[olug] kernel panics

[Ryan O'Rourke]

On Tue, 2003-06-03 at 09:50, Vincent wrote:
> That isn't much info to go on, but here are a couple standard things to try if you still have trouble after replacing the disk:
> 
> First, make sure it reboots after a panic.  Add this to /etc/sysctl.conf:
> kernel.panic = 3    # System will reboot 3 seconds after a kernel panic
> To enable this now:
> echo "3" > /proc/sys/kernel/panic
> 
> Next, try booting with each of these lilo (or grub) boot options and see if either helps:
> 
> Disable machine checking.  This tells the kernel not to check the cpu's data structures.  The BIOS is probably already doing this
> anyway and two simultaneous checks can cause problems.
> append="root=LABEL=/ nomce"
> 
> Disable apic.  This disables the use of IRQs above 15 (SMP or not).  Some hardware has trouble with APIC (Advanced Programmable
> Interrupt Control/ler)
> append="root=LABEL=/ noapic"
> 
> Run a memory tester:  http://www.memtest86.com/


Thanks, Vincent. That's pretty much what I was looking for - some advice
on what to do when a panic happens and what to keep an eye out for to
solve the problem.
This one seemed particularly scary to me because of the numerous panics
at different times during boot, fsck, and uptime.

I did find this little gem on O'Reilly's ONLamp.com about preparing a
FreeBSB machine to do a backtrace after a kernel panic to effectively
debug the problem.
Is there some way to do this in RedHat?
http://www.onlamp.com/pub/a/bsd/2002/03/21/Big_Scary_Daemons.html?page=1