* KAMEZAWA Hiroyuki <[email protected]> [2009-08-24 16:18:25]:

> On Mon, 24 Aug 2009 12:47:57 +0530
> Balbir Singh <[email protected]> wrote:
> 
> > * KAMEZAWA Hiroyuki <[email protected]> [2009-08-24 15:58:35]:
> > 
> > > On Mon, 24 Aug 2009 08:17:06 +0200
> > > Dietmar Maurer <[email protected]> wrote:
> > > 
> > > > > > how about memsw_limit for swap? :>
> > > > > > I am looking for swap usage statistics from cgroup right now from
> > > > > > memcontrol.c :) but as you did the idea is good and should be add to
> > > > > the
> > > > > > kernel and lxc-tools :)
> > > > > >
> > > > > 
> > > > > Hmm, why meminfo is necessary ? For cheating top/free/... etc ?
> > > > 
> > > > Many servers use that info to control the use of resources. For 
> > > > example, they do not span any more server processes if memory is low.
> > > > 
> > > > If you do not have that info correct, all those programs fails (they 
> > > > try to allocate much more resources than available).
> > > > 
> > > 
> > > Hm, duplicating information for keeping binary compatibility for apps ?
> > > What other stat does ? 
> > > For example, 
> > >   - load average
> > >   - cpu usage %user, %sys, %iowait, etc..
> > >   - fork-per-sec
> > > 
> > > One difficulty I think of is that memcg's limit-usage doesn't means 
> > > "free".
> > > But, hmm, showing "meminfo" in popular/user-friendly format is not very 
> > > bad.
> > > Balbir, how do you think ?
> > >
> > 
> > Sorry, I've missed the context and unable to follow the discussion.
> > IIUC, the request is to show per-container statistics?
> >  
> per-container "meminfo" i.e. the same format information as "meminfo".
> (Then, duplicating information of memory.stat")

That seems like a reasonable goal to achieve.

-- 
        Balbir
_______________________________________________
Containers mailing list
[email protected]
https://lists.linux-foundation.org/mailman/listinfo/containers

_______________________________________________
Devel mailing list
[email protected]
https://openvz.org/mailman/listinfo/devel

Reply via email to