xfs
[Top] [All Lists]

Re: 2.6.20-rc5: cp 18gb 18gb.2 = OOM killer, reproducible just like 2.16

To: Justin Piszcz <jpiszcz@xxxxxxxxxxxxxxx>
Subject: Re: 2.6.20-rc5: cp 18gb 18gb.2 = OOM killer, reproducible just like 2.16.19.2
From: Mark Hahn <hahn@xxxxxxxxxxx>
Date: Thu, 25 Jan 2007 17:34:31 -0500 (EST)
Cc: Pavel Machek <pavel@xxxxxx>, linux-kernel@xxxxxxxxxxxxxxx, linux-raid@xxxxxxxxxxxxxxx, xfs@xxxxxxxxxxx
In-reply-to: <Pine.LNX.4.64.0701250407590.3673@xxxxxxxxxxxxxxxx>
References: <Pine.LNX.4.64.0701211424170.2552@xxxxxxxxxxxxxxxx> <20070122133735.GB4493@xxxxxx> <Pine.LNX.4.64.0701221348020.11525@xxxxxxxxxxxxxxxx> <Pine.LNX.4.64.0701241837550.2930@xxxxxxxxxxxxxxxx> <20070125003242.GA23343@xxxxxxxxxx> <Pine.LNX.4.64.0701250407590.3673@xxxxxxxxxxxxxxxx>
Sender: xfs-bounce@xxxxxxxxxxx
Something is seriously wrong with that OOM killer.

do you know you don't have to operate in OOM-slaughter mode?

"vm.overcommit_memory = 2" in your /etc/sysctl.conf puts you into a mode where the kernel tracks your "committed" memory needs, and will eventually cause some allocations to fail.
this is often much nicer than the default random OOM slaughter.
(you probably also need to adjust vm.overcommit_ratio with some knowlege of your MemTotal and SwapTotal.)

regards, mark hahn.


<Prev in Thread] Current Thread [Next in Thread>