There's definitely something there using more memory than you want; then
at each GC it cleans up just enough to limp along to the next full gc
(but not enough to do anything useful).
You'll have to look into some memory tools (jmap to generate a heap
dump, and then something like jhat or the Netbeans memory tool or the
Eclipse memory analyzer) to see what objects are using all that space
and see why they aren't getting reclaimed.
-Scott
On Wed, 2009-04-08 at 19:05 -0700, glassfish_at_javadesktop.org wrote:
> Hi Scott,
>
> See my most recent reply at the bottom. It does seem to be a GC issue, and I attached a couple of gc logs from the most recent occurrences. Unfortunately, I'm not sure it's the Hibernate cache, since we do not have any second-level cache enabled. Or, could it be elsewhere in Hibrnate? It seems to me from the gc logs that something is stuck, since full GC's are running continuously when the problem occurs. It's as if they are running, but failing, and so running over and over again. Where might I be able to look to try to find the problem?
>
> Thanks for your help once again!
> [Message sent by forum member 'rwillie6' (rwillie6)]
>
> http://forums.java.net/jive/thread.jspa?messageID=341268
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: users-unsubscribe_at_glassfish.dev.java.net
> For additional commands, e-mail: users-help_at_glassfish.dev.java.net