[rrd-users] question about LSLSLOPE and LSLINT

[Marco Frattola]

Hi to everybody,
I'm trying to use rrdtool (1.3) to graph some trends (mainly for disks, 
but also cpu and ram).
I've followed what others have done (i.e. copied every example that I 
found) with LSLSLOPE and LSINT, and I thought I got it right.
Graphs seemed to prove I got it right. Thanks to all the people who 
provided examples.

Then this weird thing started happening: as I added more data sources, 
some predicted dates (reaching 75% o full capacity) looked very odd 
(meanining: the date -when reaching the limit- would not correspond to 
the graph).

I started changing the end date of the graph and noticed that the base 
parameter for the prediction (LSLSLOPE and LSLINT) changed too, which is 
something I did not expect to happen. I thought that since those values 
are based on real data (i.e. collected data), they would not change if I 
extended the graph time in the future (since they are based on data in 
the past, and they don't change).
What's wrong? What am I missing?

Thank you very much.

--