All of lore.kernel.org
 help / color / mirror / Atom feed
* Help: FIO testing problem on arm64
@ 2015-01-29  1:36 majun (F)
  2015-01-30  8:48 ` Philip Keogh
  2015-01-31  1:25 ` Kulkarni, Vasu
  0 siblings, 2 replies; 4+ messages in thread
From: majun (F) @ 2015-01-29  1:36 UTC (permalink / raw)
  To: fio@vger.kernel.org

Hi All:
	I met a problem while testing the ramdisk peformance using fio.
	The platform is arm64 , fio version is 2.1.10.
	
	The ramdisk is enabled on my platform , so there are dev/ram0 ...to dev/ram15 totall 16 ramdisk devices.
	I tried to test these ramdisks performance uisng the script file list as below.

#!/bin/bash

#disk_list=`lsscsi | awk '{print $NF}' | cut -d/ -f3`
disk_list="ram1 ram2 ram3 ram4 ram5 ram6 ram7 ram8 ram9 ram10 ram11 ram12 ram13 ram14 ram15"
numjobs=1
runtime=180
	
run_18disk_test()
{
logdir=/root/18disk_performance_test
rm -rf $logdir
mkdir $logdir
for rw in read write randread randwrite
	do
		for bs in 4K 1M
		do
			for iodepth in 1 2 4 8 16 32
			do
				for disk in `echo $disk_list`
				do
					iodepth_2=`printf "%02d" $iodepth`
					fio --time_based --filename=/dev/$disk --rw=$rw --bs=$bs --ioengine=libaio --direct=1 --numjobs=$numjobs --iodepth=$iodepth --name=test --group_reporting --runtime=$runtime &> $logdir/${disk}_${bs}_${rw}_numjobs${numjobs}_iodepth${iodepth_2}.log &
				done
				sleep $runtime
			done
		done
done
}

	

	But the test result is not ok.
	The error log is:

test: (g=0): rw=write, bs=4K-4K/4K-4K/4K-4K, ioengine=libaio, iodepth=1
fio-2.1.14
Starting 1 process
fio: cache invalidation of /dev/ram2 failed: Device or resource busy
fio: cache invalidation of /dev/ram2 failed: Device or resource busy
fio: cache invalidation of /dev/ram2 failed: Device or resource busy
fio: cache invalidation of /dev/ram2 failed: Device or resource busy
fio: cache invalidation of /dev/ram2 failed: Device or resource busy
fio: cache invalidation of /dev/ram2 failed: Device or resource busy
fio: cache invalidation of /dev/ram2 failed: Device or resource busy
fio: cache invalidation of /dev/ram2 failed: Device or resource busy

	



^ permalink raw reply	[flat|nested] 4+ messages in thread

* Re: Help: FIO testing problem on arm64
  2015-01-29  1:36 Help: FIO testing problem on arm64 majun (F)
@ 2015-01-30  8:48 ` Philip Keogh
  2015-02-02  0:51   ` majun (F)
  2015-01-31  1:25 ` Kulkarni, Vasu
  1 sibling, 1 reply; 4+ messages in thread
From: Philip Keogh @ 2015-01-30  8:48 UTC (permalink / raw)
  To: majun (F); +Cc: fio@vger.kernel.org

On Wed, Jan 28, 2015 at 5:36 PM, majun (F) <majun258@huawei.com> wrote:
> Hi All:
>         I met a problem while testing the ramdisk peformance using fio.
>         The platform is arm64 , fio version is 2.1.10.
>
>         The ramdisk is enabled on my platform , so there are dev/ram0 ...to dev/ram15 totall 16 ramdisk devices.
>         I tried to test these ramdisks performance uisng the script file list as below.
>
> #!/bin/bash
>...
>                                         fio --time_based ... --runtime=$runtime ... &
>                                 done
>                                 sleep $runtime
>                         done
>...
>         But the test result is not ok.
>         The error log is:
>...
>

Have you tried sleeping for slightly longer then the runtime passed to fio?

As a result of implementation details and the simple fact that sleep
isn't exact, nor is fio's runtime, a script not allowing a sufficient
buffer between backgrounded fio operations but intending to run
operations sequentially may step on itself.

^ permalink raw reply	[flat|nested] 4+ messages in thread

* Re: Help: FIO testing problem on arm64
  2015-01-29  1:36 Help: FIO testing problem on arm64 majun (F)
  2015-01-30  8:48 ` Philip Keogh
@ 2015-01-31  1:25 ` Kulkarni, Vasu
  1 sibling, 0 replies; 4+ messages in thread
From: Kulkarni, Vasu @ 2015-01-31  1:25 UTC (permalink / raw)
  To: majun (F); +Cc: fio@vger.kernel.org

Since the intention of the script is to run serially, you dont need
both '&' and 'sleep',  you can remove '&' and let the fio command
block and remove sleep

>                                         fio --time_based --filename=/dev/$disk --rw=$rw --bs=$bs --ioengine=libaio --direct=1 --numjobs=$numjobs --iodepth=$iodepth --name=test --group_reporting --runtime=$runtime &> $logdir/${disk}_${bs}_${rw}_numjobs${numjobs}_iodepth${iodepth_2}.log
>                                 done
>                                 #sleep $runtime

^ permalink raw reply	[flat|nested] 4+ messages in thread

* Re: Help: FIO testing problem on arm64
  2015-01-30  8:48 ` Philip Keogh
@ 2015-02-02  0:51   ` majun (F)
  0 siblings, 0 replies; 4+ messages in thread
From: majun (F) @ 2015-02-02  0:51 UTC (permalink / raw)
  To: Philip Keogh; +Cc: fio@vger.kernel.org

Hi Philip:
	Thanks for you replying.
Can you explain the case of  "may step on itself" more detail ?


在 2015/1/30 16:48, Philip Keogh 写道:
> On Wed, Jan 28, 2015 at 5:36 PM, majun (F) <majun258@huawei.com> wrote:
>> Hi All:
>>         I met a problem while testing the ramdisk peformance using fio.
>>         The platform is arm64 , fio version is 2.1.10.
>>
>>         The ramdisk is enabled on my platform , so there are dev/ram0 ...to dev/ram15 totall 16 ramdisk devices.
>>         I tried to test these ramdisks performance uisng the script file list as below.
>>
>> #!/bin/bash
>> ...
>>                                         fio --time_based ... --runtime=$runtime ... &
>>                                 done
>>                                 sleep $runtime
>>                         done
>> ...
>>         But the test result is not ok.
>>         The error log is:
>> ...
>>
> 
> Have you tried sleeping for slightly longer then the runtime passed to fio?
> 
> As a result of implementation details and the simple fact that sleep
> isn't exact, nor is fio's runtime, a script not allowing a sufficient
> buffer between backgrounded fio operations but intending to run
> operations sequentially may step on itself.
> --
> 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] 4+ messages in thread

end of thread, other threads:[~2015-02-02  0:54 UTC | newest]

Thread overview: 4+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2015-01-29  1:36 Help: FIO testing problem on arm64 majun (F)
2015-01-30  8:48 ` Philip Keogh
2015-02-02  0:51   ` majun (F)
2015-01-31  1:25 ` Kulkarni, Vasu

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.