Subject: Re: [vserver] oom-killer killapps on hosts not not guest
From: Michael <michael.auckland@gmail.com>
Date: Thu, 20 Aug 2009 07:19:43 +1200

 Thu, 20 Aug 2009 07:19:43 +1200
> you can avoid that by making host processes immune
> to the OOM killer, or by raising the badness bias of
> the guests above any host vm usage, or disable
> overcommit completely

And how to make host immune to OOM killer?

>> I've tring to setup vm.overcommit memory=2  which doesen't help
>
> probably because you left the overcommit ratio untouched,
> but you want to set this to 0 too

used overcommit 80 , or 0 with almoust the same results.

>> The only option seems to be
>> vm.oom kill allocating task = 1
>
> this is another way to address this specific case, but
> it might not be the best for the entire system
>
>> Is it a right way to do it?
>
> the OOM 'badness' which is the criterion used to choose
> the 'victim' is calculated like this:
>
>  1) total vm of a process (system wide) plus
>  2) the badness bias (if in a guest, default is 0) plus
>  3) the difference between rss curr and rss soft
How difference rss curr and rss soft works? ( which settings give more
score for a guest?)

Added 500000 to /etc/vservers/vserver1/badness
but still got process killed on host.

> this is how it is calculated right now, with one exception,
> if there is a swapoff process running, it gets full score :)
>
> we might add another 'difference' based on the vm
> settings or some kind of multiplicative factor to that
> (if you like to test it, let me know)

Yes , I've got this testing comp, so can test it.

Using latest kernel 2.6.30.5-vs2.3.0.36.14-pre5


> HTH,
> Herbert
>
>> --
>> --
>> Michael
>



-- 
--
Michael