[rrd-users] generating PRINT, no graph. start time on "graph" vs start time in DEF

Alex van den Bogaerdt alex at ergens.op.het.net
Mon Mar 24 12:24:35 CET 2008


On Mon, Mar 24, 2008 at 02:02:06AM -0700, Jack Tavares wrote:
> I am calling the graph functionality, but I am not actually generating the graph, I just want to generate PRINT output
>  
> I specify the start and end times for all DEFs
>  
> I have 2 identical calls, 
>  
> one I also specify the start and end time for "the graph" and one that I do not.

The start and end of the time series you would like to display
Not: the graph.

> I get different results for the exact same time frames.
> Why would this be?

Probably not exaxt the same time frames. Per the docs (I didn't try
myself): one of your statements specifies 24 hours, the other specifies
time 1206334663 to 1206345463 = 10800 seconds

Default is the best possible resolution, which may very well differ
between those two calls.

For example: you may be looking at 2-hour intervals in the first command,
and at 1-second intervals in the second command.

Your DEF statements ask for averages.  An example using two different
RRAs, 1 PDP per step and 3 PDPs per row:

RRA1 using 1 PDP per row:  1 1 100
RRA2 using 3 PDP per row:  34 (==average(1,1,100))

Then print MAX:

MAX(1,1,100)==100
MAX(34)==34

This is most likely how you got 103.63k vs.  20.57k


Start and end time on your DEF could be used to get only a portion of
a data series.  You could for instance

--start 1206140400 --end 1206313200
DEF:x1=port1.rrd:ds0:AVERAGE:start=1206140400:end=1206226800
DEF:x2=port2.rrd:ds1:AVERAGE:start=1206226800:end=1206313200
CDEF:x=x1,UN,x2,x1


-- 
Alex van den Bogaerdt
http://www.vandenbogaerdt.nl/rrdtool/



More information about the rrd-users mailing list