From: David Rientjes <rientjes@google.com>
To: vedran.furac@gmail.com
Cc: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>,
Hugh Dickins <hugh.dickins@tiscali.co.uk>,
linux-mm@kvack.org, linux-kernel@vger.kernel.org,
KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com>,
minchan.kim@gmail.com, Andrew Morton <akpm@linux-foundation.org>,
Andrea Arcangeli <aarcange@redhat.com>
Subject: Re: Memory overcommit
Date: Fri, 30 Oct 2009 12:24:08 -0700 (PDT) [thread overview]
Message-ID: <alpine.DEB.2.00.0910301218410.31986@chino.kir.corp.google.com> (raw)
In-Reply-To: <4AEAF145.3010801@gmail.com>
On Fri, 30 Oct 2009, Vedran Furac wrote:
> > The problem you identified in http://pastebin.com/f3f9674a0, however, is a
> > forkbomb issue where the badness score should never have been so high for
> > kdeinit4 compared to "test". That's directly proportional to adding the
> > scores of all disjoint child total_vm values into the badness score for
> > the parent and then killing the children instead.
>
> Could you explain me why ntpd invoked oom killer? Its parent is init. Or
> syslog-ng?
>
Because it attempted an order-0 GFP_USER allocation and direct reclaim
could not free any pages.
The task that invoked the oom killer is simply the unlucky task that tried
an allocation that couldn't be satisified through direct reclaim. It's
usually unrelated to the task chosen for kill unless
/proc/sys/vm/oom_kill_allocating_task is enabled (which SGI requested to
avoid excessively long tasklist scans).
> > That's the problem, not using total_vm as a baseline. Replacing that with
> > rss is not going to solve the issue and reducing the user's ability to
> > specify a rough oom priority from userspace is simply not an option.
>
> OK then, if you have a solution, I would be glad to test your patch. I
> won't care much if you don't change total_vm as a baseline. Just make
> random killing history.
>
The only randomness is in selecting a task that has a different mm from
the parent in the order of its child list. Yes, that can be addressed by
doing a smarter iteration through the children before killing one of them.
Keep in mind that a heuristic as simple as this:
- kill the task that was started most recently by the same uid, or
- kill the task that was started most recently on the system if a root
task calls the oom killer,
would have yielded perfect results for your testcase but isn't necessarily
something that we'd ever want to see.
--
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/ .
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>
next prev parent reply other threads:[~2009-10-30 19:24 UTC|newest]
Thread overview: 77+ messages / expand[flat|nested] mbox.gz Atom feed top
[not found] <hav57c$rso$1@ger.gmane.org>
[not found] ` <20091013120840.a844052d.kamezawa.hiroyu@jp.fujitsu.com>
[not found] ` <hb2cfu$r08$2@ger.gmane.org>
[not found] ` <20091014135119.e1baa07f.kamezawa.hiroyu@jp.fujitsu.com>
2009-10-20 21:52 ` Vedran Furač
2009-10-26 1:55 ` KAMEZAWA Hiroyuki
2009-10-26 16:16 ` Vedran Furač
2009-10-27 3:22 ` KAMEZAWA Hiroyuki
