The JVM is no different than any native application in terms of core dumps. The most common cause of a core dump in my experience has been one application attempting to write to an area of reserved memory. Potentially this is because of a bug in the OCI driver. Unfortunately it can be difficult to debug as there is no simple way of associating Java calls (which are interpreted by the JVM) with the actual system level calls which the JVM makes to communicate with the underlying operating system - the Java standard does not impose rules as to exactly how this communication is to be achieved.

That said, if you are familiar with core dump analysis, you *may* be able to glean some helpful information about what is happening. Refer to this site to get some information on how to perform core dump analysis.

Hope this helps...

Dermot