> > I was hoping it would never make to kernel, and if it did, it
> > would never be enable by default. But it seems I was wrong on both
> > accounts :(
>
> > VM: killing process sendmail
> > VM: killing process klogd
>
> This is *NOT* the OOM killer I submitted to Alan. My killer
> spits out different messages and tries to locate the memory
> hog first so it avoids killing sendmail and klogd.
ah, sorry. I see, it comes from arch/i386/fault.c
so what happened was that, for example, emacs requested extra mem,
it could not get it and it got killed ?!?!? .. it is not behavior I used
to know from 1.2.x kernels.
I'm perplexed.
Still, I don't like any for out of OOM killings, and it reminds me of AIX
and Windows. If that's posssible, ideally I would like to have some proc
interace to be able to diable oomk at boot time.
I used to have a latop. It was either late 1.2.x kernel or early 2.2.x
kernel. Since it has only 16 mb of ram, it used to run out of memeory
every thing. However, I used to use it heavily every day and it had uptime
of over 200 days (only so much b/c its bateries did not have enough juice
to survie trans-atlantic flight).. and I never had any problems due to oom
conditions.. at least problems which would bother me, like killing my
emacs.
-
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/
This archive was generated by hypermail 2b29 : Tue Feb 29 2000 - 21:00:17 EST