From: Andrew Morton <akpm@linux-foundation.org>
To: Mike Galbraith <efault@gmx.de>
Cc: Linus Torvalds <torvalds@linux-foundation.org>,
LKML <linux-kernel@vger.kernel.org>,
Jens Axboe <jens.axboe@oracle.com>
Subject: Re: [ext3][kernels >= 2.6.20.7 at least] KDE going comatose when FS is under heavy write load (massive starvation)
Date: Sat, 28 Apr 2007 00:01:42 -0700 [thread overview]
Message-ID: <20070428000142.2acab38a.akpm@linux-foundation.org> (raw)
In-Reply-To: <1177741953.6268.18.camel@Homer.simpson.net>
On Sat, 28 Apr 2007 08:32:32 +0200 Mike Galbraith <efault@gmx.de> wrote:
> On Sat, 2007-04-28 at 06:25 +0200, Mike Galbraith wrote:
> > On Fri, 2007-04-27 at 08:18 -0700, Linus Torvalds wrote:
> >
> > > Actually, you don't need to apply the patch - just do
> > >
> > > echo 5 > /proc/sys/vm/dirty_background_ratio
> > > echo 10 > /proc/sys/vm/dirty_ratio
> >
> > That seems to have done the trick. Amarok and GUI aren't exactly speed
> > demons while writeout is happening, but they are not hanging for
> > eternities.
>
> As promised, I tested with a kernel that I know for fact that I have
> tested heavy IO on previously, and behavior was identically horrid, so
> it's not something new that snuck in ~recently, my disk just got a _lot_
> fuller in the meantime (12k mp3s munch a lot).
Just to clarify here - you're saying that some older kernel is as sucky as
2.6.21, and that (presumably) dropping the dirty ratios makes things a bit
better on the old kernel as well?
> I also verified that I don't need to use the dirty data restrictions
> with ext2, all is just peachy using stock settings. Amarok switches
> songs quickly, and GUI doesn't hang. Behavior is that expected of a
> heavily loaded IO subsystem, and is 1000% better than ext3 with my very
> full disk.
Yes, the very full disk could explain why things are _so_ bad. Not only
does fsync() force vast amounts of writeout, it's also seeky writeout.
> Journaling is very nice, but I think I'll be much better off without it
> responsiveness wise.
Well, physical journalling with ordered data is bad here. Other forms of
journalling which don't introduce this great contention point shouldn't be
as bad.
Actually, I'm surprised that data=writeback didn't help much. If the
present theories are correct it should have helped quite a lot, because in
data=writeback mode fsync(small-file) will not cause
fdatasync(everything-else).
next prev parent reply other threads:[~2007-04-28 7:01 UTC|newest]
Thread overview: 66+ messages / expand[flat|nested] mbox.gz Atom feed top
2007-04-27 7:59 [ext3][kernels >= 2.6.20.7 at least] KDE going comatose when FS is under heavy write load (massive starvation) Mike Galbraith
2007-04-27 8:33 ` Andrew Morton
2007-04-27 9:23 ` Mike Galbraith
2007-04-27 10:17 ` Mike Galbraith
2007-04-27 11:59 ` Marat Buharov
2007-04-27 12:30 ` Peter Zijlstra
2007-04-27 13:50 ` Mark Lord
2007-04-27 12:39 ` Manoj Joseph
2007-04-27 15:30 ` Linus Torvalds
2007-04-27 19:31 ` Andreas Dilger
2007-04-27 19:44 ` Mike Galbraith
2007-04-27 19:50 ` Linus Torvalds
2007-04-27 20:05 ` Hua Zhong
2007-04-27 20:12 ` Miquel van Smoorenburg
2007-04-27 20:12 ` Bill Huey
2007-04-28 5:37 ` Mikulas Patocka
2007-04-28 5:45 ` Mikulas Patocka
2007-04-28 21:57 ` Bill Huey
2007-04-28 22:38 ` Mikulas Patocka
2007-04-27 20:29 ` Gabriel C
2007-04-27 20:45 ` Stephen Clark
2007-04-27 20:54 ` Manoj Joseph
2007-04-28 8:45 ` Matthias Andree
2007-04-27 22:18 ` Andrew Morton
2007-05-03 17:38 ` Alex Tomas
2007-05-03 23:54 ` Andrew Morton
2007-05-04 6:18 ` Alex Tomas
2007-05-04 6:38 ` Andrew Morton
2007-05-04 6:57 ` Alex Tomas
2007-05-04 7:18 ` Andrew Morton
2007-05-04 7:39 ` Alex Tomas
2007-05-04 8:02 ` Andrew Morton
2007-04-28 8:44 ` Matthias Andree
2007-04-28 20:46 ` Mikulas Patocka
2007-04-28 21:12 ` Lee Revell
2007-04-29 20:49 ` Mark Lord
2007-04-29 21:17 ` Mikulas Patocka
2007-04-27 15:18 ` Linus Torvalds
2007-04-27 15:41 ` John Anthony Kazos Jr.
2007-04-27 15:54 ` Linus Torvalds
2007-04-27 16:24 ` Chuck Ebbert
2007-04-27 19:43 ` Marko Macek
2007-04-27 18:31 ` Andrew Morton
2007-04-27 19:09 ` Zan Lynx
2007-04-27 22:07 ` Andrew Morton
2007-04-27 19:27 ` Mike Galbraith
2007-04-28 8:51 ` Matthias Andree
2007-04-28 8:59 ` Andrew Morton
2007-04-28 16:30 ` Linus Torvalds
2007-04-28 16:56 ` Paolo Ornati
2007-04-27 19:28 ` Mike Galbraith
2007-04-27 20:06 ` Jan Engelhardt
2007-04-27 21:22 ` Linus Torvalds
2007-04-28 4:25 ` Mike Galbraith
2007-04-28 6:32 ` Mike Galbraith
2007-04-28 7:01 ` Andrew Morton [this message]
2007-04-28 7:12 ` Mike Galbraith
2007-04-28 6:32 ` Mikulas Patocka
2007-04-28 16:05 ` Linus Torvalds
2007-04-28 16:37 ` Ingo Molnar
2007-04-28 17:11 ` Mikulas Patocka
2007-04-30 6:57 ` Jens Axboe
2007-04-28 17:55 ` Mikulas Patocka
2007-04-30 6:56 ` Jens Axboe
2007-05-02 6:53 ` Jens Axboe
2007-05-02 7:36 ` Mike Galbraith
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=20070428000142.2acab38a.akpm@linux-foundation.org \
--to=akpm@linux-foundation.org \
--cc=efault@gmx.de \
--cc=jens.axboe@oracle.com \
--cc=linux-kernel@vger.kernel.org \
--cc=torvalds@linux-foundation.org \
/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