Mark Thomas schrieb am 19.12.2010 um 17:43 (+0000): > On 17/12/2010 12:52, K J.Sreekumar wrote: > > > > This is the problem - > > > > 1. Our tomcat 6 server (version 6.0.29) does not respond > > after working for a couple of days. Browsers are unable to > > find application URLs (which were previously processed > > normally). > > 2. The logs does not show any error. > > 3. There are only a total of 50-60 threads in the thread > > dump, mostly in WAIT state. Also, it does not show any > > BLOCKING or LOCKED threads. > > 4. Memory utilization was 400MB only > > 5. TOMCAT restart does not fix the problem. The machine needs > > to be restarted to resume normal operation. > > 6. Netstat did not reveal any extraneous connections
> > Have anybody here faced similar issue before ? > > Frequently. > > Usually a mis-match between the number of threads configured in > httpd and the number of threads configured in Tomcat. Since mod_jk > and mod_proxy use persistent connections by default and the BIO > connector uses one thread per connection then if httpd has more > threads configured you can eventually reach a state where all > Tomcat's threads have been allocated and the next request blocks. Ah. Did you infer that from the OP's description of most threads being in the WAITING state? http://download.oracle.com/javase/1.5.0/docs/api/java/lang/Thread.State.html > There are two possible approaches. > 1. Tune the time-outs > 2. Disable connection re-use > > I'd go with 2 - it takes much less time to set up. Ignore the dire > warnings in the docs. Most users don't notice the difference. Here's another take on the same matter from two years ago: Re: mod_jk -> ajp hangs indefinitely on some requests - Rainer Jung http://www.mail-archive.com/users@tomcat.apache.org/msg54056.html -- Michael Ludwig --------------------------------------------------------------------- To unsubscribe, e-mail: users-unsubscr...@tomcat.apache.org For additional commands, e-mail: users-h...@tomcat.apache.org
