Re: Probleme mit Multiprozessorboard !

Ruediger Oertel (ro@suse.de)
Wed, 8 Sep 1999 16:22:51 +0200 (MEST)


On Wed, 8 Sep 1999, Simon Richter wrote:

> On Wed, 8 Sep 1999, Marcus wrote:
>
> > Hallo !!
>
> Bitte sprich Englisch.
>
> Rough translation in []:
>
> > Hardware :
> > - Tyam Multiprozessorboard TYN Tomcat IV, Bios V1.02 (Award)
> > - zwei Pentium 233
>
> [two P233]
>
> I hope you have the multiprocessor-capable version of these processors.
>
> > - 512 MB RAM (64MB-Module)
> > - Raid-Controller Mylex DAC960 unter RAID 5 mit 4 mal IBM 8GB-Platten
>
> [said RAID controller, using 4 harddisks of 8GB each in a RAID5 array]
>
> > und SUSE 6.1. mit Kernel 2.2.7
>
> > Das Problem : Mit dem Boot-Kernel startet das System fehlerfrei mit den 512
> > MB Ram, aber sobald der Standartkernel installiert ist und Linux anfängt mit
> > beiden Prozessoren zu arbeiten(unsere Vermutung), stürzt der Kerrnel ab.
>
> [The installation kernel from the SuSE distribution works flawlessly,
> however the kernel for normal operations supplied by SuSE crashes. This
> kernel is SMP enabled, and our assumption is that this happens when
> working with both processors]
>
> This is a bit unclear... :-)
It's a bit clearer to me. He says the kernel used for installation runs the
machine without problems. Then he refers to a kernel "supplied by SuSE",
whatever this is supposed to mean. The SuSE-6.1 does not include a precompiled
SMP kernel, so he might have gotten a different kernel from someone else here.

If the kernel crashes real early:
OTOH, I don't even think the kernel itself is the problem at all. In about 80%
of the cased I heard of this scenario, the problem was the bootloader. The
install-disks use syslinux, after installation LILO is used. Most versions
of LILO used to occupy an area in memory which contained information from
the BIOS for multiprocessor use.
AFAIR SuSE-6.1 has a version of lilo with LARGE_EBDA support on CD1 in /unsorted.
I'd suggest to try this one.

Otherwise:
> When exactly does the crash happen, is there a message, is the crash
> reproducible, does it work if you disable SMP but use this kernel, how is
> the kernel configured, what chipset do you use (I don't know that
> mainboard), what devices are there, etc.
I'd second this.
>
> Please see the file /usr/src/linux/Documentation/oops-tracing.
>
> > Hat jemand ähnlicher Erfahrungen mit dieser Hardware oder allgemein unter
> > Multiprozessorbetrieb gemacht ?
>
> [Has anyone experienced similar things with this kind of hardware or with
> multi-processing in general?]
>
> Much sh*t has hit the fan since SMP support was introduced, and many
> things have been fixed. Try to compile a recent kernel, if the machine is
> going to be a server (which I judge from the RAID), this is an absolute
> must.
>
> <HOLYWAR>(And SuSE distro is not really suited for servers BTW.)</HOLYWAR>
Well, I've seen quite a lot of them running ;-)
<kill the flame, we all need to live>

-- 
with kind regards (mit freundlichem Grinsen),
                                     Ruediger Oertel (ro@suse.de)
-----------------------------------------------------------------
Linux 2.2.12.SuSE (root@fatou) (gcc version egcs-2.91.66 ) #1 SMP
       Total of 2 processors activated (800.36 BogoMIPS)

- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.rutgers.edu Please read the FAQ at http://www.tux.org/lkml/