All of lore.kernel.org
 help / color / mirror / Atom feed
From: Phil Turmel <philip@turmel.org>
To: Dallas Clement <dallas.a.clement@gmail.com>,
	Mark Knecht <markknecht@gmail.com>
Cc: John Stoffel <john@stoffel.org>, Linux-RAID <linux-raid@vger.kernel.org>
Subject: Re: best base / worst case RAID 5,6 write speeds
Date: Tue, 15 Dec 2015 08:53:18 -0500	[thread overview]
Message-ID: <56701B4E.8060007@turmel.org> (raw)
In-Reply-To: <CAE9DZUQxT_5L0bW5m9SZ_d2GU6sZS8k0qD=g+o112qM4V=cJkw@mail.gmail.com>

Hi Dallas,

On December 14, 2015 9:36:05 PM EST, Dallas Clement

>Hi Everyone. I have some very interesting news to report. I did a
>little bit more playing around with fio, doing sequential writes to a
>RAID 5 device with all 12 disks. I kept the block size at the 128K
>chunk aligned value of 1408K. But this time I varied the queue depth.
>These are my results for writing a 10 GB of data:
>
>iodepth=1 => 642 MB/s, # of RMWs = 11
>
>iodepth=4 => 1108 MB/s, # of RMWs = 6
>
>iodepth=8 => 895 MB/s, # of RMWs = 7
>
>iodepth=16 => 855 MB/s, # of RMWs = 11
>
>iodepth=32 => 936 MB/s, # of RMWs = 11
>
>iodepth=64 => 551 MB/s, # of RMWs = 5606
>
>iodepth=128 => 554 MB/s, # of RMWs = 6333
>
>As you can see, something goes terribly wrong with async i/o with
>iodepth >= 64. Btw, not to be contentious Phil, I have checked
>multiple fio man pages and they clearly indicate that iodepth is for
>async i/o which this is (libaio). I don't see any mention of
>sequential writes being prohibited with async i/o. See
>https://github.com/axboe/fio/blob/master/HOWTO.

Hmmm. I misread that part. But do note the comment that you might not
achieve as many in-flight I/Os as you expect.

>However, maybe I'm
>missing something and it sure looks from these results that there may
>be a connection.
>
>This is my fio job config:
>
>[job]
>ioengine=libaio
>iodepth=128
>prio=0
>rw=write
>bs=1408k
>filename=/dev/md10
>numjobs=1
>size=10g
>direct=1
>invalidate=1
>
>Incidentally, the very best write speed here (1108 MB/s with
>iodepth=4) comes out to about 100 MB/s per disk, which is pretty close
>to the worst case inner disk speed of 95.5 MB/s I had recorded
>earlier.

Very interesting indeed. I wonder if the extra I/O in flight at high
depths is consuming all available stripe cache space, possibly not
consistently. I'd raise and lower that in various combinations with
various combinations of iodepth.  Running out of stripe cache will cause
premature RMWs.

Regards,

Phil

  reply	other threads:[~2015-12-15 13:53 UTC|newest]

