All of lore.kernel.org
 help / color / mirror / Atom feed
* --thinktime seems not work
@ 2015-05-13 14:23 Rodrigo Ribeiro Gomes
  2015-05-13 14:33 ` Jens Axboe
  0 siblings, 1 reply; 9+ messages in thread
From: Rodrigo Ribeiro Gomes @ 2015-05-13 14:23 UTC (permalink / raw)
  To: fio

HI guys!

I'm trying use the --thinktime parameter to sleep between each io
request for ten seconds.

But, it seems not working... I'm doing  something incorrect?

fio --name disktst  ... --thinktime 10000000

In HOWTO, this param is described as 'Stall the job x microseconds'...
But fio dont stop between each request and continues issuing. This is
complete command line that i used:


fio --name IoRequestDelayTest --directory J\:\  --iodepth 1 --bs 4k
--size 1G --io_size 1G --direct 1 --thread --runtime 1h --time_based
--rw randread --thinktime 10000000

The J: is mounted on a USB flash drive (Kingston SE9 32GB)

-- 

[]'s Rodrigo Ribeiro Gomes
MCITP: Database Administrator 2008
MCITP: Database Developer 2008

061 8106-2710
(Acesse o blog THE SQL TIMES | http://www.thesqltimes.com)

^ permalink raw reply	[flat|nested] 9+ messages in thread

* Re: --thinktime seems not work
  2015-05-13 14:23 --thinktime seems not work Rodrigo Ribeiro Gomes
@ 2015-05-13 14:33 ` Jens Axboe
  2015-05-13 14:37   ` Rodrigo Ribeiro Gomes
  0 siblings, 1 reply; 9+ messages in thread
From: Jens Axboe @ 2015-05-13 14:33 UTC (permalink / raw)
  To: Rodrigo Ribeiro Gomes, fio

On 05/13/2015 10:23 AM, Rodrigo Ribeiro Gomes wrote:
> HI guys!
>
> I'm trying use the --thinktime parameter to sleep between each io
> request for ten seconds.
>
> But, it seems not working... I'm doing  something incorrect?
>
> fio --name disktst  ... --thinktime 10000000
>
> In HOWTO, this param is described as 'Stall the job x microseconds'...
> But fio dont stop between each request and continues issuing. This is
> complete command line that i used:
>
>
> fio --name IoRequestDelayTest --directory J\:\  --iodepth 1 --bs 4k
> --size 1G --io_size 1G --direct 1 --thread --runtime 1h --time_based
> --rw randread --thinktime 10000000
>
> The J: is mounted on a USB flash drive (Kingston SE9 32GB)

Works fine for me:

./fio --name=think --size=1g --bs=4k --rw=randread --direct=1
[...]
   read : io=130244KB, bw=30114KB/s, iops=7528, runt=  4325msec

./fio --name=think --size=1g --bs=4k --rw=randread --direct=1 
--thinktime=1000
[...]
   read : io=17668KB, bw=3340.6KB/s, iops=835, runt=  5289msec

I tried your 10s as well, works too.

-- 
Jens Axboe



^ permalink raw reply	[flat|nested] 9+ messages in thread

* Re: --thinktime seems not work
  2015-05-13 14:33 ` Jens Axboe
@ 2015-05-13 14:37   ` Rodrigo Ribeiro Gomes
  2015-05-13 15:11     ` Jens Axboe
  0 siblings, 1 reply; 9+ messages in thread
From: Rodrigo Ribeiro Gomes @ 2015-05-13 14:37 UTC (permalink / raw)
  To: Jens Axboe; +Cc: fio

Interesting. It Windows version?

On Wed, May 13, 2015 at 11:33 AM, Jens Axboe <axboe@kernel.dk> wrote:
> On 05/13/2015 10:23 AM, Rodrigo Ribeiro Gomes wrote:
>>
>> HI guys!
>>
>> I'm trying use the --thinktime parameter to sleep between each io
>> request for ten seconds.
>>
>> But, it seems not working... I'm doing  something incorrect?
>>
>> fio --name disktst  ... --thinktime 10000000
>>
>> In HOWTO, this param is described as 'Stall the job x microseconds'...
>> But fio dont stop between each request and continues issuing. This is
>> complete command line that i used:
>>
>>
>> fio --name IoRequestDelayTest --directory J\:\  --iodepth 1 --bs 4k
>> --size 1G --io_size 1G --direct 1 --thread --runtime 1h --time_based
>> --rw randread --thinktime 10000000
>>
>> The J: is mounted on a USB flash drive (Kingston SE9 32GB)
>
>
> Works fine for me:
>
> ./fio --name=think --size=1g --bs=4k --rw=randread --direct=1
> [...]
>   read : io=130244KB, bw=30114KB/s, iops=7528, runt=  4325msec
>
> ./fio --name=think --size=1g --bs=4k --rw=randread --direct=1
> --thinktime=1000
> [...]
>   read : io=17668KB, bw=3340.6KB/s, iops=835, runt=  5289msec
>
> I tried your 10s as well, works too.
>
> --
> Jens Axboe
>



