[rrd-users] Re: Perl RRD graphing rounding error

[Robert Nelson]

> The problem is that for certain intervals, even though every sample
> has 100's throughout, we get a number slightly less than 100 as the
> average of B1.  I believe we have only observed it for timespans that
> span the rollup interval (for instance, from 14 days ago until today).
>  If we remove the CDEF, the result will always be 100.
> 
> The purpose of the CDEF is to convert unknown values to 0 for the
> purpose of this calculation.  It should have no effect as there are no
> unknown samples in the data we're looking at.
> 
> My question is... what am I missing?  What else can I check or look
> into?  And, is there any better way to pull the average of the
> AVERAGEs with CDEFs, rather than using the graph function and deleting
> the image?

James,

If you have 5 values of 100 and 1 value of 0, it will average 500/6.
This is why you get less than 100 for your AVERAGE. When you remove the
CDEF, it doesn't add the Nan's to the calculation, so you get 100's
only. Does that sound like what's happening?


Rob Nelson
Network Engineer
Windchannel Communications
919-538-6326 

--
Unsubscribe mailto:rrd-users-request at list.ee.ethz.ch?subject=unsubscribe
Help        mailto:rrd-users-request at list.ee.ethz.ch?subject=help
Archive     http://www.ee.ethz.ch/~slist/rrd-users
WebAdmin    http://www.ee.ethz.ch/~slist/lsg2.cgi