log_offset parameter not logging offset

log_offset parameter not logging offset

[abhishek koundal]

All,
Can someone please confirm if my usage is correct as earlier the
offset used to be logged by using the log_offset=1 in the files.
I am using FIO3.5.

fio --name=global --time_based --size=100% --direct=1 --norandommap
--randrepeat=0 --buffered=0 --refill_buffers --name=job
--ioengine=libaio --group_reporting --filename=/dev/nvme0n1
--numjobs=1 --iodepth=128 --bs=128k --rw=rw --rwmixread=70
--runtime=10800 --write_iops_log=wr.log -write_lat_log=wr.clat.log
--log_avg_msec=1000 --log_offset=1

appreciate the support.

Life is too short for silly things so invest your time in some
productive outputs!!

Re: log_offset parameter not logging offset

[Sitsofe Wheeler]

Hi,

On 9 March 2018 at 20:11, abhishek koundal <akoundal@gmail.com> wrote:
> All,
> Can someone please confirm if my usage is correct as earlier the
> offset used to be logged by using the log_offset=1 in the files.
> I am using FIO3.5.
>
> fio --name=global --time_based --size=100% --direct=1 --norandommap
> --randrepeat=0 --buffered=0 --refill_buffers --name=job
> --ioengine=libaio --group_reporting --filename=/dev/nvme0n1
> --numjobs=1 --iodepth=128 --bs=128k --rw=rw --rwmixread=70
> --runtime=10800 --write_iops_log=wr.log -write_lat_log=wr.clat.log
> --log_avg_msec=1000 --log_offset=1

To help us out a bit can you cut back as many options as possible?
Ideally you should try to get to the point where you have the minimum
number of options that still show the problem (i.e. to the point where
you can't remove any one option and still have the problem happen)?
For example I bet you can remove "--direct=1 --norandommap" and still
have the problem you describe happen. It just makes it easier to
investigate when there are less places to look...

-- 
Sitsofe | http://sucs.org/~sits/

Re: log_offset parameter not logging offset

[abhishek koundal]

fio --name=global --time_based --size=100% --filename=/dev/nvme0n1
--runtime=10 --write_iops_log=wr.log --log_avg_msec=1000
--log_offset=1

Still seeing the issue.

On Fri, Mar 9, 2018 at 1:52 PM, Sitsofe Wheeler <sitsofe@gmail.com> wrote:
> Hi,
>
> On 9 March 2018 at 20:11, abhishek koundal <akoundal@gmail.com> wrote:
>> All,
>> Can someone please confirm if my usage is correct as earlier the
>> offset used to be logged by using the log_offset=1 in the files.
>> I am using FIO3.5.
>>
>> fio --name=global --time_based --size=100% --direct=1 --norandommap
>> --randrepeat=0 --buffered=0 --refill_buffers --name=job
>> --ioengine=libaio --group_reporting --filename=/dev/nvme0n1
>> --numjobs=1 --iodepth=128 --bs=128k --rw=rw --rwmixread=70
>> --runtime=10800 --write_iops_log=wr.log -write_lat_log=wr.clat.log
>> --log_avg_msec=1000 --log_offset=1
>
> To help us out a bit can you cut back as many options as possible?
> Ideally you should try to get to the point where you have the minimum
> number of options that still show the problem (i.e. to the point where
> you can't remove any one option and still have the problem happen)?
> For example I bet you can remove "--direct=1 --norandommap" and still
> have the problem you describe happen. It just makes it easier to
> investigate when there are less places to look...
>
> --
> Sitsofe | http://sucs.org/~sits/



-- 
Life is too short for silly things so invest your time in some
productive outputs!!

Re: log_offset parameter not logging offset

[Sitsofe Wheeler]

On 9 March 2018 at 22:59, abhishek koundal <akoundal@gmail.com> wrote:
>
> On 9 March 2018 at 20:11, abhishek koundal <akoundal@gmail.com> wrote:
>> All,
>> Can someone please confirm if my usage is correct as earlier the
>> offset used to be logged by using the log_offset=1 in the files.
>> I am using FIO3.5.
>
> fio --name=global --time_based --size=100% --filename=/dev/nvme0n1
> --runtime=10 --write_iops_log=wr.log --log_avg_msec=1000
> --log_offset=1
>
> Still seeing the issue.

log_avg_msec=1000 tries to average values seen every second which
means some values will become nonsense when it's enabled. There's a
warning at the end of the "Log File Formats" section of the HOWTO
(http://fio.readthedocs.io/en/latest/fio_doc.html#log-file-formats )
but perhaps it needs to be moved elsewhere - would it have been better
under the log_avg_msec section of the HOWTO
(http://fio.readthedocs.io/en/latest/fio_doc.html#cmdoption-arg-log-avg-msec
)?

