Re: [patch] memcg: add oom killer delay

[David Rientjes <rientjes@google.com>]

On Wed, 9 Mar 2011, KAMEZAWA Hiroyuki wrote:

> For me, I want an oom-killall module Or oom-SIGSTOP-all module.
> oom-killall will be useful for killing fork-bombs and very quick recovery.
> 

That doesn't need to be in kernelspace, though, the global oom killer will 
give access to memory reserves to any task that invokes it that has a 
pending SIGKILL, so we could extend that to the memcg oom killer as well 
and then use an oom notifier to trigger a killall in userspace.

> For me, the 1st motivation of oom-disable is to taking core-dump of 
> memory leaking process and look into it for checking memory leak.
> (panic_on_oom -> kdump is used for supporting my customer.)
> 

I remember your advocacy of panic_on_oom during the oom killer rewrite 
discussion, this makes it more clear.

> Anyway, if you add notifier, please give us a user of it. If possible,
> it should be a function which can never be implemented in userland even
> with sane programmers, admins, and users.
> 
> For example, if all process's oom_score_adj was set to -1000 and oom-killer
> doesn't work, do you implement a timeout ? I think you'll say it's a
> wrong configuration. memcg's oom_disable timeout is the same thing.
> 

You know me quite well, actually :)  We've agreed to drop this patch and 
carry it internally until we can deprecate it and implement a separate oom 
handling thread in the root memcg that will detect the situation for a 
given period of time (either via an oom notifier or by checking that the 
usage of the memcg of interest equals to the limit) and then do the 
killall.

--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org.  For more info on Linux MM,
see: http://www.linux-mm.org/ .
Fight unfair telecom internet charges in Canada: sign http://stopthemeter.ca/
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>