see what happens when friends let friends use outlook? ;-)
(don't worry, i have trimmed the original considerably to make my humor
less list-disruptive)
On Feb 6, 2005, at 1:55 AM, <[EMAIL PROTECTED]> wrote:
Thank you for emailing Blizzard Entertainments Technical Support
Department. In order to provide you with greater assistance, we have
developed an automated reply system which evaluates your message and
generates a detailed response towards it.
Please read through this message as our automated reply system has
helped many of our customers. If the response given below does not
assist you or it is not relevant towards your message, please reply
back to this email and a live technician will respond to you as soon
as possible.
==
Support Information Regarding - C005:
==
[ *** EDITED FOR BREVITY *** ]
Again, if the response given above has not assisted you or it was not
relevant towards your message, please reply back to this email and a
live technician will respond to you as soon as possible.
Best Regards,
Blizzard Support Team
http://www.blizzard.com/support
Blizzard Entertainment
If you respond to this email, please attach all previous messages and
files relating to this issue.
-Original Message-
From: tomcat-user@jakarta.apache.org
Sent: 2/5/2005 5:25:51 AM
To: tomcat-user@jakarta.apache.org
Subject: Diagnosing tomcat problems
Hello
I'm running a tomcat 5.5.4 server with sun jdk1.5 update 1, on a RH
enterprise linux 3.0 (with 2.4.21 kernel). The server processes
about 20-30 requests per second, most of which include a database
query (to a postgresql DB). The configuration I'm using is:
JAVA_OPTS='-Dbuild.compiler.emacs=true -Xmx240m -Xms240m -Xmn60m
-server -XX:+UseParallelGC'
LD_ASSUME_KERNEL=2.2.5
(the problems occur with or without the -XX:+UseParallelGC switch)
The problem I have is that approximately every 3 days the server
stops processing requests. The only trace that is left in the logs
is the following exception (which seems to be totally non-related but
for some reason always occurs on such a hang):
java.lang.IllegalStateException: Cannot create a session after the
response has been committed
at
org.apache.catalina.connector.Request.doGetSession(Request.java:2190)
at
org.apache.catalina.connector.Request.getSession(Request.java:2012)
And my automatic script for checking if the server is alive tries to
restart it (with shutdown -force), which doesn't work (it doesn't
kill the server for some reason and it keeps hogging the listen port
and not processing requests) - so I switched to 'killall -9 java'.
This does not seem to be a out-of-memory problem - I had those with
tomcat 4.x, java 1.4.2 and the same webapp - and after I moved to
tomcat 5.5 and java 1.5 those problems ceased - so this suggests that
the webapp is OK at least memory-wise. But now this one came up.
So I'm wondering how to diagnose this problem and figure out what is
going wrong. The catalina.log files seem to be missing in tomcat 5.5
and that could perhaps give me some more hints than the
IllegalStateException from the catalina.out logfile.
Can you give me some hints what this problem could be caused by? Or
perhaps ideas how to figure out what is going wrong?
I've noticed in tomcat docs that I can list the worker threads using
the jmxproxy in the manager webapp - but I can't really decipher the
output - maybe that would give me some info on why it stops
responding. Maybe I'm the one causing the problem - by some deadlock
in my webapp stopping the threads one by one (which is unlikely but
possible).
I also diagnosed the webapp using sun's jvmstat, but that just showed
that the webapp is running fine and there is enough memory and GC is
taking only a small fraction of running time..
Thanks,
Wojtek
-
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