-- 
Sitsofe | http://sucs.org/~sits/

Re: log_offset parameter not logging offset

[Sitsofe Wheeler]

Hi,

Glad to hear you found the cause. Nonetheless, do you think the
documentation needs changing and if so how?

On 10 March 2018 at 08:07, abhishek koundal <akoundal@gmail.com> wrote:
> yes you are correct. thanks for the information, i can see the offset
> w/o the log_avg_msec.
> appreciate your prompt support.
>
> On Fri, Mar 9, 2018 at 9:48 PM, Sitsofe Wheeler <sitsofe@gmail.com> wrote:
>> On 9 March 2018 at 22:59, abhishek koundal <akoundal@gmail.com> wrote:
>>>
>>> On 9 March 2018 at 20:11, abhishek koundal <akoundal@gmail.com> wrote:
>>>> All,
>>>> Can someone please confirm if my usage is correct as earlier the
>>>> offset used to be logged by using the log_offset=1 in the files.
>>>> I am using FIO3.5.
>>>
>>> fio --name=global --time_based --size=100% --filename=/dev/nvme0n1
>>> --runtime=10 --write_iops_log=wr.log --log_avg_msec=1000
>>> --log_offset=1
>>>
>>> Still seeing the issue.
>>
>> log_avg_msec=1000 tries to average values seen every second which
>> means some values will become nonsense when it's enabled. There's a
>> warning at the end of the "Log File Formats" section of the HOWTO
>> (http://fio.readthedocs.io/en/latest/fio_doc.html#log-file-formats )
>> but perhaps it needs to be moved elsewhere - would it have been better
>> under the log_avg_msec section of the HOWTO
>> (http://fio.readthedocs.io/en/latest/fio_doc.html#cmdoption-arg-log-avg-msec
>> )?

-- 
Sitsofe | http://sucs.org/~sits/

Re: log_offset parameter not logging offset

[abhishek koundal]

Where  log_offset description is given It will be best to have one
liner saying if log_avg_msec parameter is used offset wont be logged.
This will make it very clear.

On Sat, Mar 10, 2018 at 1:18 AM, Sitsofe Wheeler <sitsofe@gmail.com> wrote:
> Hi,
>
> Glad to hear you found the cause. Nonetheless, do you think the
> documentation needs changing and if so how?
>
> On 10 March 2018 at 08:07, abhishek koundal <akoundal@gmail.com> wrote:
>> yes you are correct. thanks for the information, i can see the offset
>> w/o the log_avg_msec.
>> appreciate your prompt support.
>>
>> On Fri, Mar 9, 2018 at 9:48 PM, Sitsofe Wheeler <sitsofe@gmail.com> wrote:
>>> On 9 March 2018 at 22:59, abhishek koundal <akoundal@gmail.com> wrote:
>>>>
>>>> On 9 March 2018 at 20:11, abhishek koundal <akoundal@gmail.com> wrote:
>>>>> All,
>>>>> Can someone please confirm if my usage is correct as earlier the
>>>>> offset used to be logged by using the log_offset=1 in the files.
>>>>> I am using FIO3.5.
>>>>
>>>> fio --name=global --time_based --size=100% --filename=/dev/nvme0n1
>>>> --runtime=10 --write_iops_log=wr.log --log_avg_msec=1000
>>>> --log_offset=1
>>>>
>>>> Still seeing the issue.
>>>
>>> log_avg_msec=1000 tries to average values seen every second which
>>> means some values will become nonsense when it's enabled. There's a
>>> warning at the end of the "Log File Formats" section of the HOWTO
>>> (http://fio.readthedocs.io/en/latest/fio_doc.html#log-file-formats )
>>> but perhaps it needs to be moved elsewhere - would it have been better
>>> under the log_avg_msec section of the HOWTO
>>> (http://fio.readthedocs.io/en/latest/fio_doc.html#cmdoption-arg-log-avg-msec
>>> )?
>
> --
> Sitsofe | http://sucs.org/~sits/



-- 
Life is too short for silly things so invest your time in some
productive outputs!!