[mrtg] MRTG/RRD Performance tests: preliminary results

[Eric Brander]

On Sun, Sep 12, 2010 at 8:00 PM, Steve Shipway <s.shipway at auckland.ac.nz> wrote:
> MRTG Daemon mode
>
> Running MRTG in daemon mode allows it to keep file descriptors open, cache
> the configuration file contents, and so on.  The performance gain seems to
> be consistently in the region of 20%, mo matter which version of RRDTool you
> are using.
>

Steve, et al,

I avoid using daemon mode because I lose a lot of visibility into how
MRTG is running. I run a cron job every minute that searches for
config files and calls MRTG against each one individually. I keep a
log for each poll as well as a total log for each time the script is
run. I also keep track of the seconds it take to run the entire
script, save that to a text file, and run MRTG against that to build a
graph showing how long polling took. It's a bit cumbersome but I know
at a glance how long my 1-minute polling jobs are taking to run. An
added side effect is that one broken cfg file doesn't break polling
for everything.  I also have a separate (but functionally identical)
script for 5-minute polling.

Maybe you can suggest another way to do this while taking advantage of
the daemon mode performance increase? Since in daemon mode MRTG never
"ends" there doesn't seem to be a way for me to graph how long a cycle
takes.

BTW, thanks for doing this research Steve. Very informative!

Regards,

Eric N. Brander