From: "Stephen C. Tweedie" <sct@redhat.com>
To: Daryll Strauss <daryll@valinux.com>
Cc: "Stephen C. Tweedie" <sct@redhat.com>, Dan Kegel <dank@kegel.com>,
"linux-kernel@vger.kernel.org" <linux-kernel@vger.kernel.org>
Subject: Re: O_DIRECT please; Sybase 12.5
Date: Tue, 3 Jul 2001 16:48:33 +0100 [thread overview]
Message-ID: <20010703164833.F28793@redhat.com> (raw)
In-Reply-To: <3B3C4CB4.6B3D2B2F@kegel.com> <20010703104253.B29868@redhat.com> <20010703081039.C3942@newbie>
In-Reply-To: <20010703081039.C3942@newbie>; from daryll@valinux.com on Tue, Jul 03, 2001 at 08:10:39AM -0700
Hi,
On Tue, Jul 03, 2001 at 08:10:39AM -0700, Daryll Strauss wrote:
> I recall hearing about a problem with the md device and raw IO. It was
> something about the block sizes not matching causing performance
> problems. Has anything been done to improve those issues?
The problem is a combination of two things. First, raw IO is always
fully synchronous, so with raw IO (and O_DIRECT) you are, in effect,
explicitly instructing the kernel not to do any readahead. That makes
it hard to keep two disks running in parallel with soft raid if you
are using small IOs, obviously.
Secondly, raw IO pins buffers in physical memory, and to avoid
causing serious VM problems due to having too much unswappable memory
pinned by arbitrary applications, the current raw IO driver limits the
pinned chunk size to 64k. That, combined with the sequential nature
of raw IO, can limit performance, certainly.
Raw IO is quite capable of running with larger chunk sizes, but we
really need a kernel limiter of some description to prevent users from
using this mechanism to pin massive amounts of memory for raw IO at
once. There are several candidate mechanisms for that, but none in
the main kernel right now.
Cheers,
Stephen
next prev parent reply other threads:[~2001-07-03 16:14 UTC|newest]
Thread overview: 15+ messages / expand[flat|nested] mbox.gz Atom feed top
2001-06-29 9:39 O_DIRECT please; Sybase 12.5 Dan Kegel
2001-06-29 9:50 ` Alan Cox
2001-06-29 10:16 ` Dan Kegel
2001-06-29 12:49 ` Mike Harrold
2001-07-05 13:59 ` Andrea Arcangeli
2001-06-29 15:23 ` Steve Lord
2001-07-03 9:42 ` Stephen C. Tweedie
2001-07-03 15:10 ` Daryll Strauss
2001-07-03 15:48 ` Stephen C. Tweedie [this message]
2001-07-05 13:53 ` Andrea Arcangeli
2001-07-05 14:28 ` Andrew Morton
2001-07-05 14:37 ` Andrea Arcangeli
2001-07-05 15:06 ` Andrew Morton
2001-07-06 0:25 ` Keith Owens
[not found] <3B3C4CB4.6B3D2B2F@kegel.com.suse.lists.linux.kernel>
2001-06-29 10:42 ` Andi Kleen
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=20010703164833.F28793@redhat.com \
--to=sct@redhat.com \
--cc=dank@kegel.com \
--cc=daryll@valinux.com \
--cc=linux-kernel@vger.kernel.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).