[mrtg] Re: 14all.cgi and performance

[Chris Green]

Here I'm monitoring 4000+ targets with mrtg/rrdtool and the mrtg
process takes ~30-40 seconds with Forks: 4.  The processing time has
gone down even more that I'm running mrtg in daemon mode.

Machine is a dual 733 pIII w/ 1 gig of ram and scsi drives.

A couple 14allv1.1 releases ago, I added a quick and dirty cache to
the config parsing and Joe Gooch <mrwizard at psu.edu> worked on
generating mrtg graphs at run the first cgi call rather than waiting
and parsing the config file for each individual stat.

This makes one huge config file work pretty well.  There is a slight
pause on huge target directories but nothing major.

Logarithmic graphs are also a runtime option.

One configuration change is there is a 14all*htmlbasedir for the
basedirecotory to your pngs.  The log and config directories also will
have to be writable by the cgi process for both the images and the
cache.  

http://helium.tucc.uab.edu/~sprout/14all_cmg/ has the code

Rainer.Bawidamann at rz.uni-ulm.de (Rainer Bawidamann) writes:
> 
> 14all v1.0 (for mrtg 2.8) contains code to cache the config. I didn't
> get a response from sites with big mrtg installations (they will use
> mrtg 2.9) so I don't know how good this is. I plan to port this to 14all
> v1.1/mrtg 2.9 but this is a major change to mrtg (MRTG_lib) so I habe to
> talk with Tobi about this and I don't have the time for it now.
> 
> Chris Green did some work on performance for 14all, maybe he listens and
> wants to provide his work to the list...

-- 
Chris Green <cmg at uab.edu>
A good pun is its own reword.

--
Unsubscribe mailto:mrtg-request at list.ee.ethz.ch?subject=unsubscribe
Archive     http://www.ee.ethz.ch/~slist/mrtg
FAQ         http://faq.mrtg.org    Homepage     http://www.mrtg.org
WebAdmin    http://www.ee.ethz.ch/~slist/lsg2.cgi