2009-10-27 6:10 ` KOSAKI Motohiro
2009-10-27 6:34 ` Minchan Kim
2009-10-27 6:36 ` KAMEZAWA Hiroyuki
2009-10-27 6:55 ` Minchan Kim
2009-10-27 7:45 ` [RFC][PATCH] oom_kill: avoid depends on total_vm and use real RSS/swap value for oom_score (Re: " KAMEZAWA Hiroyuki
2009-10-27 7:56 ` Minchan Kim
2009-10-27 12:38 ` Andrea Arcangeli
2009-10-28 0:22 ` KAMEZAWA Hiroyuki
2009-10-28 0:45 ` Vedran Furač
2009-10-27 7:56 ` KAMEZAWA Hiroyuki
2009-10-27 8:14 ` Minchan Kim
2009-10-27 8:33 ` KAMEZAWA Hiroyuki
2009-10-27 8:52 ` Minchan Kim
2009-10-27 8:56 ` KAMEZAWA Hiroyuki
2009-10-27 17:41 ` Vedran Furač
2009-10-28 0:13 ` KAMEZAWA Hiroyuki
2009-10-27 18:39 ` Hugh Dickins
2009-10-27 18:47 ` Andrea Arcangeli
2009-10-28 0:32 ` KAMEZAWA Hiroyuki
2009-11-05 19:02 ` Pavel Machek
2009-10-28 0:28 ` KAMEZAWA Hiroyuki
2009-10-27 6:46 ` KOSAKI Motohiro
2009-10-27 6:56 ` Minchan Kim
2009-10-27 17:12 ` Vedran Furač
2009-10-27 18:02 ` KOSAKI Motohiro
2009-10-27 18:30 ` Vedran Furač
2009-10-27 20:44 ` Hugh Dickins
2009-10-27 21:04 ` David Rientjes
2009-10-28 0:08 ` Vedran Furač
2009-10-28 0:25 ` David Rientjes
2009-10-28 0:39 ` Vedran Furač
2009-10-28 4:08 ` David Rientjes
2009-10-28 4:55 ` KAMEZAWA Hiroyuki
2009-10-28 5:13 ` David Rientjes
2009-10-28 6:05 ` KAMEZAWA Hiroyuki
2009-10-28 6:17 ` David Rientjes
2009-10-28 6:20 ` KAMEZAWA Hiroyuki
2009-10-29 8:38 ` David Rientjes
2009-10-29 11:11 ` Vedran Furač
2009-10-29 19:53 ` David Rientjes
2009-10-29 23:48 ` KAMEZAWA Hiroyuki
2009-10-30 9:10 ` David Rientjes
2009-10-30 9:36 ` KAMEZAWA Hiroyuki
2009-11-03 20:49 ` David Rientjes
2009-11-04 0:50 ` KAMEZAWA Hiroyuki
2009-11-04 1:58 ` David Rientjes
2009-11-04 2:17 ` KAMEZAWA Hiroyuki
2009-11-04 3:10 ` David Rientjes
2009-11-04 3:19 ` KAMEZAWA Hiroyuki
2009-10-30 13:59 ` Vedran Furač
2009-10-30 19:24 ` David Rientjes [this message]
2009-11-02 19:58 ` Vedran Furač
2009-10-28 13:28 ` Vedran Furač
2009-10-28 20:10 ` David Rientjes
2009-10-29 3:05 ` Vedran Furač
2009-10-29 8:35 ` David Rientjes
2009-10-29 11:01 ` Vedran Furač
2009-10-29 19:42 ` David Rientjes
2009-10-30 13:53 ` Vedran Furač
2009-10-30 14:08 ` Thomas Fjellstrom
2009-10-30 15:13 ` Vedran Furač
2009-10-30 14:12 ` Andrea Arcangeli
2009-10-30 14:41 ` Vedran Furač
2009-10-30 15:15 ` Andrea Arcangeli
2009-10-30 16:24 ` Hugh Dickins
2009-11-02 19:56 ` Vedran Furač
2009-10-30 19:44 ` David Rientjes
2009-11-02 19:56 ` Vedran Furač
2009-10-28 0:43 ` KAMEZAWA Hiroyuki
2009-10-28 2:47 ` KOSAKI Motohiro
2009-10-28 3:17 ` KAMEZAWA Hiroyuki
2009-10-28 4:12 ` David Rientjes
2009-10-28 8:10 ` Hugh Dickins
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=alpine.DEB.2.00.0910301218410.31986@chino.kir.corp.google.com \
--to=rientjes@google.com \
--cc=aarcange@redhat.com \
--cc=akpm@linux-foundation.org \
--cc=hugh.dickins@tiscali.co.uk \
--cc=kamezawa.hiroyu@jp.fujitsu.com \
--cc=kosaki.motohiro@jp.fujitsu.com \
--cc=linux-kernel@vger.kernel.org \
--cc=linux-mm@kvack.org \
--cc=minchan.kim@gmail.com \
--cc=vedran.furac@gmail.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox