[mrtg] Re: Weird Cisco issue

[Keith]

On Wed, 25 Aug 1999, Tim Kennedy wrote:

> First off, I am assuming that the individual graphs are
> correct.  Or are they having problems?  If those are really high speed
> interfaces, maybe the counters are rolling on some of them, and mrtg
> is adjusting for that when you graph the individual line, but I don't
> know if, when you are adding interfaces together, MRTG is smart enough to
> make the distinction, or if it just throws a zero in there.

I was told there was an IOS bug that caused IOS to return weird data if
there are too many SNMP polls in a small amount of time. This is word of
mouth; I have heard nothing from Cisco.

The individual graphs corrolate to the other side of the interfaces, so
they are correct.

They are not pushing _that_ much traffic -- a lot less then above.net does
and they are using MRTG (though, I have a feeling it's heavily modified by
Dave Rand for AboveNet, but maybe not.)

> I don't poll ifInOctets and ifOutOctets, though.  I use, instead, cisco's
> own locIfInterfaces stats.  ifInBps, and ifOutBps.  Set the bits option.
> And away I go.  You also don't have to worry about the counter rolling,
> since it's already giving you a bps number.  you could try that.

Hmm. What would a sample interface look like? This sounds like a good
workaround.

Thanks for the advice,

--
Keith McCallion
keith at mccallion.com

--
* To unsubscribe from the mrtg mailing list, send a message with the
  subject: unsubscribe to mrtg-request at list.ee.ethz.ch
* The mailing list archive is at http://www.ee.ethz.ch/~slist/mrtg