[rrd-users] creating a rrd database with specific RRAs

[Simon Hobson]

grewskee wrote:

>I apologize, it's actually a 5 minute resolution I am looking for.  It
>sounds like storing data for exactly 3 days would not be the best
>thing to do for the reasons that you described.  Probably keeping a
>weeks worth of data, and then just simply graphing for the last 3 days
>would be the best bet I assume.
>
>So, essentially what I want is:
>
>* Last hour (5 minute average)
>* Last day (5 minute average)
>* Last week (30 minute average)
>
>Again, I'd like to have the ability to determine the MIN/MAX values as
>well.  Can you assist me with the formula in this particular case?

Again, it's fairly simple.

You probably want something like this :
RRA:AVERAGE:0.5:1:290
RRA:AVERAGE:0.5:6:336
RRA:MINIMUM:0.5:6:336
RRA:MAXIMUM:0.5:6:336

That's 288 samples * 5 minutes = 1 day. If you are precise with your 
graph start & end times then 288 will work. 5 minutes means 12 
samples/hour, times 24 hours, gives 288. That will do you for your 
hour and day graphs.

For your longer graphs, 6 PDP/CDP gives 30 minute bins. 2 bins/hour, 
times 24 hours, times 7 days, gives 336.

And of course, if you want to be bale to graph Max and Min in 
addition to average, then you need to store it when consolidating.

To recap. You need to say now many primary data point are in each 
consolidated data point. Each consolidated data point will cover step 
time times PDP per CDP periods of time - eg 300s times 6 is 1800s.
Then you need to say how many of these consolidated points to keep.

-- 
Simon Hobson

Visit http://www.magpiesnestpublishing.co.uk/ for books by acclaimed
author Gladys Hobson. Novels - poetry - short stories - ideal as
Christmas stocking fillers. Some available as e-books.