[httperf] request rate steps

Tim Wood twwood at gmail.com
Sat Aug 25 10:46:52 PDT 2007


You are correct that autobench just runs httperf multiple times
consecutively.  I would think that the delay between rates would be
small enough to be insignificant for your test (ie, much less than a
second).

On 8/25/07, rahul gundecha <rahoolgundecha at yahoo.co.in> wrote:
> thanks Tim for prompt reply.
> while going through link I found - "Autobench runs httperf a number of times
> against each host"
> Does that mean Autobench calls httperf several times with different rates?
> If this is a case then there is no difference between that and a script
> which calls httperf no. of times.
> I want that there should be any time lag (not a single second) between
> httperf calls with two different request rates. Is it possible? My aim to
> study saturation effect which will get vanish between 2 httperf calls if I
> call them in a script.
>
> thanks,
> -Rahul
>
>
> Tim Wood <twwood at gmail.com> wrote:
>  On 8/25/07, rahul gundecha wrote:
> > I am using httperf to measure webserver performance. I want to increase
> the
> > load in steps and decrease the load after reaching certain value. This
> > should be done in single experiment. As I want to see saturation effect on
> > decreased value of load as well.
>
> You might take a look at autobench
> (http://www.xenoclast.org/autobench/) it is a wrapper for
> httperf
> which lets you specify workloads which increase in steps like you
> describe.
>
> Otherwise you would probably just have to create a script which runs
> httperf several times sequentially with your different rates.
>
> -Tim
> _______________________________________________
> httperf mailing list
> httperf at linux.hpl.hp.com
> http://www.hpl.hp.com/hosted/linux/mail-archives/httperf/
>
>
>
>  ________________________________
>  DELETE button is history. Unlimited mail storage is just a click away.
>
>


More information about the httperf mailing list