap and captured histograms of the live java objects and this object still has 1,573,474 new instances. (There are no active sessions in glassfish as checked with asadmin.)
How and when is glassfish using this class? It will help me track down this memory leak in glassfish?
Anyone have ideas for a good dtrace/btrace script to run while I have the server in this memory-depleted state?
[Message sent by forum member 'rpetruzzelli' (rpetruzzelli)]
http://forums.java.net/jive/thread.jspa?messageID=324100