Thread overview: 60+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2015-12-10  1:34 best base / worst case RAID 5,6 write speeds Dallas Clement
2015-12-10  6:36 ` Alexander Afonyashin
2015-12-10 14:38   ` Dallas Clement
2015-12-10 15:14 ` John Stoffel
2015-12-10 18:40   ` Dallas Clement
     [not found]     ` <CAK2H+ed+fe5Wr0B=h5AzK5_=ougQtW_6cJcUG_S_cg+WfzDb=Q@mail.gmail.com>
2015-12-10 19:26       ` Dallas Clement
2015-12-10 19:33         ` John Stoffel
2015-12-10 22:19           ` Wols Lists
2015-12-10 19:28     ` John Stoffel
2015-12-10 22:23       ` Wols Lists
2015-12-10 20:06 ` Phil Turmel
2015-12-10 20:09   ` Dallas Clement
2015-12-10 20:29     ` Phil Turmel
2015-12-10 21:14       ` Dallas Clement
2015-12-10 21:32         ` Phil Turmel
     [not found]     ` <CAK2H+ednN7dCGzcOt8TxgNdhdDA1mN6Xr5P8vQ+Y=-uRoxRksw@mail.gmail.com>
2015-12-11  0:02       ` Dallas Clement
     [not found]         ` <CAK2H+efF2dM1BsM7kzfTxMdQEHvbWRaVe7zJLTGcPZzafn2M6A@mail.gmail.com>
2015-12-11  0:41           ` Dallas Clement
2015-12-11  1:19             ` Dallas Clement
     [not found]               ` <CAK2H+ec-zMbhxoFyHXLkdM-z-9cYYzNbPFhn19XjTHqrOMDZKQ@mail.gmail.com>
2015-12-11 15:44                 ` Dallas Clement
2015-12-11 16:32                   ` John Stoffel
2015-12-11 16:47                     ` Dallas Clement
2015-12-11 19:34                       ` John Stoffel
2015-12-11 21:24                         ` Dallas Clement
2015-12-11 23:30                           ` Dallas Clement
2015-12-12  0:00                             ` Dallas Clement
2015-12-12  0:38                               ` Phil Turmel
2015-12-12  2:55                                 ` Dallas Clement
2015-12-12  4:47                                   ` Phil Turmel
2015-12-14 20:14                                     ` Dallas Clement
     [not found]                                       ` <CAK2H+edazVORrVovWDeTA8DmqUL+5HRH-AcRwg8KkMas=o+Cog@mail.gmail.com>
2015-12-14 20:55                                         ` Dallas Clement
     [not found]                                           ` <CAK2H+ed-3Z8SR20t8rpt3Fb48c3X2Jft=qZoiY9emC2nQww1xQ@mail.gmail.com>
2015-12-14 21:20                                             ` Dallas Clement
2015-12-14 22:05                                               ` Dallas Clement
2015-12-14 22:31                                                 ` Tommy Apel
     [not found]                                                 ` <CAK2H+ecMvDLdYLhMtMQbP7Ygw-VohG7LGZ2n7H+LAXQ1waJK3A@mail.gmail.com>
2015-12-14 23:25                                                   ` Dallas Clement
2015-12-15  2:36                                                     ` Dallas Clement
2015-12-15 13:53                                                       ` Phil Turmel [this message]
2015-12-15 14:09                                                       ` Robert Kierski
2015-12-15 15:14                                                       ` John Stoffel
2015-12-15 17:30                                                         ` Dallas Clement
2015-12-15 19:22                                                           ` Phil Turmel
2015-12-15 19:44                                                             ` Dallas Clement
2015-12-15 19:52                                                               ` Phil Turmel
2015-12-15 21:54                                                           ` John Stoffel
2015-12-15 23:07                                                             ` Dallas Clement
2015-12-16 15:31                                                               ` Dallas Clement
     [not found]                                                                 ` <CAK2H+eeD2k4yzuvL4uF_qKycp6A=XPe8pVF_J-7Agi8Ze89PPQ@mail.gmail.com>
2015-12-17  5:57                                                                   ` Dallas Clement
2015-12-17 13:41                                                                   ` Phil Turmel
2015-12-17 21:08                                                                     ` Dallas Clement
2015-12-17 22:40                                                                       ` Phil Turmel
2015-12-17 23:28                                                                         ` Dallas Clement
2015-12-18  0:54                                                                           ` Dallas Clement
     [not found]                                                                             ` <CAFx4rwT8xgwZ0OWaLLsZvhMskiwmY54MzHgnnEPaswByeRrXxQ@mail.gmail.com>
2015-12-22  6:15                                                                               ` Doug Dumitru
2015-12-22 14:34                                                                                 ` Robert Kierski
2015-12-22 16:48                                                                                 ` Dallas Clement
2015-12-22 18:33                                                                                   ` Doug Dumitru
2016-01-04 18:56                                                                                     ` Robert Kierski
2016-01-04 19:13                                                                                       ` Doug Dumitru
2016-01-04 19:33                                                                                         ` Robert Kierski
2016-01-04 19:43                                                                                           ` Doug Dumitru
2016-01-15 16:53                                                                                             ` Robert Kierski

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=56701B4E.8060007@turmel.org \
    --to=philip@turmel.org \
    --cc=dallas.a.clement@gmail.com \
    --cc=john@stoffel.org \
    --cc=linux-raid@vger.kernel.org \
    --cc=markknecht@gmail.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 an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.