Re: Misleading OOM messages

From: Balbir Singh
Date: Fri May 15 2009 - 14:20:30 EST


* Dave Hansen <dave@xxxxxxxxxxxxxxxxxx> [2009-05-15 11:15:07]:

> On Fri, 2009-05-15 at 13:57 -0400, Christoph Lameter wrote:
> > > If you misconfigured cgroups, you give more memory to them.
> >
> > If you do not have enough memory in a cgroup then your application should
> > slow down (because of page evictions) but the system should not OOM.
> > Are cgroups broken or why are you getting OOMs when using them?
>
> See mm/oom_kill.c::mem_cgroup_out_of_memory(). A group itself can have
> an OOM done on it. It's not a system-wide oom. We just need to ensure
> that we continue to differentiate the cgroup-specific oom message from
> the general one. Maybe also include some more cgroup info in the debug
> outbut.
>

A long pending action item is to allow user space to handle cgroup
OOM. A cgroup can OOM, if the tasks in the group is way over their limit
and reclaim fails. I'll look to see if we enhance the output.

--
Balbir
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/