-- 

[]'s Rodrigo Ribeiro Gomes
MCITP: Database Administrator 2008
MCITP: Database Developer 2008

061 8106-2710
(Acesse o blog THE SQL TIMES | http://www.thesqltimes.com)


^ permalink raw reply	[flat|nested] 9+ messages in thread

* Re: --thinktime seems not work
  2015-05-13 14:37   ` Rodrigo Ribeiro Gomes
@ 2015-05-13 15:11     ` Jens Axboe
  2015-05-13 15:33       ` Rodrigo Ribeiro Gomes
  2015-05-13 23:01       ` Bruce Cran
  0 siblings, 2 replies; 9+ messages in thread
From: Jens Axboe @ 2015-05-13 15:11 UTC (permalink / raw)
  To: Rodrigo Ribeiro Gomes; +Cc: fio, Bruce Cran

On 05/13/2015 10:37 AM, Rodrigo Ribeiro Gomes wrote:
>
> On Wed, May 13, 2015 at 11:33 AM, Jens Axboe <axboe@kernel.dk> wrote:
>> On 05/13/2015 10:23 AM, Rodrigo Ribeiro Gomes wrote:
>>>
>>> HI guys!
>>>
>>> I'm trying use the --thinktime parameter to sleep between each io
>>> request for ten seconds.
>>>
>>> But, it seems not working... I'm doing  something incorrect?
>>>
>>> fio --name disktst  ... --thinktime 10000000
>>>
>>> In HOWTO, this param is described as 'Stall the job x microseconds'...
>>> But fio dont stop between each request and continues issuing. This is
>>> complete command line that i used:
>>>
>>>
>>> fio --name IoRequestDelayTest --directory J\:\  --iodepth 1 --bs 4k
>>> --size 1G --io_size 1G --direct 1 --thread --runtime 1h --time_based
>>> --rw randread --thinktime 10000000
>>>
>>> The J: is mounted on a USB flash drive (Kingston SE9 32GB)
>>
>>
>> Works fine for me:
>>
>> ./fio --name=think --size=1g --bs=4k --rw=randread --direct=1
>> [...]
>>    read : io=130244KB, bw=30114KB/s, iops=7528, runt=  4325msec
>>
>> ./fio --name=think --size=1g --bs=4k --rw=randread --direct=1
>> --thinktime=1000
>> [...]
>>    read : io=17668KB, bw=3340.6KB/s, iops=835, runt=  5289msec
>>
>> I tried your 10s as well, works too.
 >
 > Interesting. It Windows version?

(please don't top post!)

It's possible, the Windows part does have its own nanosleep() function, 
which fio calls for this part. That's the only OS dependent bit in this 
equation.

Bruce?

-- 
Jens Axboe



^ permalink raw reply	[flat|nested] 9+ messages in thread

* Re: --thinktime seems not work
  2015-05-13 15:11     ` Jens Axboe
@ 2015-05-13 15:33       ` Rodrigo Ribeiro Gomes
  2015-05-13 17:28         ` Jens Axboe
  2015-05-13 23:01       ` Bruce Cran
  1 sibling, 1 reply; 9+ messages in thread
From: Rodrigo Ribeiro Gomes @ 2015-05-13 15:33 UTC (permalink / raw)
  To: Jens Axboe; +Cc: fio, Bruce Cran

I did some new test and removed the parameters '--runtime' and
'--time_based' .... Without it, the thinktime works!
(sorry for top post)


^ permalink raw reply	[flat|nested] 9+ messages in thread

* Re: --thinktime seems not work
  2015-05-13 15:33       ` Rodrigo Ribeiro Gomes
@ 2015-05-13 17:28         ` Jens Axboe
  0 siblings, 0 replies; 9+ messages in thread
From: Jens Axboe @ 2015-05-13 17:28 UTC (permalink / raw)
  To: Rodrigo Ribeiro Gomes; +Cc: fio, Bruce Cran

On 05/13/2015 11:33 AM, Rodrigo Ribeiro Gomes wrote:
> I did some new test and removed the parameters '--runtime' and
> '--time_based' .... Without it, the thinktime works!
> (sorry for top post)

Neither of those should have any impact on the thinktime. Doesn't 
reproduce for me. What version of fio are you using?

