Am 11.12.2008 09:43, schrieb Jesse Klaasse:
Rainer Jung-3 wrote:
When using big heaps, you need to take extra effort to get your GC
settings right. Do you have GC-Logs? What are the JVM options you use to
start Tomcat?
The JVM options:
-Dcatalina.base=D:\tomcat -Dcatalina.home=D:\tomcat
-Djava.endorsed.dirs=D:\tomcat\common\endorsed
-Djava.io.tmpdir=D:\tomcat\temp
-Djavax.sql.DataSource.Factory=org.apache.commons.dbcp.BasicDataSourceFactory
-Dcom.sun.management.jmxremote -XX:MaxPermSize=512m
-Xloggc:D:\logs\gc\tomcat-gc.log -XX:+PrintGCDetails -XX:+UseParNewGC
-XX:+UseConcMarkSweepGC -XX:+UseTLAB -XX:+CMSClassUnloadingEnabled
-XX:+CMSPermGenSweepingEnabled
OK, so you are using CMS, good.
Today's GC log: http://www.nabble.com/file/p20951581/20081211_tomcat-gc.zip
20081211_tomcat-gc.zip
Does that file contain any situation, were you had problems? I'm asking
because the date in the file name suggests it's younger than the problem
event you noticed. If it contains a problem: the time stamps are relativ
to JVM start, so at what time was the problem, and when did you start
the JVM (or in short: which offset do we have to look at?). I don't want
to dig into it, if it is just an example file, but not directly related
to an observed problem.
Starting with 1.6.0_07 there is a -XX:+PrintGCDateStamps that adds
absolute timestamps, but unfortunately it isn't available in 1.5 and
earlier versions.
Rainer Jung-3 wrote:
- Check without APR
This seems impossible to do, I had once removed the APR, and the system got
REALLY unstable under a normal day's load. I'm quite hesitant to try this
again, as you may understand.
OK.
So let's hope you manage to collect Thread Dumps of Tomcat next time.
Regards,
Rainer
---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
