I have the CRUX-2.1 distribution with a custom kernel installed on a computer
that I built from an HP Pavilion's core (Trigem Cognac MB, Celeron 400, 64MB
SDRAM, etc.) and I have a fishy little problem with it. When I used "make
menuconfig" I would sometimes get restarts out of the blue, without warning. I
used "make config" and built a kernel the hard way, but despite that, I can
reproduce this unexpected restarting by logging in and then typing in:

# dmesg | less

and holding the "down" scroll button. Never fails. I flashed the BIOS to the
latest revision that can be flashed to. Sometimes this restart causes ext3
journal corruption which results in boot failure.

I've tweaked BIOS settings until they bled. Everything looks like it should
work. I think HP's older products have something terribly wrong with them at
this point.

Help? If you respond to the group, please also CC the reply to my E-mail.

Thanks!
Jody

Not much longer after I wrote this, the 13.6 GB IBM hard drive in this system
went berserk and died. I replaced with a 40 GB Seagate and have not managed to
reproduce this problem again.

The question must be asked, however...why can dmesg | less with a bad hard drive
cause the system to restart?



Jody wrote:
> I have the CRUX-2.1 distribution with a custom kernel installed on a
> computer that I built from an HP Pavilion's core (Trigem Cognac MB,
> Celeron 400, 64MB SDRAM, etc.) and I have a fishy little problem with
> it. When I used "make menuconfig" I would sometimes get restarts out of
> the blue, without warning. I used "make config" and built a kernel the
> hard way, but despite that, I can reproduce this unexpected restarting
> by logging in and then typing in:
>
> # dmesg | less
>
> and holding the "down" scroll button. Never fails. I flashed the BIOS
> to the latest revision that can be flashed to. Sometimes this restart
> causes ext3 journal corruption which results in boot failure.
>
> I've tweaked BIOS settings until they bled. Everything looks like it
> should work. I think HP's older products have something terribly wrong
> with them at this point.
>
> Help? If you respond to the group, please also CC the reply to my E-mail.
>
> Thanks!
> Jody

Jody <jbruchon@nc.rr.com> wrote:
> The question must be asked, however...why can dmesg | less with a bad hard drive
> cause the system to restart?

The question must be asked "wy do you ask that"? It has to read the
code of less from SOMEWHERE, no? And it has to search SOMEWHERE to
find that somewhere no? And the somewhere that it is searching is a
file system on your faulty disk, no? Hence corrupt. Hence not in an
expected state. And it has to place other stuff in SWAP, no? And that
swap is where? Somewhere?

Fix that.

Peter

"Jody" <jbruchon@nc.rr.com> wrote in message
news:3lf_e.76068$Jp.2024537@twister.southeast.rr.c om...
> Not much longer after I wrote this, the 13.6 GB IBM hard drive in this
> system went berserk and died. I replaced with a 40 GB Seagate and have
> not managed to reproduce this problem again.
>
> The question must be asked, however...why can dmesg | less with a bad hard
> drive cause the system to restart?

First rule of this newsgroup. Ignore Peter snarking at you.

"dmesg" and other log probing tools like "last" have to read two different
chunks of disk, at least:

1: The bit with the program on it.
2: The bit with the logs on it.

I'd suspect the bit with the logs on it as failing: logs get written to
quite a lot, and thus wind up getting more physical wear and tear. Many
moons ago, when I worked with Sun's, it was invariably /var/adm that would
fail and require re-partitioning and disk replacement or rebuilding from
backup.

On Tue, 27 Sep 2005 15:37:37 GMT, Jody
<jbruchon@nc.rr.com> wrote:
> I have the CRUX-2.1 distribution with a custom kernel installed on a
> computer that I built from an HP Pavilion's core (Trigem Cognac MB,
> Celeron 400, 64MB SDRAM, etc.) and I have a fishy little problem with
> it. When I used "make menuconfig" I would sometimes get restarts out
> of the blue, without warning. I used "make config" and built a kernel
> the hard way, but despite that, I can reproduce this unexpected
> restarting by logging in and then typing in:
>
> # dmesg | less
>
> and holding the "down" scroll button. Never fails. I flashed the
> BIOS to the latest revision that can be flashed to. Sometimes this
> restart causes ext3 journal corruption which results in boot failure.
>

Try memtest86.


--
You will be given a post of trust and responsibility.