[mrtg] missing data from one datasource

[Daniel J McDonald]

On Wed, 2007-07-04 at 09:53 +0100, Neil Lathwood wrote:
> Hi Folks,
> 
> Currently running mrtg 2.15.2 on a CentOS 5 box using rrd as a
> backend. Everything works fine for the most, data is being collected
> fine and graphs generated fine. However every so often the graphs will
> stop generating from one of the datasources, if I enable logging for
> the rrd updates I get the following: 
> 
> 2007-07-04 09:25:24 -- --log: RRDs::tune(/home/mrtg/switch1_16.rrd -a
> ds0:1250000 -a ds1:1250000 -d ds0:COUNTER -d ds1:COUNTER)

If I am counting the zeros properly, this is gigabit, but you are using
the 32bit snmp counters instead of the 64 bit counters.  So, every time
you go above ~130 MB/s mrtg will give you bizarre results, since it will
roll over twice in 5 minutes.

Switch to snmp V2 (or v3, if you prefer to have the polling encrypted)
and it should work fine.


-- 
Daniel J McDonald, CCIE # 2495, CISSP # 78281, CNX
Austin Energy
http://www.austinenergy.com