All of lore.kernel.org
 help / color / mirror / Atom feed
* FIO r/w to 2 disks
@ 2012-06-08 19:52 Strickland, Carl
  2012-06-14  7:54 ` Jens Axboe
  0 siblings, 1 reply; 5+ messages in thread
From: Strickland, Carl @ 2012-06-08 19:52 UTC (permalink / raw)
  To: fio@vger.kernel.org

Hi All,

I'm performing a series of time-based storage tests and need FIO to read and
write to separate disks, in the ratio specified (e.g. 80/20).

I've performed initial tests with two FIO instances (one read, one write), but
found that in one scenario I get a ratio of ~50/50 and in another I get
~60/40. However, in order to compare the results of these tests, their ratios
must be approximately equal.

Is it possible to achieve this with a single FIO instance? Do you have any
recommendations for achieving a specific ratio with two FIO instances?

All help is greatly appreciated!


Thanks,
Carl

^ permalink raw reply	[flat|nested] 5+ messages in thread

* Re: FIO r/w to 2 disks
  2012-06-08 19:52 FIO r/w to 2 disks Strickland, Carl
@ 2012-06-14  7:54 ` Jens Axboe
  2012-06-14 23:41   ` Strickland, Carl
  0 siblings, 1 reply; 5+ messages in thread
From: Jens Axboe @ 2012-06-14  7:54 UTC (permalink / raw)
  To: Strickland, Carl; +Cc: fio@vger.kernel.org

On 2012-06-08 21:52, Strickland, Carl wrote:
> Hi All,
> 
> I'm performing a series of time-based storage tests and need FIO to read and
> write to separate disks, in the ratio specified (e.g. 80/20).
> 
> I've performed initial tests with two FIO instances (one read, one write), but
> found that in one scenario I get a ratio of ~50/50 and in another I get
> ~60/40. However, in order to compare the results of these tests, their ratios
> must be approximately equal.
> 
> Is it possible to achieve this with a single FIO instance? Do you have any
> recommendations for achieving a specific ratio with two FIO instances?

It's not clear from your email exactly what kind of job(s) you ran. Have
you tried a single job with rwmix set? rwmixread=80 and rw=randrw should
give you the requested ratio, 80% reads and 20% writes.

-- 
Jens Axboe


^ permalink raw reply	[flat|nested] 5+ messages in thread

* RE: FIO r/w to 2 disks
  2012-06-14  7:54 ` Jens Axboe
@ 2012-06-14 23:41   ` Strickland, Carl
  2012-06-15  6:49     ` Jens Axboe
  0 siblings, 1 reply; 5+ messages in thread
From: Strickland, Carl @ 2012-06-14 23:41 UTC (permalink / raw)
  To: Jens Axboe; +Cc: fio@vger.kernel.org

Hi Jens,

The test I'm interested in performing would require (I believe) FIO to read
from one device and write to another in a specified ratio. This is different
from the rwmix feature in FIO, as I can specify only a single device where the
reads and writes will take place.

Assuming that I could achieve the same read/write ratio with two FIO instances
(one reading from one device and the other writing to a different device) as I
get using a single FIO instance and rwmix, is there a reason--in terms of how
reads and writes would be performed--that would prevent a valid comparison
between the two?


Thanks,
Carl


^ permalink raw reply	[flat|nested] 5+ messages in thread

* Re: FIO r/w to 2 disks
  2012-06-14 23:41   ` Strickland, Carl
@ 2012-06-15  6:49     ` Jens Axboe
  2012-06-18 15:48       ` Strickland, Carl
  0 siblings, 1 reply; 5+ messages in thread
From: Jens Axboe @ 2012-06-15  6:49 UTC (permalink / raw)
  To: Strickland, Carl; +Cc: fio@vger.kernel.org

On 06/15/2012 01:41 AM, Strickland, Carl wrote:
> Hi Jens,
> 
> The test I'm interested in performing would require (I believe) FIO to
> read from one device and write to another in a specified ratio. This
> is different from the rwmix feature in FIO, as I can specify only a
> single device where the reads and writes will take place.
> 
> Assuming that I could achieve the same read/write ratio with two FIO
> instances (one reading from one device and the other writing to a
> different device) as I get using a single FIO instance and rwmix, is
> there a reason--in terms of how reads and writes would be
> performed--that would prevent a valid comparison between the two?

Maybe the flow options would help you? Set up two jobs in a single file,
one reader and one writer. Set up the flow parameters to get your
desired split. See HOWTO, and/or examples/flow to get an idea for how to
use that.

-- 
Jens Axboe


^ permalink raw reply	[flat|nested] 5+ messages in thread

* RE: FIO r/w to 2 disks
  2012-06-15  6:49     ` Jens Axboe
@ 2012-06-18 15:48       ` Strickland, Carl
  0 siblings, 0 replies; 5+ messages in thread
From: Strickland, Carl @ 2012-06-18 15:48 UTC (permalink / raw)
  To: Jens Axboe; +Cc: fio@vger.kernel.org

Thanks Jens, I will look into it!


Cheers,
Carl Strickland

-----Original Message-----
From: fio-owner@vger.kernel.org [mailto:fio-owner@vger.kernel.org] On Behalf Of Jens Axboe
Sent: Thursday, June 14, 2012 11:50 PM
To: Strickland, Carl
Cc: fio@vger.kernel.org
Subject: Re: FIO r/w to 2 disks

On 06/15/2012 01:41 AM, Strickland, Carl wrote:
> Hi Jens,
> 
> The test I'm interested in performing would require (I believe) FIO to 
> read from one device and write to another in a specified ratio. This 
> is different from the rwmix feature in FIO, as I can specify only a 
> single device where the reads and writes will take place.
> 
> Assuming that I could achieve the same read/write ratio with two FIO 
> instances (one reading from one device and the other writing to a 
> different device) as I get using a single FIO instance and rwmix, is 
> there a reason--in terms of how reads and writes would be 
> performed--that would prevent a valid comparison between the two?

Maybe the flow options would help you? Set up two jobs in a single file, one reader and one writer. Set up the flow parameters to get your desired split. See HOWTO, and/or examples/flow to get an idea for how to use that.

--
Jens Axboe

--
To unsubscribe from this list: send the line "unsubscribe fio" in the body of a message to majordomo@vger.kernel.org More majordomo info at  http://vger.kernel.org/majordomo-info.html


^ permalink raw reply	[flat|nested] 5+ messages in thread

end of thread, other threads:[~2012-06-18 15:48 UTC | newest]

Thread overview: 5+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2012-06-08 19:52 FIO r/w to 2 disks Strickland, Carl
2012-06-14  7:54 ` Jens Axboe
2012-06-14 23:41   ` Strickland, Carl
2012-06-15  6:49     ` Jens Axboe
2012-06-18 15:48       ` Strickland, Carl

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.