-- 
Jens Axboe



^ permalink raw reply	[flat|nested] 9+ messages in thread

* Re: --thinktime seems not work
  2015-05-13 15:11     ` Jens Axboe
  2015-05-13 15:33       ` Rodrigo Ribeiro Gomes
@ 2015-05-13 23:01       ` Bruce Cran
  2015-05-14 15:33         ` Jens Axboe
  1 sibling, 1 reply; 9+ messages in thread
From: Bruce Cran @ 2015-05-13 23:01 UTC (permalink / raw)
  To: Jens Axboe, Rodrigo Ribeiro Gomes; +Cc: fio

On 5/13/2015 9:11 AM, Jens Axboe wrote:
>
> It's possible, the Windows part does have its own nanosleep() 
> function, which fio calls for this part. That's the only OS dependent 
> bit in this equation.
>
> Bruce?
>

Well, on Windows the nanosleep() function appears to be working with the 
latest fio code, so I'm not sure what might be going wrong.

-- 
Bruce


^ permalink raw reply	[flat|nested] 9+ messages in thread

* Re: --thinktime seems not work
  2015-05-13 23:01       ` Bruce Cran
@ 2015-05-14 15:33         ` Jens Axboe
  0 siblings, 0 replies; 9+ messages in thread
From: Jens Axboe @ 2015-05-14 15:33 UTC (permalink / raw)
  To: Bruce Cran, Rodrigo Ribeiro Gomes; +Cc: fio

On 05/13/2015 07:01 PM, Bruce Cran wrote:
> On 5/13/2015 9:11 AM, Jens Axboe wrote:
>>
>> It's possible, the Windows part does have its own nanosleep()
>> function, which fio calls for this part. That's the only OS dependent
>> bit in this equation.
>>
>> Bruce?
>>
>
> Well, on Windows the nanosleep() function appears to be working with the
> latest fio code, so I'm not sure what might be going wrong.

I did read the code, and it looked OK to me. So ditto, no real idea on 
what is going on here. Rodrigo, are you using a 32-bit build? Maybe the 
ints overflow. Does it work if you use smaller delays, the 10s delay per 
IO is pretty massive.

-- 
Jens Axboe



^ permalink raw reply	[flat|nested] 9+ messages in thread

* --thinktime seems not work
@ 2015-05-13 14:29 rodrigo
  0 siblings, 0 replies; 9+ messages in thread
From: rodrigo @ 2015-05-13 14:29 UTC (permalink / raw)
  To: fio


HI guys!

I'm trying use the --thinktime parameter to sleep between each io
request for ten seconds.

But, it seems not working... I'm doing  something incorrect?

fio --name disktst  ... --thinktime 10000000

In HOWTO, this param is described as 'Stall the job x microseconds'...
But fio dont stop between each request and continues issuing. This is
complete command line that i used:


fio --name IoRequestDelayTest --directory J\:\  --iodepth 1 --bs 4k
--size 1G --io_size 1G --direct 1 --thread --runtime 1h --time_based
--rw randread --thinktime 10000000

The J: is mounted on a USB flash drive (Kingston SE9 32GB)


^ permalink raw reply	[flat|nested] 9+ messages in thread

end of thread, other threads:[~2015-05-14 15:33 UTC | newest]

Thread overview: 9+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2015-05-13 14:23 --thinktime seems not work Rodrigo Ribeiro Gomes
2015-05-13 14:33 ` Jens Axboe
2015-05-13 14:37   ` Rodrigo Ribeiro Gomes
2015-05-13 15:11     ` Jens Axboe
2015-05-13 15:33       ` Rodrigo Ribeiro Gomes
2015-05-13 17:28         ` Jens Axboe
2015-05-13 23:01       ` Bruce Cran
2015-05-14 15:33         ` Jens Axboe
2015-05-13 14:29 rodrigo

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.