[Orca-users] Re: Is ORCA/RRD the right tool?

[Waltner, Steve]

> --- In orca-users at y..., "pcorchary" <pcorchary at y...> wrote:
> > Requirement is to plot against time readings taken MANUALLY at 
> > differing times. (Approximately 2x/day - eventually there will be 
> > remote data capture via SNMP - in about 2-4 months).
> 
> Maybe I'm dense, but I'm massivly confused:
> 
> 1) For ORCA to work, do I need to define/create the RRD 'database', 
> or does ORCA do that from the log file data that I feed it?
> 
I would suggest just using RRD and not adding the Orca layer on top of RRD.
Orca is nice in that it automates the collection of numerous counters using
SEToolkit and generation of numerous graphs. Since you are just plotting one
variable, the setup and customization of Orca is overkill.

Just use RRD to setup the database and put your data points into the
database. By only taking 2 readings a day, you aren't going to get very
accurate graphs, but it's a start. RRD (and thus Orca since it just uses
RRD) interpolate you data. It assumes that if you moved 60 miles in one hour
that you moved 60 MPH for the whole hour and obviously can't tell that you
moved 180 MPH for the first 20 minutes and stopped for the last 40 minutes.
Regarding your question about interpolating data, this works fine as well.
RRD will interpolate data to figure out what the counter would have been at
the recording interval even though you didn't take your measurement at the
actual interval. Download RRD from http://www.rrdtool.org/ and go through
the examples. They discuss this behavior in detail. The nice thing about RRD
is that once you automate the data collection process, you just need to get
your data collector setup to run the same command to put your data into the
round-robin database.

Steve