From: Herbert Poetzl (herbert_at_13thfloor.at)
Date: Tue 14 Jan 2003 - 14:35:43 GMT
On Tue, Jan 14, 2003 at 01:17:49PM -0000, Lu?s Miguel Silva wrote:
> Hmm...we are also having problems on one of our servers!
>
> It misteriously crashes! (we dont know if it is a kernel crash or a
> hardware crash...but this email made me suspicious!)
well, you could for example change the blankinterval
in drivers/char/console.c to zero, or export
poke_blanked_console and call it from panic.c or use
some script to keep the console from blanking ...
... you should see a kernel oops
> We have three servers with "exactly" the same configurations:
> p4 2ghz, 512MBram, 2 disks ata100 (doing raid) (80GB each).
> Kernel 2.4.19ctx-15 (SMP support alltough there is only one CPU per box).
you shouldn't do that, SMP support on non SMP machines
is inefficient and almost useless (except for APIC which
is better served by applying the ACPI patches ...
> The only thing that changes on the boxes are the disks. One of them has a
> scsi controller + scsi hdd.
this means that one of the machines has an scsi
controller the other do not have/use?
which of the machines is crashing?
>
> Did anybody find out what causes that "kernel crash" somebody mentioned a
> month ago?
I suggested a INCIDENT REPORT FORM but it was neither
discussed nor used (as far as I know) to report any
configs/incidents ... so no new information has been
gathered ...
best,
Herbert
>
> Best regards,
> +-----------------------------------------
> | Lu?s Miguel Silva
> | Network Administrator@ ISPGaya.pt
> | Rua Ant?nio Rodrigues da Rocha, 291/341
> | Sto. Ov?dio ? 4400-025 V. N. de Gaia
> | Portugal
> | T: +351 22 3745730/3/5 F: +351 22 3745738
> | G: +351 93 6371253 E: lms_at_ispgaya.pt
> | H: http://lms.ispgaya.pt/
> +-----------------------------------------
>
> >
> > There was a discussion about a month back about machines hanging with
> > various vserver kernels. I have 3 that are rock-solid stable, and one
> > that is a DOG (no offense to dogs intended) despite repeated kernel
> > recompiles, multiple versions, hardware swaps, etc.
> >
> > Check the archives at www.paul.sladen.org/vserver, and specifically
> > http://www.paul.sladen.org/vserver/archives/200212/0238.html for the
> > thread.
> >
> > On Mon, 13 Jan 2003, Hans-J?rgen Schwarz wrote:
> >
> >> Hello Cathy,
> >>
> >> Am Sonntag, 12. Januar 2003 02:17 schrieb Cathy Sarisky:
> >>
> >> [...]
> >> > that has NEVER done the mysterious vserver-related hang and has done
> >> 90+ days uptime between reboots.
> >>
> >> I never heard about such behavior. Can you tell me more about it? Does
> >> that mean vservers are not realy reliable for production purposes?
> >>
> >> thank you very much
> >>
> >> Hans-Juergen
> >>
>
>