[mrtg] Re: sum of OID's

[McDonald, Dan]

 >From: kurt van Buynder
> Hello all,   Since a while I monitoring traffic on two different links.
>These two graphs are absolutely great! Now I wanted to graph the sum of
>thesetwo links by doing the following   33:public at IP +
>34:public at same_IP.
>Theresult is that I'm getting a graph that consist the sum of these
>links
>butI have a lot of spikes in this graph (not in the other ones) for a
>period
>sometimes de one hour.   Can anyone help me?   Thx, Kurt

Unfortunately, that is a side-effect of roll-overs that cannot be accounted
for.  With just one counter rolling over, you can figure out how much it
changed.  But with two counters that might roll over at any time, there is
little way to determine what the correct values are - particularly if one
rolled over in time period "T" and the other rolled in time period "T+1".
It's even worse if they both roll over in the same period.  Since only the
aggregate is known, mrtg can't figure out if it should roll over once or
twice - or even at all.

A solution might be to use RRDTOOL as the backend, log the data separately,
and  then create a stacked graph from the two rrd's as a custom graph.  

Or use 64 bit counters so that rollovers occur very seldom.

--
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