Bug#341801: independent confirmation
Anthony DeRobertis wrote:
> Kevin Brown wrote:
>
> >>PS: Since you have the same hardware as me, I'm curious if you're able
> >>to use kernels > 2.6.12.
> >
> >
> > Yeah. I'm currently using 2.6.15. Note that I'm using the
> > amd64-generic kernel, not the amd64-k8 kernel. I don't expect that
> > makes any difference at all with respect to this particular bug,
> > though it might change how much memory is visible.
>
> I'm using the k8-smp version. I'd try generic, but there is no generic
> SMP, AFAIK (and having bought a >$500 dual-core chip, I damn well want
> to use it!)
>
> The problem I've always seen with > 2.6.12 is that after a short while,
> either Ethernet or SATA (with the promise chip) dies. I've filed
> http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=347412 on this (a while
> ago).
>
> I'm sort of curious why you use -generic instead of -k8; unless, of
> course, this is why.
Well, the main reason is that I didn't see any advantage, since I'm
running a non-SMP system.
The secondary reason is that I'm running the NVidia proprietary
drivers, and getting them set up was a pain.
And finally, with the exception of this bug, everything appears to
work.
> > One other thing: I'm using ECC memory, and have ECC enabled.
>
> Why they heck would you do that... ?
Two reasons: the first is that I want as much stability as I can get,
and ECC increases that. I'll leave my systems running for hundreds of
days at a time unless there's a compelling need to reboot. The second
is that I value my data, and don't want an undetected memory error
screwing me over. I also run my disks in a mirrored configuration for
the same basic reason.
--
Kevin Brown kevin@sysexperts.com
Reply to: