[mrtg] Re: 1500 interfaces traffic and errors in/out How fast a machine do I need

[Bjorn Nordbo]

ChgoWinger at aol.com wrote:
> 
> I have mrtg  running in a small multiple wan link and multiple Ethernet 
> switched envronment. I monitor the traffic  and errors on about 80 interfaces 
> as well as netware server parameters. monitoring machine NT4 sp6  pent 400 64 
> mb ram mrtg-2.8.11
> 
> I would like to monitor about 40 devices containing about 1700 interfaces for 
> traffic and errors (in and out) with about 1500 interfaces.
> 
> Can I do this on one machine ?

Yes, I am monitoring about that number of targets on a dual CPU Ultra
2, and I guess it would have had a load of <10 if it was otherwise idle.
This is using MRTG 2.9pre23 with rrdtool btw, we also have ~1000 targets
on the same box using MRTG 2.8.12, and they accounts for most of the
other load.

> How fast does it have to be ?

Any old Ultra 1 could do this if you use MRTG 2.9 and rrdtool. If you
want to use the old log format (rateup) you would need something with
a lot more power.

> Can I still use NT to do it?

Hrmpf. :)

> Would it be faster if I implemented it on a platform such as Redhat
> Should I have a separate config for each device?

Our configs contains up to 500 targets each, and I find that very
convenient to work with. OTOH, if you are going to maintain these
configs by hand splitting it up in smaller files would probably be
easier to work with.

> Do I need to make the interval longer then 5 minutes ?

No need, we do this every 5 minutes. Also, as of MRTG 2.9 you can
use SNMPv2c for polling hispeed interfaces so that 5 minutes will
be sifficent even for a Gbit interface. Very convenient.

-- 
Bjørn Nordbø  -  IP Development  -  Nextra Norway

--
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