Re: [jetty-users] ThreadMonitor memory leak

[Greg Wilkins <gregw@xxxxxxxxxxx>]

Carey,

sounds like a possible bug.  Can you  raise a Bugzilla: Jetty so we can put some resourced onto investigating/fixing.

thanks

On 23 December 2015 at 08:30, Carey Evans <carey@xxxxxxxxxxxxx> wrote:

Hi.

We're seeing a slow memory leak on our Jetty 9.2.14 server, which
currently makes heavy use of the JDK's common ForkJoinPool. We've got
the default "monitor" module enabled, and a heap dump after a couple of
days shows that ThreadMonitor._monitorInfo is using 350 MB of heap, with
13230 entries, even though there's only 42 active threads.

Is what I'm looking at a bug? It seems like the ThreadMonitor should be
releasing its reference to old threads soon after they've stopped,
rather than keeping thousands of 32KB ForkJoinWorkerThread instances
alive forever. I guess it's more noticeable with ForkJoinWorkerThread,
since it extends the normal java.lang.Thread and makes it much larger.




--
  Carey Evans
  carey@xxxxxxxxxxxxx
_______________________________________________
jetty-users mailing list
jetty-users@xxxxxxxxxxx
To change your delivery options, retrieve your password, or unsubscribe from this list, visit
https://dev.eclipse.org/mailman/listinfo/jetty-users

--

Greg Wilkins <gregw@xxxxxxxxxxx> CTO http://webtide.com