Java VM core dump in Weblogic

[Yu Zhen_1]

I have run into a tough case: weblogic often generate core dump during its run time.
Scenario:
HP9000/820/D280???2CPU???180MHz???512MB Mem
JDK???1.3.1.0.1
Weblogic: 6.1sp2
Error message:
MetaTP, 2002-01-21 01:04:35,26,27,381
docommittolog2, 2002-01-21:04:35,1,2,384
bp_name:00001B step:0 logid:null
bp_name:00001B step:0 logid:1015750564
./startManagedWebLogic.sh[175]: 4486 (coredump)

I think I have already had the necessary patches installed. Any suggestions?

[Keith Hughes_1 · Accepted solution]

Yu,

The first thing to try is to upgrade to java c.1.3.1.02 at:-

http://www.hp.com/products1/unix/java/java2/sdkrte1_3/downloads/index.html

Please also check out the patching for Java on HPUX at:-

http://www.hp.com/products1/unix/java/infolibrary/patches.html.

If you still are getting core dumps and you have a software support contract with HP then we will be please to analyse your core dump to find the cause of your problem. The problem may arise due to the JVM falling over, kernel parameters set incorrectly, coding issues or parameters not set (null pointers). We may find out more from the core dump.

Please also check out hp tools HPjconfig and HPjmeter at:-

http://www.hp.com/products1/unix/java/


HPJconfig is run while your java application is running and will give details of the recommended kernel settings.

HPJmeter is a tools used as a performance monitor during the running of the application. There are also instructions on how to look for the bottle necks to enhance performance.


I hope that this helps you.

Keith