java process getting SIGSEGV, pthread issue?

Hello,

We're seeing an issue with a java application in which it gets into a
state in which one of the threads consumes 99% of the CPU.  What appears
to be happening is that one of the threads accesses a bad address and
is signaled with a SIGSEGV, and in the course of handling this signal,
another SIGSEGV is generated.  The result is a tight loop causing the
CPU consumption.

Further debugging seems to reveal that the SIGSEGV is originating from
the libpthread library (pthread_sighandler_rt).  Unfortunately we
do not have the EIP of where the initial SIGSEGV occurred and we cannot
reliably cause the problem to happen.

We're seeing this on a Redhat Advance Server box (both P3 and Intel
Xeon).  Here are some vital statistics:

	JRE:	Sun JRE version 1.4.0_00
	glibc:	glibc-2.2.4-26 (/lib/i686/libpthread-0.9.so)
	kernel:	RH AS 2.1 (2.4.9-e10)

Has anyone else encountered a similar problem?  Any suggestions or
insight would be greatly appreciated.

Thanks in advance.

Pat

p.s.	Please include me on any replies because I am not subscribed to
	this list.

-- 
Patrick O'Rourke
[email protected]