On Sun, 13 Feb 2000, Bernd Paysan wrote:
> I've been using Rik van Riel's OOM patch for quite a time, and
> kept it up to date with the 2.3.x kernels. I added a /proc/sys/vm
> entry to configure it (out-of-memory-kill: limit-to-kill
> limit-to-term, in percentage of overall free memory).
The "free memory" thing was kind of a design mistake by me.
It usually works great, but I can see it fail in a situation
where we have a whole bunch of busy mmap()ed zones, or don't
have swap, etc...
> I must say I need it. My programs go wild too often. This patch
> selects the worst offender and kills it. There are two limits: one
> to send a SIGTERM, one to send a SIGKILL. So far, it hasn't caused
> any false kills with me. I tried to live without it, and the rest
> of OOM handling in the kernel just doesn't do it. An OOM situation
> without this patch still results too in a machine that needs a
> hard reset.
Good to hear that the old patch still works :)
Anyway, you might want to look into the new version of my
patch, against 2.2.15pre7. You can grab that patch from my
home page: http://www.surriel.com/patches/
And yes, I think that we want the OOM handling in the kernel.
It doesn't interfere with the main code path and really saves
your butt when you're out of memory...
Alan, could you please integrate the 2.2 version (the one I
sent you last week, please dingusclick the url above) into
2.2.15? It seems to work quite well, at least I haven't managed
to break it in 3 days of testing various situations...
cheers,
Rik
-- The Internet is not a network of computers. It is a network of people. That is its real strength.- 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 15 2000 - 21:00:26 EST