Re: [Devmon] Devmon not populating rrd file correctly

[Richard F. <gol...@gm...>]

http://pastebin.com/WY7k37WM is what I've been using for graphing, in
case you get an RRD generated via NCV again. I'm guessing you still
have TEST2RRD with "power=devmon" instead of "power=ncv" and weren't
aware that it doesn't take effect immediately (I think one needs to at
least kill|signal the channel=status process).

Add to oids:
upsVoltageActual: .1.3.6.1.4.1.318.1.1.1.2.2.8.0        : leaf
upsReplaceDate  : .1.3.6.1.4.1.318.1.1.1.2.1.3.0        : leaf
upsCurrentOut   : .1.3.6.1.4.1.318.1.1.1.4.2.4.0        : leaf

and add at least the first one somewhere in the message.
Alternatively, remove it from the graph definition if you're not using
it.

Your proposed graph definition seems to have some sort of volt byte
units. If you're converting from volt bits to volt bytes, you should
divide by 8 rather than multiplying. :)  It is trying to use a
nonexistent "in" value, which doesn't exist as a DS in your RRD, so
the units are moot.

In my experience, there's no need to patch the message file as Buchan
suggested (because, simply stated, the NCV parser starts when it sees
a digit or "-" and stops when it sees a non-digit non-period), but I'm
not seeing the "C" he mentioned. OTOH, I'm still on 4.2.3.

Here's what I use for the NCV parser:
NCV_power="BatteryTemperature:GAUGE,RuntimeRemaining:GAUGE,BatteryCapacity:GAUGE,UPSLoad:GAUGE,Voltageactual:GAUGE,Voltagein:GAUGE,Voltageout:GAUGE,Currentout:GAUGE,Timeonbattery:GAUGE,Lastreplacedon:GAUGE"
Now that I look at it again, that Lastreplacedon value really oughtn't
be there; as implied above, for a date of "05/01/07", it's parsed as
(and results in) "5". A drawback of the NCV parser (and a raison
d'etre of SPLIT_NCV) is that an RRD can't easily have an DS pruned,
last I checked, so I can't remove it without starting over.

My TEST2RRD includes "power=ncv" and GRAPHS includes
",power:power|power1|power2," but, again, that's on 4.2.3.