linux-raid.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: Charles Polisher <cpolish@surewest.net>
To: Peter Rabbitson <rabbit+list@rabbit.us>
Cc: Phil Turmel <philip@turmel.org>, linux-raid@vger.kernel.org
Subject: Re: Suboptimal raid6 linear read speed
Date: Tue, 15 Jan 2013 09:09:38 -0800	[thread overview]
Message-ID: <20130115170937.GA8831@kevin> (raw)
In-Reply-To: <20130115125507.GA12184@rabbit.us>

On Tue, Jan 15, 2013 at 11:55:07PM +1100, Peter Rabbitson wrote:
> On Tue, Jan 15, 2013 at 07:49:10AM -0500, Phil Turmel wrote:
> > You are neglecting each drive's need to skip over parity blocks.  If the
> > array's chunk size is small, the drives won't have to seek, just wait
> > for the platter spin.  Larger chunks might need a seek.
> 
> > Either way, you
> > won't get better than (single drive rate) * (n-2) where "n" is the
> > number of drives in your array. (Large sequential reads.)
> 
> This can't be right. As far as I know the md layer is smarter than that, and
> includes various anticipatory codepaths specifically to leverage multiple
> drives in this fashion. Fwiw raid5 does give me the near-expected speed
> (n * single drive).

Happen to be working with comparative benchmarks looking for
relative throughput, varying the number of active drives in the
array and the RAID level. Clearly in this data RAID6 sequential
writes are bottlenecked by the 2 parity stripes. RAID6 setup
increases from 2 non-parity drives in the 4 drive configuration
to 6 non-parity drives in the 8 drive configuration, so one
might hope for 3x advantage. Yet the data show an advantage of
only 1.83 for reads. My guess is the need to read the parity
stripes is again a limiting factor. Next benchmark will vary
stripe and stride.

                                        Advantage     Advantage
                                        vs 4 drives   vs RAID0
Config  Drives  Seq write   Seq  read   Write  Read   Write Read
------  ------  ----------  ----------  ----- -----   ----  ----
RAID0   4        8.1MB/sec   9.3MB/sec   1.00  1.00   1.00  1.00
RAID0   8       16.8MB/sec  15.0MB/sec   2.07  1.61   1.00  1.00
 
RAID1   4        2.1MB/sec   3.6MB/sec   1.00  1.00   0.25  0.38
RAID1   8        1.6MB/sec   3.6MB/sec   0.76  1.00   0.09  0.24

RAID5   4       16.8MB/sec   9.1MB/sec   1.00  1.00   2.07  0.97
RAID5   8       17.2MB/sec  14.9MB/sec   1.02  1.63   2.12  1.60

RAID6   4       12.6MB/sec   7.9MB/sec   1.00  1.00   1.55  0.84
RAID6   8       14.4MB/sec  14.5MB/sec   1.63  1.83   1.77  1.55

RAID10  4        4.0MB/sec   7.3MB/sec   1.00  1.00   0.49  0.78
RAID10  8        6.3MB/sec  13.4MB/sec   1.57  1.83   0.37  0.89

Yes, these drives are *really* slow (Connor CP 30548). 
The math doesn't change.
-- 
Charles



  reply	other threads:[~2013-01-15 17:09 UTC|newest]

Thread overview: 46+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2013-01-15 12:33 Suboptimal raid6 linear read speed Peter Rabbitson
2013-01-15 12:45 ` Mikael Abrahamsson
2013-01-15 12:56   ` Peter Rabbitson
2013-01-15 16:13     ` Mikael Abrahamsson
2013-01-15 12:49 ` Phil Turmel
2013-01-15 12:55   ` Peter Rabbitson
2013-01-15 17:09     ` Charles Polisher [this message]
2013-01-15 19:57       ` keld
2013-01-16  4:43         ` Charles Polisher
2013-01-16  6:37           ` Tommy Apel Hansen
2013-01-16  9:36           ` keld
2013-01-16 16:09             ` Charles Polisher
2013-01-16 20:40               ` EJ Vincent
2013-01-15 23:17     ` Phil Turmel
2013-01-16  2:48     ` Stan Hoeppner
2013-01-16  2:58       ` Peter Rabbitson
2013-01-16 20:29         ` Stan Hoeppner
2013-01-16 21:20           ` Roy Sigurd Karlsbakk
2013-01-17 15:51           ` Mikael Abrahamsson
2013-01-18  8:31             ` Stan Hoeppner
2013-01-18  9:18               ` Mikael Abrahamsson
2013-01-18 22:56                 ` Stan Hoeppner
2013-01-19  7:43                   ` Mikael Abrahamsson
2013-01-19 22:48                     ` Stan Hoeppner
2013-01-19 23:51                       ` Maarten
2013-01-20  0:16                         ` Chris Murphy
2013-01-20  0:49                           ` Maarten
2013-01-20  1:37                             ` Phil Turmel
2013-01-20  9:44                             ` Chris Murphy
2013-01-20  6:26                           ` Mikael Abrahamsson
2013-01-20  9:39                             ` Chris Murphy
2013-01-20 16:55                               ` Mikael Abrahamsson
2013-01-20 17:15                                 ` Chris Murphy
2013-01-20 17:17                                   ` Mikael Abrahamsson
2013-01-20 17:20                                     ` Chris Murphy
2013-01-19 23:53                       ` Phil Turmel
2013-01-20  9:04                     ` Wolfgang Denk
2013-01-20 19:28                     ` Peter Grandi
2013-01-20 21:09                       ` Mikael Abrahamsson
2013-01-20 21:50                         ` Peter Grandi
2013-01-21  5:24                           ` Mikael Abrahamsson
2013-01-21 14:40                       ` Peter Rabbitson
2013-01-21 20:32                         ` Peter Grandi
2013-01-21 20:55                           ` Peter Grandi
2013-01-21 22:00                         ` Peter Grandi
2013-01-19 13:21                   ` Roy Sigurd Karlsbakk

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=20130115170937.GA8831@kevin \
    --to=cpolish@surewest.net \
    --cc=linux-raid@vger.kernel.org \
    --cc=philip@turmel.org \
    --cc=rabbit+list@rabbit.us \
    /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).