Bug#307457: kernel-image-2.6.10-1-686-smp: running into oom-killer: gfp_mask=0xd0 often
On Tue, 03 May 2005, Toni Mueller wrote:
> On Tue, 03.05.2005 at 12:16:51 +0200, maximilian attems <debian@sternwelten.at> wrote:
> > On Tue, 03 May 2005, Toni Mueller wrote:
> > > Package: kernel-image-2.6.10-1-686-smp
> > > Severity: important
> > >
> > > I find error messages like the one above (not sure about the actual
> > > value of gfp_mask) on a regular basis in the logs. Sometimes I just find
> > > out hard because the machine simply crashes shortly after. This makes
> > > the machine in question unreliable, crashing at least every few weeks.
> > > Therefore I'm flagging this bug as "important".
> > >
> > > Hardware is a rather new IBM x205, with ServeRAID4, GigE, and 1GB RAM.
> >
> > you didn't report the error message?!
> > dmesg after boot would also be usefull.
>
> Ok. I thought I was clear enough. The error messages I see are at
> least:
>
> oom-killer: gfp_mask=0xd0
> oom-killer: gfp_mask=0x1d2
> oom-killer: gfp_mask=0x4d0
>
>
> > anyway please try kernel-image-2.6.11 from unstable!
>
> This is currently impossible, due to #306930. Currently, I try to fix
> this (ie, create a local kernel-nonfree-... package).
in oom cases 2.6.11 is very recommended as it has nice fixes from -aa
in that release.
here you'lll find the intersting non-free bits:
http://www.acm.cs.rpi.edu/~dilinger/kernel-source-nonfree-2.6.11/
do you have enough swap, is it activated?
thanks for your feedback.
--
maks
Reply to: