segmentation fault

["Osmundsen, Helge" <He...@usa.xerox.com>]

We are running tomcat 3.1 on solaris 2.6 with connection pool with 200 max
threads, 100 min, 120 max spare threads.  We have @ 40 users (all heavy
users), and about once per day we get the following

SIGSEGV   11*  segmentation violation
    si_signo [11]: SIGSEGV   11*  segmentation violation
    si_errno [0]: Error 0
    si_code [1]: SEGV_MAPERR [addr: 0xbbbbfa64]

        stackbase=E5891D6C, stackpointer=E58905FC

Full thread dump:

We have an auto-restart script set up so we are only affected for a minute
or so, but we would like to prevent this altogether.  The memory heap size
is 120MB (ms+mx).  The last garbage collection shows we have plenty of
memory available (90-95 % free).

I see connection expiration messages in the full thread dump for a couple of
threads that had RMI connections.  Don't know if that means anything.  We
are running JRE 1.1.6 with native threads. 

 Any suggestions would be appreciated,

Helge Osmundsen