Flexible I/O Tester development
 help / color / mirror / Atom feed
From: Jens Axboe <jens.axboe@oracle.com>
To: Bill Broadley <bill@broadley.org>
Cc: fio@vger.kernel.org
Subject: Re: --blocksize bug?
Date: Tue, 13 Apr 2010 10:40:19 +0200	[thread overview]
Message-ID: <20100413084019.GY27497@kernel.dk> (raw)
In-Reply-To: <4BC42842.7000304@broadley.org>

On Tue, Apr 13 2010, Bill Broadley wrote:
>
> I downloaded 1.38, and then the  
> http://brick.kernel.dk/snaps/fio-git-latest.tar.gz.  Both exhibited  
> identical behavior.
>
> Both built with no problems on a centos-5.4 x86-64 box.
>
> REPORTING-BUGS mentioned cat .git/HEAD, so I'll report it:
> ref: refs/heads/master
>
> Goal:
>   To benchmark randomly accessing a 1GB file with different blocksizes
>   to see how blocksize effects bandwidth.
>
> rand.fio:
> [random-read]
> rw=randread
> size=128m # just enough to get the numbers
> filesize=1G # make sure the seek distance is randomly over 1GB file
> directory=/mnt/tmp
> blocksize=1024k
>
> BTW, the fio.1 page says:
>   blocksize=int[,int], bs=int[,int]
>      Block size for I/O units.  Default: 4k.  Values  for  reads  and
>      writes  can  be  specified  seperately in the format read,write,
>      either of which may be empty to leave that value at its default.
>
> BTW, that should be separately not "seperately"

Thanks, will fix that.

> So since I wanted to vary the blocksize I figured I'd do it on the  
> command line so I said:
> ./fio --showcmd ~/rand.fio
> fio --name=random-read --rw=randread --size=128m --filesize=1G  
> --directory=/mnt/tmp --blocksize=1024k
>
> To I tried to run it:
> ./fio --name=random-read --rw=randread --size=128m --filesize=1G  
> --directory=/mnt/tmp --blocksize=1024k
> ./fio: unrecognized option `--blocksize=1024k'
>
> I tried to use the config file except for the blocksize.  I deleted the  
> blocksize= line from rand.fio and:
> ./fio --blocksize=1m  ~/rand.fio
> ./fio: unrecognized option `--blocksize=1m'
>
> Potential fixes:
> 1) fix showcmd to report a different flag for setting --blocksize
> 2) fix fio to accept --blocksize
>
> Expected behavior:
> I expected ./fio ~/rand.fio to work exactly like the arguments reported  
> by ./fio --showcmd.

And it should, apparently the alias names are not duped to the command
line options. I have fixed this up now, if you download the latest
snapshot it should work.

http://brick.kernel.dk/snaps/fio-git-latest.tar.gz

-- 
Jens Axboe


      reply	other threads:[~2010-04-13  8:40 UTC|newest]

Thread overview: 2+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2010-04-13  8:16 --blocksize bug? Bill Broadley
2010-04-13  8:40 ` Jens Axboe [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=20100413084019.GY27497@kernel.dk \
    --to=jens.axboe@oracle.com \
    --cc=bill@broadley.org \
    --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 a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox