PJS = Peter J. Stieber
PJS>> I have been having the "memory.c bad pmds" with
PJS>> a Tyan S2885 motherboard.
PJS>>
PJS>> https://www.redhat.com/archives/fedora-list/2005-May/msg01690.html
PJS>>
http://www.lib.uaa.alaska.edu/linux-kernel/archive/2005-Week-19/1397.html
PJS>>
PJS>> <Skip a lot>
PJS>> I'm just trying anything I can to get rid of the
PJS>> bad pmd messages.
CW=Christopher Warner
CW> Just read the other existing thread linked. Is
CW> everyone running the same model opteron on these
CW> Tyan boards (246)? To update even further. Besides
CW> the parallel port problem I've sent back about 5 of
CW> these tyan motherboards. A couple of then simply
CW> didn't have the gigabit network adapters available
CW> via bios. In some cases linux loaded the drivers for
CW> the adapters regardless of their setting in BIOS. In
CW> other cases they simply were not available.
CW>
CW> Also, are the processes you run swapping out/in memory
CW> repeatedly? Or using a large amount of threading?
CW> Thread pools? I'm trying to get rid of this problem
CW> myself. I've seen it extensively in 2.6.11.5 and not as
CW> much in 2.6.11.8 but I haven't tested lately past .8
I'm running two 244 processors not 246 processors. The process that
causes problems is a bash shell script that builds a suite of codes. The
scipt call other scripts that call autoreconfig then call configure and
the cofigure dies sometimes. If I repeat the process it works. It might
be swapping/paging related. I'm not sure how I could tell. I normally
use stock Fedora distributions. I started seeing the problem in April
after I installed 2.6.11-1.14_FC3smp. I'm currently running
2.6.11-1.29_FC3smp provided by Dave Jones and I still see the problem.
Pete