[Bloat] [Codel] better testing, linux 3.6.1, cerowrt credits, other stuff

Toke Høiland-Jørgensen toke at toke.dk
Wed Oct 10 10:27:22 EDT 2012


Rick Jones <rick.jones2 at hp.com> writes:

> Mostly. The smallest step size in rrdtool is one second, so I create
> rrd's with that as the step size, but I try for sub-second samples
> from the interval results to mitigate (or try to) some of rrdtools
> averaging.

Right. I now have the runner script parsing netperf demo output and
aligning the time steps, complete with interpolation and everything (I
think; I've looked at the numbers and it seems reasonable, but I'll try
graphing the raw numbers along with the interpolated values and see if
it looks good there too). I've tried running it with a long _RR test and
then after a few seconds starting the _STREAM and _MAERTS tests. That
produces output like the attached table.

The next step will probably be adding some kind of graphing capability;
probably using matplotlib. It should also be fairly straight forward to
add parsers for different kinds of data, time series or otherwise. But
that will probably be something I'll look at as I need them, and/or when
I get around to poking around Dave's test repository.

Anyway, feedback is quite welcome. Running the sample config files in
the repository should be possible by just changing the -H parameter for
netperf.

The code is, as before, at: https://github.com/tohojo/netperf-wrapper

-Toke


-- 
Toke Høiland-Jørgensen
toke at toke.dk

-------------- next part --------------
A non-text attachment was scrubbed...
Name: netperf_timeseries.org
Type: text/x-org
Size: 1394 bytes
Desc: Netperf wrapper output
URL: <https://lists.bufferbloat.net/pipermail/bloat/attachments/20121010/b2477426/attachment-0003.bin>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 489 bytes
Desc: not available
URL: <https://lists.bufferbloat.net/pipermail/bloat/attachments/20121010/b2477426/attachment-0001.sig>


More information about the Bloat mailing list