Repeatable kernel panic

[Darryl Clarke]

On 31/03/06, stan <stanb at panix.com> wrote:
> On Thu, Mar 30, 2006 at 10:23:55PM -0500, Darryl Clarke wrote:
> > On 30/03/06, stan <stanb at panix.com> wrote:
> > >
> > > I've installed ubuntu 5.10 (64 bit X86 version) on a Sun Ultra 40.
> > > In this machine I have a Symbios LVD SCSI card. I have an HP DLT80
> > > tape drive connected to this card.
> > >
> > > Doing "dd if=/dev/nst1 bs=32k skip=1" Will reliably panic the machine
> > > I have the AMD SMP kernel installed, but this does not seem to be
> > > kernel specific, as I built a 2.6.16 kernel with the sources from
> > > kernel.org, and I can still reproduce this panic.
> > >
> > > What should I do to properly report this?
> >
> > Since you've tested on a kernel outside of ubuntu's realm, I'd suggest
> > reporting it directly to the kernel guys.
> >
> > http://bugzilla.kernel.org/
> >
> > If I had to guess, I'd throw it in the category IO/Storage -> SCSI
>
> I was really asking for information as to how to best capture useful
> data for the bug report.
>
> Does Linux generate crash dumps for instance? Is there a way to set
> up so that when the panic occurs, the mesages that are sent to the
> console get logged? etc.


Any noise the kernel makes before it's last hurrah, will be in /var/log/kern.log

If in the case the kernel dies before anything is logged there's some
trickery required to get information.

There's more information here:
http://kernel.org/pub/linux/docs/lkml/reporting-bugs.html

--
~ Darryl  ~ smartssa at gmail.com
~ http://darrylclarke.com