[rrd-users] Re: Newbie help need
serge.maandag at staff.zeelandnet.nl
Wed Nov 21 10:11:39 MET 2001
>>As for the averages: They determine the resolution of your data. rrd's
are made for storing data that has been measured in fairly constant
>Is that another way of saying the averages control how many dots you
Well, at least it determines the maximum No of dots you could print.
If you have 480 datasamples in your rrd and you're graphing it with
>How do you decide what xff to use? What is it?
It's a tricky one. The manual says:
xff -> The xfiles factor defines what part of a consolidation interval
may be made up from *UNKNOWN* data while the consolidated value is still
regarded as known.
That's quite an X-files explanation, but if I have understood it
correctly, it works like this:
You're inputting data every 180 seconds. That also is the step size for
An xff of 0.5 says that if you input two consecutive points with an
interval of less than 90 secs (=180-0.5x180) or more than 270 secs
(=180+0.5x180), they will be regarded as invalid and a "value unknown"
will be inserted in your rrd.
If your interval is within valid range, the data will be extrapolated as
if it was measured at the 180 secs interval.
> People talk about rows...is the data in a row column format in the
rrd? With my data, how would the table look?
Create the RRD and do a "rrdtool dump", open it in your browser and
you'll see. A row is a placeholder for a measured or averaged value. In
your case the rrd holds 480 rows for the day timespan.
>> You're base interval is 1800 secs, so that's the rrd stepsize.
>How do I use this 1800 stepsize ... this is different than the --step
parameter? >Do I need a --step 1800?
Err, typo. That's 180, since that is the time between every measured
value. I forgot to mention in the example you had to add --step 180 to
the rrdtool create command.
>>Based on all this you should try:
>I tried this and ran the graph command listed below. I get a graph
that has week 47 - week 50 on the horizontal and 100 through 300 on the
vertical. Why did it talk about weeks when I put in a start of 0 hr
today and end of time I ran the command?
Hmm, your start and end time will have to correlate with the timestamps
of your data. Try graphing the first day by setting start to your first
timestamp (1008780618 below) and set end to that value + 86400 secs (= 1
Unsubscribe mailto:rrd-users-request at list.ee.ethz.ch?subject=unsubscribe
Help mailto:rrd-users-request at list.ee.ethz.ch?subject=help
More information about the rrd-users