Re: [PATCH 2/2] memcg: add memory.numastat api for numa statistics

[Ying Han <yinghan@google.com>]

[-- Attachment #1: Type: text/plain, Size: 1587 bytes --]

On Tue, May 17, 2011 at 4:52 PM, KAMEZAWA Hiroyuki <
kamezawa.hiroyu@jp.fujitsu.com> wrote:

> On Tue, 17 May 2011 15:25:51 -0700
> Ying Han <yinghan@google.com> wrote:
>
> > The new API exports numa_maps per-memcg basis. This is a piece of useful
> > information where it exports per-memcg page distribution across real numa
> > nodes.
> >
> > One of the usecase is evaluating application performance by combining
> this
> > information w/ the cpu allocation to the application.
> >
> > The output of the memory.numastat tries to follow w/ simiar format of
> numa_maps
> > like:
> >
> > <total pages> N0=<node 0 pages> N1=<node 1 pages> ...
> >
> > $ cat /dev/cgroup/memory/memory.numa_stat
> > 292115 N0=36364 N1=166876 N2=39741 N3=49115
> >
> > Note: I noticed <total pages> is not equal to the sum of the rest of
> counters.
> > I might need to change the way get that counter, comments are welcomed.
> >
> > Signed-off-by: Ying Han <yinghan@google.com>
>
> Hmm, If I'm a user, I want to know file-cache is well balanced or where
> Anon is
> allocated from....Can't we have more precice one rather than
> total(anon+file) ?
>
> So, I don't like this patch. Could you show total,anon,file at least ?
>

Ok, then this is really becoming per-memcg numa_maps. Before I go ahead
posting the next version, this is something we are looking for:

total=<total pages> N0=<node 0 pages> N1=<node 1 pages> ...
anon=<total anon pages> N0=<node 0 pages> N1=<node 1 pages> ...
file=<total file pages> N0=<node 0 pages> N1=<node 1 pages> ...

please confirm?

thanks

--Ying

> Thanks,
> -Kame
>
>

[-- Attachment #2: Type: text/html, Size: 2635 bytes --]