Re: linux crashes

[Rodolfo Brasnarof <rodob@datafull.com>]

On Sun, 09 Jan 2005 23:23:28 -0000
Ian Haver wrote:

> On Sun, 9 Jan 2005 20:59:51 -0000, Tom Vavra <wb8zrl@avalon.net>
> wrote:
> 
> > In mid December I replaced my DOS based packetcluster system with a
> > Linux based DXSpider system.  Since that time I have had 5 system
> > crashes.  I think all are the same.  Enclosed at the end is the
> > transcribed crash screen.  I tried to proof it, but I would not bet
> > too much that I got it correct.
> >
> > I am running RedHat 9.0 with a 2.4.20 kernel.  The netrom/ax25
> > packages  are:
> >
> >   ax25-apps-0.0.6
> >   ax25-tools-0.0.8
> >   libax25-0.0.11
> >   node-0.3.1
> >
> > I am using three serial ports to drive tncs.  Two of the tncs are
> > tapr-2(MFJ-1270b) with BPQKISS eproms, and the other tnc is a two
> > port  KPC-9612,
> > using both ports.  So I am driving all three ports with mkiss,
> > followed  by
> > kissattach.  Two netrom ports are defined.  And I am running
> > mheardd.
> >
> > The cpu is a 450mhz pentium that idles at less than 2%, and there is
> >  
> > 256Mb
> > of memory so swap is almost nonexistant.
> >
> > Anyone have any ideas about what I need to do to eliminate the
> > crashes?
> >
> >              Tom  WB8ZRL
> >
> 
> HI Tom
> I had similar problem with my suse distro.
> I had to download a new kernel and patch it up.
> The files can be found on my server here at YFS but i dont have the  
> fastest ADSL link.
> You are more than welcome to download them.
> they can be found at:
> http://gb7yfs.org.uk/public/linux/new_kernel_bits/
> 
> also in that directory is all the ax25 bits i used here.
> When i was running 2.4.20 i could average around 10 days uptime, since
> i  upgraded it runs and runs and runs till i reboot it. LOL
> 
> Hope this helps a little

You can still set your system to automatically reboot in case of a
kernel panic. Of course is better to fix the problem, but for an
unattended system this setting can be useful.

You can do this by adding a line

	echo 1 > /proc/sys/kernel/panic

to your startup scripts (maybe rc.local)

or adding a line

	kernel.panic = 1

to sysctl.conf

The 1 means the number of seconds to wait for reboot after a kernel
panic. 0 means no reboot.