From: Daniel Phillips <phillips@bonn-fries.net>
To: Ed Tomlinson <tomlins@cam.org>, "Stephen C. Tweedie" <sct@redhat.com>
Cc: Chris Mason <mason@suse.com>,
linux-kernel@vger.kernel.org, linux-mm@kvack.org
Subject: Re: [RFC] using writepage to start io
Date: Tue, 7 Aug 2001 20:36:20 +0200 [thread overview]
Message-ID: <0108072036200D.02365@starship> (raw)
In-Reply-To: <755760000.997128720@tiny> <20010807120234.D4036@redhat.com> <20010807113944.D229E7B53@oscar.casa.dyndns.org>
In-Reply-To: <20010807113944.D229E7B53@oscar.casa.dyndns.org>
On Tuesday 07 August 2001 13:39, Ed Tomlinson wrote:
> On August 7, 2001 07:02 am, Stephen C. Tweedie wrote:
> > On Mon, Aug 06, 2001 at 11:18:26PM +0200, Daniel Phillips wrote:
> > > > On Monday, August 06, 2001 09:45:12 PM +0200 Daniel Phillips
> > > >
> > > > Grin, we're talking in circles. My point is that by having two
> > > > threads, bdflush is allowed to skip over older buffers in favor
> > > > of younger ones because somebody else is responsible for writing
> > > > the older ones out.
> > >
> > > Yes, and you can't imagine an algorithm that could do that with
> > > *one* thread?
> >
> > FWIW, we've seen big performance degradations in the past when
> > testing different ext3 checkpointing modes. You can't reuse a disk
> > block in the journal without making sure that the data in it has
> > been flushed to disk, so ext3 does regular checkpointing to flush
> > journaled blocks out. That can interact very badly with normal VM
> > writeback if you're not careful: having two threads doing the same
> > thing at the same time can just thrash the disk.
> >
> > Parallel sync() calls from multiple processes has shown up the same
> > behaviour on ext2 in the past. I'd definitely like to see at most
> > one thread of writeback per disk to avoid that.
>
> Be carefull here. I have a system (solaris) at the office that has 96
> drives on it. Do we really want 96 writeback threads? With 96
> drives, suspect the bus bandwidth would be the limiting factor.
Surely these are grouped into some kind of raid? You'd have one queue
per raid. Since the buffer submission is nonblocking[1] it's a matter
of taste whether that translates into multiple threads or not.
[1] So long as you don't run into the low level request limits which
should never happen if you pay attention to how much IO is in flight.
--
Daniel
next prev parent reply other threads:[~2001-08-07 18:30 UTC|newest]
Thread overview: 30+ messages / expand[flat|nested] mbox.gz Atom feed top
2001-08-05 18:34 [RFC] using writepage to start io Chris Mason
2001-08-05 22:38 ` Daniel Phillips
2001-08-05 23:32 ` Chris Mason
2001-08-06 5:39 ` Daniel Phillips
2001-08-06 13:24 ` Chris Mason
2001-08-06 16:13 ` Daniel Phillips
2001-08-06 16:51 ` Chris Mason
2001-08-06 19:45 ` Daniel Phillips
2001-08-06 20:12 ` Chris Mason
2001-08-06 21:18 ` Daniel Phillips
2001-08-07 11:02 ` Stephen C. Tweedie
2001-08-07 11:39 ` Ed Tomlinson
2001-08-07 18:36 ` Daniel Phillips [this message]
2001-08-07 12:07 ` Anton Altaparmakov
2001-08-07 12:02 ` Anton Altaparmakov
2001-08-07 13:29 ` Daniel Phillips
2001-08-07 13:31 ` Alexander Viro
2001-08-07 15:52 ` Daniel Phillips
2001-08-07 14:23 ` Stephen C. Tweedie
2001-08-07 15:51 ` Daniel Phillips
2001-08-08 14:49 ` Stephen C. Tweedie
2001-08-06 15:13 ` Eric W. Biederman
-- strict thread matches above, loose matches on Subject: below --
2001-08-07 15:19 Chris Mason
2001-07-28 16:01 Chris Mason
2001-07-28 16:18 ` Alexander Viro
2001-07-28 16:25 ` Chris Mason
2001-07-31 19:07 ` Chris Mason
2001-08-01 1:01 ` Daniel Phillips
2001-08-01 2:05 ` Chris Mason
2001-08-01 14:57 ` Daniel Phillips
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=0108072036200D.02365@starship \
--to=phillips@bonn-fries.net \
--cc=linux-kernel@vger.kernel.org \
--cc=linux-mm@kvack.org \
--cc=mason@suse.com \
--cc=sct@redhat.com \
--cc=tomlins@cam.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;
as well as URLs for NNTP newsgroup(s).