JVM crashes on one box and not on another with similar setup

[Roshan George Cheereth]

Hi,

We are running around 10 application instances of ATG Dynamo v6.4 on 2 high end HP boxes with a similar setup. We had never had any problems with them till 3-4 weeks back where the instances go down suddenly with some giving vm dumps. We are still investigating if there were any changes made recently but till now we do not see any. I am providing few vm dumps and other info thrown in the logs before it goes down.

What we want to understand is if there are any issues with the patches or any of the library files are corrupt causing this to happen on one particular box?
Is there a possibility that we are allocating less MaxPermSize during the startup?
Do you see any common factor in these dumps which could say that there is any kind of environmental setting problem. ( I did verifiy that the instances which went down had the same settings of heap against the instances in the box which was healthy).
There are times its not able to load the configuration ( properties files ) properly forcing us to restart it again and then it comes up fine. This never happens on the other healthy box with a similar setup.

Attached are the logs and the JVM args.

Thanks,
Roshan

[Steve Steel]

Hi


This is known and fixed in 1.4.2.04


Get a more uptodate version

see www.hp.com/go/java

latest is


SDK & RTE 1.4.2.06 (Sun 1.4.2_06) X X


Steve Steel

[Roshan George Cheereth]

Thanks for the reply.

My question here would be :
If it is an issue with JDK version why doesn't it occur on the other box which has the same issue?

Were you able to see the attachment of dumps I have attached here?

regards,
Roshan.

[Steve Steel]

Hi


Similar yes but never exactly the same

maybe something got corrupted or as you say memory

either way
www.hp.com/go/java

get a new version and look at hpjtune as well


Steve Steel