[mrtg] Strange vertical scale.

[Fred Zwarts (KVI)]

Maybe I should add that the common factor for graphs with this strange
behavior  is that they have the option kMG set to "M,G,T,P". Still I don't
understand why the result for the weekly, monthly and yearly graphs would be
different from the daily graphs. Don't these graphs use common code for the
scale along the axis?


"Fred Zwarts (KVI)"  wrote in message news:kvf5k8$6kn$1 at ger.gmane.org...

Recently, we upgraded our MRTG to version 2.17.4.
I now see that there are several graphs for which the vertical scales for
the Weekly, Monthly and Yearly graphs look strange. They do not convert to
the right postfix (k, M, G, T), but the daily graphs are OK. As an example
there is a graph which monitors the disk contents of a certain disk. There
is no change on this disk, so all graphs show the same average, maximum and
current values (10.9TB and 11.2TB for the two channels, representing the
actual use and the capacity of the disk, resp.).
For these graphs, the daily graph has a maximum scale of 12.0T, whereas the
weekly, monthly and yearly graphs show a maximum of 12000000.0M. In the
previous version all graphs showed a maximum of 12.0T. Although the scale is
not really incorrect, it looks strange, because the large number now
overlaps the graph. Is this simply a bug in this version, or is there
something in the setup that I could change to get back the behavior of the
previous version?