On 10/10/2012 19:30, Kirill Kireyev wrote: > Thanks Mark! > > So currently ContainerBackgroundProcessor is sleeping: > 0x5f31 0 27.4 > "ContainerBackgroundProcessor[StandardEngine[Catalina]]" daemon prio=10 > tid=0x000000000306c800 nid=0x5f31 sleeping[0x00007fa2e9dcb000] > > What should I be looking for in terms of that thread?
Look at it over time. Make sure it isn't deadlocked. That is the thread that does session expiration. I have seen deadlocks stop session expiration. The other check is to attach a debugger and put a break point in org.apache.catalina.session.ManagerBase.backgroundProcess() and then check what is going on. Mark > > -Kirill > > > On 10/10/12 12:27 PM, Mark Thomas wrote: >> On 10/10/2012 19:24, Kirill Kireyev wrote: >>> Thanks Shanti! >>> >>> I think your hypothesis is that these sessions remain active, due to >>> client-side polling. However, I don't think that's the case - these >>> sessions are listed as *inactive* (or rather inactive for longer than 20 >>> mins) in the Tomcat manager web application. And hitting "Expire >>> (sessions > 20mins)" causes them to expire! So they are inactive >>> sessions, but are somehow not expired automatically. >> Again: >> >> Thread dumps. >> >> Look at the ContainerBackgroundProcessor thread. >> >> If you continue to ignore folks giving you advice don't expect the >> advice to continue. >> >> Mark >> >> --------------------------------------------------------------------- >> To unsubscribe, e-mail: users-unsubscr...@tomcat.apache.org >> For additional commands, e-mail: users-h...@tomcat.apache.org >> > > > -- > *Kirill Kireyev, PhD* > Founder/CTO instaGrok.com <http://www.instagrok.com> > kir...@instagrok.com > Twitter: @instaGrok <http://twitter.com/InstaGrok> > FB: facebook.com/instagrok <http://www.facebook.com/instagrok> > <http://www.instagrok.com> --------------------------------------------------------------------- To unsubscribe, e-mail: users-unsubscr...@tomcat.apache.org For additional commands, e-mail: users-h...@tomcat.apache.org
