linux-btrfs.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: Bob Marley <bobmarley@shiftmail.org>
To: linux-btrfs <linux-btrfs@vger.kernel.org>
Subject: Performance reduces with nodatasum
Date: Sat, 04 Oct 2014 12:26:27 +0200	[thread overview]
Message-ID: <542FCB53.1070201@shiftmail.org> (raw)

Hello,
apparently I have found an issue with btrfs: performance reduces with 
nodatasum and multi-device "raid0" or "single".

I was testing with a series of 8 LIO ramdisks, with btrfs on those in 
multi-device single mode, and writing zeroes on the filesystem with 16 
dd in parallel.
Performance decreases significantly if the filesystem is mounted with 
nodatasum, or with nodatacow which implies nodatasum.
CPU occupation also reduces, together with speed, as seen with htop.

At first I thought it was my problem, but then I saw this web page
http://www.linux-mag.com/id/7308/3/
which also reports reduced performance with nodatasum and multi-device 
raid0 or single, e.g. see these two lines:

Btrfs
two disks,
single
standard
50.144    50.264    126.984    131.130

Btrfs
two disks,
single
nodatacow,
nodatasum
43.834    47.603    131.612    131.470

similarly with raid0

even more with compression:

Btrfs
two disks,
raid0
-o compress
70.234    69.048    130.852    129928
Btrfs

two disks,
raid0
-o compress
nodatacow,
nodatasum
48.762    48.831    130.812    130.202


If you go higher with the performances, such as with ramdisks, in the 
GB/sec range, it reduces more than that. I have noticed upto 50% reduction.

It would be important to fix this problem for high-performance usages of 
btrfs.

Best regards
BM


             reply	other threads:[~2014-10-04 10:34 UTC|newest]

Thread overview: 3+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2014-10-04 10:26 Bob Marley [this message]
2014-10-04 10:36 ` Performance reduces with nodatasum Bob Marley
2014-10-04 10:52   ` Bob Marley

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=542FCB53.1070201@shiftmail.org \
    --to=bobmarley@shiftmail.org \
    --cc=linux-btrfs@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).