perfsonar development work

[Loukik Kudarimoti <>]

Roman Lapacz wrote:
> Loukik,
>
> long time ago (when all pS developers were working on the framework 
> and rrd ma) you implemented methods for writing data to rrd files 
> (they are still the same). How did you test them to see if they work 
> well? I could test those methods the same way to see if there is a 
> problem.
>
> Roman
I remember testing it. I created an rrd file using specific parameters 
(heartbeat, interval, etc). I can't remember the details but these 
parameters were very important as it defines at what rate you can put in 
the data. Then, I sent test data at an interval used to construct the 
rrd file (otherwise, rrd file aggregates values into the same row). I 
kept varying the test data value that I asked rrd file to store. I 
remember that the results were okay (nothing different from the command 
line interface)

The store interface is a tricky one because the rate at which you send 
data is very much dependent on the rate that you configure RRD files 
with. If the value is sent late, the RRD file ignores the interval. If 
the value is sent early, it gets aggregated into the previous row than 
the one expected. The stored value depends on what you configure the 
file as (COUNTER, ABSOLUTE, etc)...The MA of course has no clue and no 
grip on this. Further, if you have multiple data sources in the same rrd 
file, all of them need to be updated at the same time. Otherwise you 
have NaNs for the datasources (or columns) for which no data was provided.

Loukik.

--

---------------------------------------------------------------
L o u k i k   K u d a r i m o t i

     * *              Network Engineer
   *     *            City House, 126 - 130, Hills Road
  *                   Cambridge CB2 1PQ, United Kingdom
  *                   WWW: http://www.dante.net
D  A  N  T  E          Tel:+44 1223 371300 Fax:+44 1223 371371