how to understand the Q2Q in btt result?

how to understand the Q2Q in btt result?

[Ming Zhang]

from btt user guide

"Q2Q which measures the time between queue traces in the system.
This provides some idea as to how quickly IOs are being handed to
the block IO layer."

does this mean the time between queuing N request and N
+1 request?

thanks!

Ming

Re: how to understand the Q2Q in btt result?

[Alan D. Brunelle]

Ming Zhang wrote:
> from btt user guide
>
> "Q2Q which measures the time between queue traces in the system.
> This provides some idea as to how quickly IOs are being handed to
> the block IO layer."
>
> does this mean the time between queuing N request and N
> +1 request?
>   

Hi Ming -

This is the average time between the Q traces being logged. We take the 
difference between each succeeding Q trace, and divide the sum by the total.

Alan
> thanks!
>
> Ming
>
>
> -
> To unsubscribe from this list: send the line "unsubscribe linux-btrace" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html
>   

Re: how to understand the Q2Q in btt result?

[Ming Zhang]

On Sun, 2007-03-25 at 23:51 -0400, Alan D. Brunelle wrote:
> Ming Zhang wrote:
> > from btt user guide
> >
> > "Q2Q which measures the time between queue traces in the system.
> > This provides some idea as to how quickly IOs are being handed to
> > the block IO layer."
> >
> > does this mean the time between queuing N request and N
> > +1 request?
> >   
> 
> Hi Ming -
> 
> This is the average time between the Q traces being logged. We take the 
> difference between each succeeding Q trace, and divide the sum by the total.

so like block layer idle indicator. maybe a similar D2D will be a nice
device indicator as well.


> 
> Alan
> > thanks!
> >
> > Ming
> >
> >
> > -
> > To unsubscribe from this list: send the line "unsubscribe linux-btrace" in
> > the body of a message to majordomo@vger.kernel.org
> > More majordomo info at  http://vger.kernel.org/majordomo-info.html
> >   
> 
> -
> To unsubscribe from this list: send the line "unsubscribe linux-btrace" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re: how to understand the Q2Q in btt result?

[Alan D. Brunelle]

Ming Zhang wrote:
> On Sun, 2007-03-25 at 23:51 -0400, Alan D. Brunelle wrote:
>   
>> Ming Zhang wrote:
>>     
>>> from btt user guide
>>>
>>> "Q2Q which measures the time between queue traces in the system.
>>> This provides some idea as to how quickly IOs are being handed to
>>> the block IO layer."
>>>
>>> does this mean the time between queuing N request and N
>>> +1 request?
>>>   
>>>       
>> Hi Ming -
>>
>> This is the average time between the Q traces being logged. We take the 
>> difference between each succeeding Q trace, and divide the sum by the total.
>>     
>
> so like block layer idle indicator. 

Not quite - the block layer %util value indicates how much the 
underlying device (driver) has active commands. Q2Q is more of a measure 
of upper-level activity - application & file system & VM (page cache). 
It indicates how quickly commands are being submitted to the block IO layer.


> maybe a similar D2D will be a nice
> device indicator as well.
>   

This would be an indicator of how quickly lower-layer device drivers are 
getting issued IOs. It would be trivial to implement, if you are 
interested in it, I can do it...

Alan

>
>   
>> Alan
>>     
>>> thanks!
>>>
>>> Ming
>>>
>>>
>>> -
>>> To unsubscribe from this list: send the line "unsubscribe linux-btrace" in
>>> the body of a message to majordomo@vger.kernel.org
>>> More majordomo info at  http://vger.kernel.org/majordomo-info.html
>>>   
>>>       
>> -
>> To unsubscribe from this list: send the line "unsubscribe linux-btrace" in
>> the body of a message to majordomo@vger.kernel.org
>> More majordomo info at  http://vger.kernel.org/majordomo-info.html
>>     
>
> -
> To unsubscribe from this list: send the line "unsubscribe linux-btrace" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html
>   

Re: how to understand the Q2Q in btt result?

