Flexible I/O Tester development
 help / color / mirror / Atom feed
From: Jens Axboe <jens.axboe@oracle.com>
To: Michael O'Sullivan <michael.osullivan@auckland.ac.nz>
Cc: fio@vger.kernel.org
Subject: Re: fio doesn't seem to generate IO
Date: Tue, 23 Feb 2010 10:20:30 +0100	[thread overview]
Message-ID: <20100223092029.GO1025@kernel.dk> (raw)
In-Reply-To: <4B792645.8020109@auckland.ac.nz>

On Mon, Feb 15 2010, Michael O'Sullivan wrote:
> Hi everyone,
>
> I am having a problem using iolog files to generate I/O. The attached  
> test_fio file and spc_bsu0_str3 iolog files should run together to  
> generate 6 reads and 10 writes on /mnt/xphost/asu1. Fio runs via
>
> fio test_fio
>
> and generates the following output that looks good
>
> stream0: (groupid=0, jobs=1): err= 0: pid=19187
>  read : io=24,576B, bw=18,745B/s, iops=4, runt=  1311msec
>    clat (usec): min=35, max=564, avg=125.17, stdev=214.99
>    bw (KB/s) : min=   12, max=   31, per=119.44%, avg=21.50, stdev=13.44
>  write: io=40,960B, bw=31,243B/s, iops=7, runt=  1311msec
>    clat (usec): min=25, max=72, avg=49.00, stdev=12.35
>    bw (KB/s) : min=   25, max=   25, per=83.33%, avg=25.00, stdev= 0.00
>  cpu          : usr=0.00%, sys=0.31%, ctx=27, majf=0, minf=22
>  IO depths    : 1=100.0%, 2=0.0%, 4=0.0%, 8=0.0%, 16=0.0%, 32=0.0%,  
> >=64=0.0%
>     submit    : 0=0.0%, 4=100.0%, 8=0.0%, 16=0.0%, 32=0.0%, 64=0.0%,  
> >=64=0.0%
>     complete  : 0=0.0%, 4=100.0%, 8=0.0%, 16=0.0%, 32=0.0%, 64=0.0%,  
> >=64=0.0%
>     issued r/w: total=6/10, short=0/0
>     lat (usec): 50=62.50%, 100=31.25%, 750=6.25%
>
>
> Run status group 0 (all jobs):
>   READ: io=24KB, aggrb=18KB/s, minb=18KB/s, maxb=18KB/s, mint=1311msec,  
> maxt=1311msec
>  WRITE: io=40KB, aggrb=30KB/s, minb=31KB/s, maxb=31KB/s, mint=1311msec,  
> maxt=1311msec
>
> Also, if I use inotifywatch shows the correct number of access and  
> modify (read and write) on /mnt/xphost/asu1. However, iotop shows that  
> only 8KB of reads and 8KB of writes have been performed by the IO  
> generating process...! I have run soe other tests and fio shows much  
> better performance than expected because it is only generating a small  
> fraction of the requested IO. I have been tinkering with the fio code,  
> so I'm worried I broke something and would like to know what. Can anyone  
> help?

Caching? If you uncomment the direct=1, does it do as you want?

Or is it fio not actually issuing those reads and writes?

-- 
Jens Axboe


      reply	other threads:[~2010-02-23  9:20 UTC|newest]

Thread overview: 4+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2010-02-10  9:15 Integrate SPC-1 with fio, version 2 ready for options Michael O'Sullivan
2010-02-10 10:08 ` Jens Axboe
2010-02-15 10:47 ` fio doesn't seem to generate IO Michael O'Sullivan
2010-02-23  9:20   ` 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=20100223092029.GO1025@kernel.dk \
    --to=jens.axboe@oracle.com \
    --cc=fio@vger.kernel.org \
    --cc=michael.osullivan@auckland.ac.nz \
    /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