[rrd-users] peak after NaN?

[Nuno Gonçalves]

I have a rrdtool database created with:

>rrdtool create "$rrdfile" --step 60 DS:pressure:GAUGE:90:0:1100 DS:temperature:GAUGE:90:-50:100 DS:humidity:GAUGE:90:0:100

The samples are entered in 60s steps with:

>rrdtool update "$rrdfile" N:$P:$T:$H

After a computer restart, that made 2 missed samples, I noticed a
down-peak on the graph and did a rrdtool fetch to check:

>1239617640: 9.9586069400e+02 1.7373722021e+01 5.6504290012e+01
>1239617700: 9.9586000000e+02 1.7776053280e+01 5.7130494367e+01
>1239617760: nan nan nan
>1239617820: nan nan nan
>1239617880: 9.8159131106e+02 1.9585499172e+01 4.9668510483e+01
>1239617940: 9.9584083364e+02 2.0126658206e+01 5.0490830009e+01
>1239618000: 9.9583079547e+02 2.0435340373e+01 5.1061476218e+01

The same samples are also recorded to a text file to I could confirm
that it was not a sensor problem, the values feed to rrdtool where
between 996.1-995.8.

In >1239617880: 9.8159131106e+02 1.9585499172e+01 4.9668510483e+01,
where did come the 981.59?

Thanks,
Nuno

rrdtool 1.2.30