[Ming Zhang]

On Mon, 2007-03-26 at 09:58 -0400, Alan D. Brunelle wrote:
> Ming Zhang wrote:
> > On Sun, 2007-03-25 at 23:51 -0400, Alan D. Brunelle wrote:
> >   
> >> Ming Zhang wrote:
> >>     
> >>> from btt user guide
> >>>
> >>> "Q2Q which measures the time between queue traces in the system.
> >>> This provides some idea as to how quickly IOs are being handed to
> >>> the block IO layer."
> >>>
> >>> does this mean the time between queuing N request and N
> >>> +1 request?
> >>>   
> >>>       
> >> Hi Ming -
> >>
> >> This is the average time between the Q traces being logged. We take the 
> >> difference between each succeeding Q trace, and divide the sum by the total.
> >>     
> >
> > so like block layer idle indicator. 
> 
> Not quite - the block layer %util value indicates how much the 
> underlying device (driver) has active commands. Q2Q is more of a measure 
> of upper-level activity - application & file system & VM (page cache). 
> It indicates how quickly commands are being submitted to the block IO layer.
> 

sorry, after hit the send button, i realized my fault. yes, it is a
incoming system idleness indicator. like a incoming rate.

> 
> > maybe a similar D2D will be a nice
> > device indicator as well.
> >   
> 
> This would be an indicator of how quickly lower-layer device drivers are 
> getting issued IOs. It would be trivial to implement, if you are 
> interested in it, I can do it...

i think this will be interesting when compare with Q2Q to see how block
layer smooth/intense the io.


> 
> Alan
> 
> >
> >   
> >> Alan
> >>     
> >>> thanks!
> >>>
> >>> Ming
> >>>
> >>>
> >>> -
> >>> To unsubscribe from this list: send the line "unsubscribe linux-btrace" in
> >>> the body of a message to majordomo@vger.kernel.org
> >>> More majordomo info at  http://vger.kernel.org/majordomo-info.html
> >>>   
> >>>       
> >> -
> >> To unsubscribe from this list: send the line "unsubscribe linux-btrace" in
> >> the body of a message to majordomo@vger.kernel.org
> >> More majordomo info at  http://vger.kernel.org/majordomo-info.html
> >>     
> >
> > -
> > To unsubscribe from this list: send the line "unsubscribe linux-btrace" in
> > the body of a message to majordomo@vger.kernel.org
> > More majordomo info at  http://vger.kernel.org/majordomo-info.html
> >   
> 
> -
> To unsubscribe from this list: send the line "unsubscribe linux-btrace" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re: how to understand the Q2Q in btt result?

[Alan D. Brunelle]

Ming Zhang wrote:
> On Mon, 2007-03-26 at 09:58 -0400, Alan D. Brunelle wrote:
> <snip
>   
>>> maybe a similar D2D will be a nice
>>> device indicator as well.
>>>   
>>>       
>> This would be an indicator of how quickly lower-layer device drivers are 
>> getting issued IOs. It would be trivial to implement, if you are 
>> interested in it, I can do it...
>>     
>
> i think this will be interesting when compare with Q2Q to see how block
> layer smooth/intense the io.
>
>   
I will look into adding this later today.

Alan

Re: how to understand the Q2Q in btt result?

[Ming Zhang]

On Mon, 2007-03-26 at 11:09 -0400, Alan D. Brunelle wrote:
> Ming Zhang wrote:
> > On Mon, 2007-03-26 at 09:58 -0400, Alan D. Brunelle wrote:
> > <snip
> >   
> >>> maybe a similar D2D will be a nice
> >>> device indicator as well.
> >>>   
> >>>       
> >> This would be an indicator of how quickly lower-layer device drivers are 
> >> getting issued IOs. It would be trivial to implement, if you are 
> >> interested in it, I can do it...
> >>     
> >
> > i think this will be interesting when compare with Q2Q to see how block
> > layer smooth/intense the io.
> >
> >   
> I will look into adding this later today.

thanks.


> 
> Alan