Re: Beating a dead Horse I know but.....

[Peter Hrastnik <pe...@1012surf.net>]

I had the same problem as you but I got a feedback from SUN:

<SNIP>
Hi Peter Hrastnik,

Thank you for using our bug submit page. The bug you have reported
is a duplicate of the Bug ID: 4372197. For more information on this,
please refer to the Bug ID: 4372197 at:

http://developer.java.sun.com/developer/bugParade/bugs/4372197.html

The bug has been fixed for ladybird (1.3.1).  You can download the 
early beta
of ladybird  at:

http://developer.java.sun.com/developer/earlyAccess/j2sdk131/

The beta may NOT already contain the fix, unfortunately.  (Our records
do not definitively show whether the fix made it into the beta version.)
</SNIP>

So if there are to much threads (~ too much traffic), the JVM crashes.

Instead of switching to an early beta (I use SUN Java 1.3.0v2), I added
this to my tomcat environment (as proposed in one of the bug comments):
_JAVA_SR_SIGNUM=16

And it worked! The JVM runs stable now. Instead of crashing 5 times a
day, it doesn't crash anymore.

Bye,
	Peter. 


--
Mag. Peter Hrastnik
tele.ring Telekom Service GmbH
A-1030 Wien, Hainburgerstr. 33
Tel.: +43/1/931012-3277


Simon Chatfield wrote:
> 
> 
> I and a few other people have posted these problems in the past to no
> real conclusion, but I'm getting this problem almost daily with tomcat
> in ssl mode.
> 
> Using jdk version 1.2.2, 1.3.0_v2 and everything between, with and
> without the JIT activated. After a day or two of uptime the
> JVM crashes with nothing but this error...
> 
> #
> # HotSpot Virtual Machine Error, Internal Error
> # Please report this error at
> # http://java.sun.com/cgi-bin/bugreport.cgi
> #
> # Error ID: 43113F32554E54494D45110E43505002C5
> #
> # Problematic Thread: prio=1 tid=0x8148e78 nid=0x7ee9 suspended
> #
> 
> I have submitted this bug to SUN but still haven't gotten a response.
> This error is not issolatable to a particular line of code or JSP
> page, but I can count on it failing within 3 days of uptime. This ONLY
> happens with tomcat's jvm, (rmi services on the same machine are fine)
> This happens on both linux and solaris. It only appears to happen when
> running in SSL mode.
> 
> Once again, sorry for these repeated posts, but I was hoping this
> extra info might sparc someone's though processes for a solution or
> one of the tomcat developers to look into the problem.
> 
> Thanks
> 
> 
> -Simon
> 
>