mirror of https://lore.kernel.org/lkml/
 help / color / mirror / Atom feed
From: Miquel van Smoorenburg <miquels@cistron.nl>
To: Jens Axboe <axboe@suse.de>
Cc: Miquel van Smoorenburg <miquels@cistron.nl>,
	linux-lvm@sistina.com, linux-kernel@vger.kernel.org,
	Joe Thornber <thornber@redhat.com>
Subject: Re: IO scheduler, queue depth, nr_requests
Date: Thu, 19 Feb 2004 00:52:43 +0100	[thread overview]
Message-ID: <20040218235243.GA30621@drinkel.cistron.nl> (raw)
In-Reply-To: <20040217145716.GE30438@traveler.cistron.net> (from miquels@cistron.nl on Tue, Feb 17, 2004 at 15:57:16 +0100)

On Tue, 17 Feb 2004 15:57:16, Miquel van Smoorenburg wrote:
> For some reason, when using LVM, write requests get queued out
> of order to the 3ware controller, which results in quite a bit
> of seeking and thus performance loss.
[..]
> Okay I repeated some earlier tests, and I added some debug code in
> several places.
> 
> I added logging to tw_scsi_queue() in the 3ware driver to log the
> start sector and length of each request. It logs something like:
> 3wdbg: id 119, lba = 0x2330bc33, num_sectors = 256
> 
> With a perl script, I can check if the requests are sent to the
> host in order. That outputs something like this:
> 
> Consecutive: start 1180906348, length 7936 sec (3968 KB), requests: 31
> Consecutive: start 1180906340, length 8 sec (4 KB), requests: 1
> Consecutive: start 1180914292, length 7936 sec (3968 KB), requests: 31
> Consecutive: start 1180914284, length 8 sec (4 KB), requests: 1
> Consecutive: start 1180922236, length 7936 sec (3968 KB), requests: 31
> Consecutive: start 1180922228, length 8 sec (4 KB), requests: 1
> Consecutive: start 1180930180, length 7936 sec (3968 KB), requests: 31
> 
> See, 31 requests in order, then one request "backwards", then 31 in order, etc.

I found out what causes this. It's get_request_wait().

When the request queue is full, and a new request needs to be created,
__make_request() blocks in get_request_wait().

Another process wakes up first (pdflush / process submitting I/O itself /
xfsdatad / etc) and sends the next bio's to __make_request().
In the mean time some free requests have become available, and the bios
are merged into a new request. Those requests are submitted to the device.

Then, get_request_wait() returns but the bio is not mergeable anymore -
and that results in a backwards seek, severely limiting the I/O rate.

Wouldn't it be better to allow the request allocation and queue the
request, and /then/ put the process to sleep ? The queue will grow larger
than nr_requests, but it does that anyway.

Mike.

       reply	other threads:[~2004-02-18 23:53 UTC|newest]

Thread overview: 29+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
     [not found] <20040216131609.GA21974@cistron.nl>
     [not found] ` <20040216133047.GA9330@suse.de>
     [not found]   ` <20040217145716.GE30438@traveler.cistron.net>
2004-02-18 23:52     ` Miquel van Smoorenburg [this message]
2004-02-19  1:24       ` Nick Piggin
2004-02-19  1:52         ` Miquel van Smoorenburg
2004-02-19  2:01           ` Nick Piggin
2004-02-19  1:26       ` Andrew Morton
2004-02-19  2:11         ` Miquel van Smoorenburg
2004-02-19  2:26           ` Andrew Morton
2004-02-19 10:15             ` Miquel van Smoorenburg
2004-02-19 10:19               ` Jens Axboe
2004-02-19 20:59                 ` Miquel van Smoorenburg
2004-02-19 22:52                   ` Nick Piggin
2004-02-19 23:53                     ` Miquel van Smoorenburg
2004-02-20  0:15                       ` Nick Piggin
2004-02-20  1:12                       ` [PATCH] per process request limits (was Re: IO scheduler, queue depth, nr_requests) Nick Piggin
2004-02-20  1:26                         ` Andrew Morton
2004-02-20  1:40                           ` Nick Piggin
2004-02-20  2:32                             ` Andrew Morton
2004-02-20 14:40                               ` [PATCH] bdi_congestion_funp (was: Re: [PATCH] per process request limits (was Re: IO scheduler, queue depth, nr_requests)) Miquel van Smoorenburg
2004-02-20 14:57                                 ` Jens Axboe
2004-02-20 14:59                                 ` Joe Thornber
2004-02-20 15:00                                   ` Jens Axboe
2004-02-22 14:02                                     ` Miquel van Smoorenburg
2004-02-22 19:55                                       ` Andrew Morton
2004-02-20  1:45                         ` [PATCH] per process request limits (was Re: IO scheduler, queue depth, nr_requests) Nick Piggin
2004-02-19  2:51           ` IO scheduler, queue depth, nr_requests Nick Piggin
2004-02-19 10:21             ` Jens Axboe
     [not found] <1qJVx-75K-15@gated-at.bofh.it>
     [not found] ` <1qJVx-75K-17@gated-at.bofh.it>
     [not found]   ` <1qJVw-75K-11@gated-at.bofh.it>
     [not found]     ` <1qLb8-6m-27@gated-at.bofh.it>
     [not found]       ` <1qLXl-XV-17@gated-at.bofh.it>
     [not found]         ` <1qMgF-1dA-5@gated-at.bofh.it>
     [not found]           ` <1qTs3-7A2-51@gated-at.bofh.it>
     [not found]             ` <1qTBB-7Hh-7@gated-at.bofh.it>
     [not found]               ` <1r3AS-1hW-5@gated-at.bofh.it>
     [not found]                 ` <1r5jD-2RQ-31@gated-at.bofh.it>
     [not found]                   ` <1r6fH-3L8-11@gated-at.bofh.it>
     [not found]                     ` <1r6S4-6cv-1@gated-at.bofh.it>
2004-02-25 20:17                       ` Bill Davidsen
2004-02-25 21:39                         ` Miquel van Smoorenburg
2004-02-26  0:39                         ` Nick Piggin

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=20040218235243.GA30621@drinkel.cistron.nl \
    --to=miquels@cistron.nl \
    --cc=axboe@suse.de \
    --cc=linux-kernel@vger.kernel.org \
    --cc=linux-lvm@sistina.com \
    --cc=thornber@redhat.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

all inboxes | Powered by JetHome®