Hi Dan,

I suppose this could be one of our problems, Could this "illegal
instruction" problem listed here be the cause of our sig 11 errors?  
(there is no mention of -i "illegal" in any of the error files that the
JVM saves on each crash).

The "patch_final" link at the bottom of this page doesn't work, and I
tried applying the contextual diffs to RH 7.3 (2.4.18-03), but most of
them failed.

Craig

On Thu, 23 May 2002, Dan Kegel wrote:

> That reminds me: there is another important patch
> for x86 SMP.  It has not yet been blessed by Red Hat, though.
> See http://marc.theaimsgroup.com/?l=linux-kernel&m=102208353523931&w=2
> - Dan
> 
> Calvin Austin wrote:
> > 
> > There is a very important patch to Redhat 7.2 SMP kernel, you need the
> > 2.4.9-31smp patch available under 'security patches'  from redhat.com
> > 
> > Unpatched redhat 7.2 SMP will create duplicate process ids. That is not good!
> > The bug is also fixed in 2.4.18 kernels. Hui Huang tracked this down a while
> > back
> > 
> > regards
> > calvin
> > 
> > Jesse Stockall wrote:
> > >
> > > On Thu, 2002-05-23 at 04:44, Craig O'Shannessy wrote:
> > > >
> > > > Sun JDK 1.4.0
> > > > Sun JDK 1.3.1
> > > > Sun JDK 1.3.1_03
> > > > IBM JDK 1.3.1_01  (had to use "LD_ASSUME_KERNEL=2.2.5" for smp machine)
> > > > Blackdown JDK 1.3.1_01 (both native and green threads, both give sig 11)
> > >
> > > In my experience you should always be using "LD_ASSUME_KERNEL=2.2.5"
> > > with any of these JRE's and RH 7.2. (It was mentioned in the release
> > > notes for RH 7.1)
> > >
> > > I run Blackdown's and Sun's 1.3.1 on Debian & RH both single & dual cpu
> > > & have never encountered any of the problems you seem to be having. Well
> > > I used to get the odd crash & many defunct java processes, but they were
> > > resolved after setting "LD_ASSUME_KERNEL=2.2.5".
> 


----------------------------------------------------------------------
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

Reply via email to