Hi Miles,
	Did you try the RH8 or stock kernel on those machines as a test?

							regards,

							Stephen.

On Tue, 25 Oct 2005, Miles O'Neal wrote:

> After the upgrade from RH8 to SL304, most
> of our few stability problems seem to have
> gone away.  However, we gained some new
> stability problems.
>
> We have 10 systems that ar eover two years
> old that have never given us a single
> problem, no matter how hard we flogged
> them.  These are 2.4GHz P4 systems using
> the Intel D845PE boards (no onboard LAN,
> audio or video).  They all have some form
> of 3Com 509 and a variety of video cards,
> including S3 Virge, ATI and Matrox.
>
> Since the SL304 upgrade a couple of months
> ago, 7 of these boxes have locked up or
> shown a kernel panic, most of them more
> than once, two on an almost daily basis.
> The shop who builds our systems can't
> find anyhting wrong with them, and ran
> the two that die daily here for two weeks
> at their shop with no glitches.  All their
> diag tools run under Windows or DOS) so
> this appears to be OS-related.  (Other
> systems on the same UPSes show no problems,
> so it's not a power issue).
>
> It can happen when they're really busy
> or (it seems) even when they're idle.
>
> I wondered about microcode issues; the
> microcode_ctl startup task is on and the
> logs show it ran:
>
>    Oct 25 10:24:48 sim026 kernel: microcode: No suitable data for cpu 0
>
> There's nothing suspicious in the logs;
> they just lock up.  Both systems that
> locked up today had the CAPS LOCK and
> SCROLL LOCK keyboard LEDs flashing.
>
> Has anyone else seen anything like this?
>
> Our AMD systems are all fine, and the
> other Intel boxes (faster SuperMicros
> using I forget what board) are fine...
>
> Thanks,
> Miles
>
>

--
 /------------------------------------+-------------------------\
|Stephen J. Gowdy                     | SLAC, MailStop 34,       |
|http://www.slac.stanford.edu/~gowdy/ | 2575 Sand Hill Road,     |
|http://calendar.yahoo.com/gowdy      | Menlo Park CA 94025, USA |
|EMail: [log in to unmask]       | Tel: +1 650 926 3144     |
 \------------------------------------+-------------------------/