On 10/30/12 4:23 PM, Keith Fetterman wrote: > Thanks Scott, > > We are currently running 4.0.23 Pro. I am getting ready to deploy > 4.0.32 Pro into production. Do you think the reporting in 4.0.32 Pro is > better and would help troubleshoot the problem?
It should be about the same. The 4.0.23 heap dump will show the same information. There are a few reporting changes in 4.0.32, but nothing major. -- Scott > > Keith > > > On 10/30/2012 02:04 PM, Scott Ferguson wrote: >> On 10/30/12 10:14 AM, Keith Fetterman wrote: >>> Hi Everyone, >>> >>> I could use a little help troubleshooting some messages in my >>> jvm-default.log. I am seeing a lot of the following messages in our >>> Resin logs. We are currently running Resin 4.0.23 Pro. >>> >>>> INFO com.caucho.env.health.HealthService >>>> MemoryTenuredHealthCheck[WARNING:0.48% free (10.127M free, 2097.152M max, >>>> 2087.025M used)] >>> I did a bit of research this morning about adjusting tenured memory >>> limits, but I didn't find anything specific to adjusting the value. Does >>> anyone know what property I need to adjust? I currently have the >>> following Java settings in my resin.xml file: >> As Knut said, that size should be plenty. >> >> I'd look to see if you have some kind of memory leak. Take a look at >> either the heap section of Resin's PDF or /resin-admin. Specifically >> look for the ZombieClassLoaderMarker. >>>> <jvm-arg>-J-server</jvm-arg> >>>> <jvm-arg>-J-d64</jvm-arg> >>>> <jvm-arg>-Xms3000M</jvm-arg> >>>> <jvm-arg>-Xmx3000M</jvm-arg> >>>> <jvm-arg>-XX:MaxPermSize=250M</jvm-arg> >>> I could increase the -Xmx property, but I am wondering if there is a >>> better property to tweak. >>> >>> I am also seeing the following warning messages: >>> >>>> WARNING com.caucho.health.analysis.AnomalyAnalyzer AnomalyAnalyzer >>>> JVM|Thread|JVM Blocked Count WARNING >>>> WARNING com.caucho.health.analysis.AnomalyAnalyzer AnomalyAnalyzer >>>> JVM|Thread|JVM Runnable Count WARNING >>> Does anyone have some advice on how to resolve these warnings? I read >>> the Resin documentation, but I didn't find something specific. >> Again, take a look at the PDF report (or thread graphs at that time.) >> >> If it's a little spike (2-3 threads), it's not a big problem, but if >> there are lots of threads in that spike you might have some kind of >> locking issue. >> >> -- Scott >> >>> Thanks for your help >>> >>> Keith >>> >> _______________________________________________ >> resin-interest mailing list >> email@example.com >> http://maillist.caucho.com/mailman/listinfo/resin-interest _______________________________________________ resin-interest mailing list firstname.lastname@example.org http://maillist.caucho.com/mailman/listinfo/resin-interest