On Mon, Feb 10, 2003 at 12:39:51PM -0800, John Giordano wrote:
> 
> no beginning and no end.  within is contained the knowledge we seek.

true, provided you also have a clue about the pointer to "now".

> what I am not sure on is what I have heard on how RRD "compresses" the data 
> over time.  What does this mean?

It doesn't compress data.  It does however do the following:

inputs:  A, B, C, D, E, F, G, H, I, J, K, L ...
archives: p,q,r ...

in archive p, each individual input is stored
in archive q, the average of two inputs is stored
in archive r, the average of four inputs is stored


A --> p
B --> p  >and<  avg(A,B) --> q
C --> p
D --> p  >and<  avg(C,D) --> q  >and<  avg(A,B,C,D) --> r
E --> p
F --> p  >and<  avg(E,F) --> q
G --> p
H --> p  >and<  avg(G,H) --> q  >and<  avg(E,F,G,H) --> r
...and so on...


> I understand that the .RRD logs don't grow and this is one of the many things 
> that are cool about RRD and the fact it only generates the graphs when need 
> be.

If you need to graph one day per pixel (say for a yearly graph), it is
nice to have the data available >>>in that resolution<<<.  You don't
want to spend time on computing the average of (86400/300) samples
for each pixel at graph time.  RRDtool will do this averaging (when
instructed!) at storage.

Also, if you never need to graph the data of last year's January in
a high resolution image, why keep that data ?

What this means is that it usually is a good thing to have, for instance:

about 600 lines of 300 seconds per bucket
about 600 lines of 30 minutes per bucket
about 600 lines of 2 hours per bucket
about 600 lines of 1 day per bucket

This doesn't allow you to fetch data in a 300-second-per-bucket resolution
from more than (300*600) seconds ago.  It does allow you to get the
same average in a lower resolution.

If you want to know more, there's detailed documentation available.

-- 
Much of what looks like rudeness in hacker circles is not intended to give
offence. Rather, it's the product of the direct, cut-through-the-bullshit
communications style that is natural to people who are more concerned about
solving problems than making others feel warm and fuzzy.

http://www.tuxedo.org/~esr/faqs/smart-questions.html

--
Unsubscribe mailto:[EMAIL PROTECTED]
Help        mailto:[EMAIL PROTECTED]
Archive     http://www.ee.ethz.ch/~slist/rrd-users
WebAdmin    http://www.ee.ethz.ch/~slist/lsg2.cgi

Reply via email to