On 23/03/17 05:07, Igal @ Lucee.org wrote: > Hi Mark, > > On 3/22/2017 1:00 PM, Mark Thomas wrote: >> On 22/03/17 16:20, Igal @ Lucee.org wrote: >>> Looks like all the threads are waiting on the same lock? >> Agreed. But I don't see anything holding that lock. That looks like a >> JVM bug to me. > > Thank you for replying. I managed to "do" it again. This time on > Tomcat 8.5.11 so I modified the subject line slightly. But this time I > took both a thread dump (attached) and a heap dump. > > I should mention that my code utilizes JSR-356 WebSockets, I'm not sure > if that's related, but it seems to be a factor in reproducing the issue. > > I will try to create a reduced test case. > > I am hoping that the screenshots below, showing references to and from > the lock object at 0x00000006c0dbc1b8, can give you some more > information that may help to diagnose this issue, or at least give some > ideas about what to look for next:
I took another look and those threads are in the normal state you'd expect for threads in the pool waiting for work. If Tomcat is unresponsive, that would suggest a problem with either the Acceptor or the Poller but both of those threads look normal to. You are going to need to do some more debugging. Ideally, you need to configure your favourite IDE to do remote debugging and then once the problem occurs, step through the code as a single connection is handled and see where things go wrong. Based on what I see in the thread dumps (all normal) I'd also check that your test is doing what it is meant to. Mark --------------------------------------------------------------------- To unsubscribe, e-mail: users-unsubscr...@tomcat.apache.org For additional commands, e-mail: users-h...@tomcat.apache.org
