xfs
[Top] [All Lists]

Re: Kernel oops

To: Juha K Kallio <bunnyh@xxxxxxxxxxxxxx>
Subject: Re: Kernel oops
From: Steeve McCauley <smccauley@xxxxxxxxx>
Date: Thu, 14 Mar 2002 06:58:59 -0500
Cc: linux-xfs@xxxxxxxxxxx
In-reply-to: <20020314075218.GA2662@bunny.ihme.org>; from bunnyh@banix.ihme.org on Thu, Mar 14, 2002 at 09:52:18AM +0200
References: <20020314055144.GA4312@bunny.ihme.org> <29292.1016085499@kao2.melbourne.sgi.com> <20020314075218.GA2662@bunny.ihme.org>
Sender: owner-linux-xfs@xxxxxxxxxxx
User-agent: Mutt/1.2.5i
On Thu, Mar 14, 2002 at 09:52:18AM +0200, Juha K Kallio wrote:
> On Thu, Mar 14, 2002 at 04:58:19PM +1100, Keith Owens wrote:
> > On Thu, 14 Mar 2002 07:51:44 +0200, 
> > Juha K Kallio <bunnyh@xxxxxxxxxxxxxx> wrote:
> > >I got a kernel oops on the display this morning when i woke up. It had
> > >something about kupdated being killed, but i can't get the whole error,
> > >since it could't log anymore. 
> > >Right after i tried to read my mail, it crashed. Strangely i could still
> > >exec binaries.. Any idea what might have caused this?
> > 
> > To do any debugging we need the first oops (all of it) run through
> > ksymoops.  Without that plus your kernel version, XFS version, gcc
> > etc., there is no hope of debugging this.
> >
> 
> 
> I've never had a kernel oops before, so I'm really not experienced in
> debugging. But I have /var/log/ksymoops and some files in there, do I
> have to do something with those? Strangely, some of them are dated three
> days back, I hadn't any problems before this morning.

Check in /var/log/messages for the Oops.  If it's there copy
it to another file and run it through ksymoops.  Make sure that
System.map is symlinked to the correct System.map-xxx file in
/boot, normally this will be true while running the same kernel
that caused the oops.  See ksymoops manpage for details.

It's possible that the oops was not written to the syslog
though and if that's the case and you do not have a photograhpic
memory, you'll have to wait until it happens again (type in
the oops or write it down, painful but the fs programmers will
find it invaluable in tracing the cause of the oops.)

-- 
Steeve McCauley                         Maximum Throughput
Programmer & Team Leader                smccauley@xxxxxxxxx
:wq


<Prev in Thread] Current Thread [Next in Thread>