All of lore.kernel.org
 help / color / mirror / Atom feed
From: Robert Balogh <ethrbh@gmail.com>
To: Ben England <bengland@redhat.com>
Cc: Damien Le Moal <damien.lemoal@opensource.wdc.com>,
	fio <fio@vger.kernel.org>
Subject: Re: FIO performance measurement between volumes
Date: Mon, 17 Jan 2022 08:48:50 +0100	[thread overview]
Message-ID: <CAA4OwwVv+364MrZ8=1zUrWf2-+uechnkmdLVj_aa176H3pA2sw@mail.gmail.com> (raw)
In-Reply-To: <CAM5EYbmExSVAtEKzYWE+quTQk-5J4iSnnW4+qoMLg0+MF5sbJw@mail.gmail.com>

hello Ben,

Thanks for your reply. I will try out how dd works for me.

thanks again,
/Robi

On Fri, Jan 14, 2022 at 2:59 PM Ben England <bengland@redhat.com> wrote:
>
> dd can do this.   Beware of caching effects, you can drop cache + use dd conv=fsync  to ensure the data makes it to the target volume.
>
> On Fri, Jan 14, 2022 at 2:48 AM Robert Balogh <ethrbh@gmail.com> wrote:
>>
>> hello Damien,
>>
>> Thanks for your quick reply. I am afraid I made a mistake to not
>> describe clearly what my goal is. So I can perform the FIO test for
>> the 2nd attached volume similar to what is done for the 1st volume.
>> But now I would like to do a performance measurement between the
>> volumes.
>>
>> So my idea is, there is a e.g: 10G size file stored in volume1, and
>> this file is going to write to volume2, then read back. But now, I am
>> not sure this can be done with the FIO tool.
>>
>> Thanks for your help.
>> /Robi
>>
>> On Fri, Jan 14, 2022 at 12:20 AM Damien Le Moal
>> <damien.lemoal@opensource.wdc.com> wrote:
>> >
>> > On 1/13/22 20:47, Robert Balogh wrote:
>> > > hello FIO experts,
>> > >
>> > > I am a beginner in FIO, and I met with a problem.
>> > >
>> > > Currently in my Ubuntu based server the "/dev/vdb1" volume is attached
>> > > to the "/home/batman/fio/cindervolume" path. FIO job is configured
>> > > like this
>> > >
>> > >     [global]
>> > >     filename=fio-172_20_2_13
>> > >     directory=/home/batman/fio/cindervolume rw=rw
>> > >     rwmixread=60
>> > >     rwmixwrite=40
>> > >     bs=4k
>> > >     rate=500k
>> > >     direct=1
>> > >     numjobs=1
>> > >     time_based=1
>> > >     runtime=14d
>> > >     verify=crc32c
>> > >     continue_on_error=all
>> > >     group_reporting=1
>> > >
>> > >     [file1]
>> > >     iodepth=1
>> > >     ; -- end job file --
>> > >
>> > > FIO is started like this and works well:
>> > >     /usr/bin/fio --size=10G
>> > > --output=/home/batman/fio/cindervolume/fio-172_20_2_13-process.log
>> > > /home/batman/fioApp/fio-seq-RW.job &
>> > >
>> > > FIO version I use: fio-3.25
>> > >
>> > > My next step would be to attach a 2nd volume for example "/dev/vdc" to
>> > > "/home/batman/fio/cindervolume-2" path, and I would do performance
>> > > measurement between the volumes. I was checking the FIO user’s guide,
>> > > https://fio.readthedocs.io/en/latest/fio_doc.html but unfortunately I
>> > > cannot figure out which parameter might help me to solve this topic.
>> > >
>> > > By-the-way is this possible to do with FIO? If so, could you please
>> > > help me by giving some direction/hints?
>> >
>> > Move the directory=/... option from global section into the job section
>> > and add a job for your other volume with directory= again added in that
>> > job section to point to the new volume directory. The 2 jobs will run
>> > simultaneously targetting different volumes, unless you add "stonewall",
>> > in which case the jobs will run one after the other.
>> >
>> > Or you could just write another fio script for the second volume.
>> >
>> > It is not clear if you want to run the perf tests on the 2 volumes in
>> > parallel or one after the other.
>> >
>> > >
>> > > thanks for your help,
>> > >
>> > > /Robi
>> >
>> >
>> > --
>> > Damien Le Moal
>> > Western Digital Research
>>

      parent reply	other threads:[~2022-01-17  7:49 UTC|newest]

Thread overview: 4+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2022-01-13 11:47 FIO performance measurement between volumes Robert Balogh
2022-01-13 23:20 ` Damien Le Moal
2022-01-14  7:47   ` Robert Balogh
     [not found]     ` <CAM5EYbmExSVAtEKzYWE+quTQk-5J4iSnnW4+qoMLg0+MF5sbJw@mail.gmail.com>
2022-01-17  7:48       ` Robert Balogh [this message]

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='CAA4OwwVv+364MrZ8=1zUrWf2-+uechnkmdLVj_aa176H3pA2sw@mail.gmail.com' \
    --to=ethrbh@gmail.com \
    --cc=bengland@redhat.com \
    --cc=damien.lemoal@opensource.wdc.com \
    --cc=fio@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 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.