I've been using xnmgraph extensively. Recently, I have a need to
use the Input Data Extensibility feature. Towards that end, I have
run into a few issues. I would be grateful if anyone could offer
me a solution to any or all...
1. Rate of change. The input data I'm using for xnmgraph is actually
representing a counter (octet count), although it does not come
from a MIB var. However, I didn't find any way to tell xnmgraph
that the variable is a counter (and so, graph it as a "rate of change").
xnmgraph insists on drawing the actual measured data - which is,
obviously, not useful. I could write a script to pipe stdin through,
that will calculate the diffs etc. - but, is this really the only
2. "printOnly" option. I'm using this option extensively with the SNMP
collection (drawing graphs in the background, that eventually make
their way into web pages as gifs). For SNMP variable monitoring,
it works nicely: it will display the complete graph, then issue the
print command. When I use it with the Input Data Extensibility, though,
the graph displays only half-way through (not all points being displayed!),
and then the "printOnly" kicks in. I end up with a partial graph -
even though I have data points until 12:00 PM, the lines terminate
at 09:30 (for example).
How can I deal with this? Is there a way to tell xnmgraph to go through
all data points and *then* print? (I would hate to go into forking
and then doing SIGs - just for that...).
Thanks for any enlightment!