Dear pmacct users,

I'm trying to connect pmacct on a linux system to a machine running 
netflowanalyser v7.

My linux system configuration is a PC with 3 gigabit ports, one for 
administrive use: the other two are bridged. 

I plan to implement some bandwidth control on the egress interface of 
this bridge, but for now i'm just finding a suitable monitoring package.
Netflow analyser does everything I need.


To get enough and the correct data to this host using pmacct is proving 
difficult.
I have a single /24 IP network and wish to report on upstream and 
downstream traffic and link utilisation from this.

I imagine it will be a simple configuration file but all that I've tried 
have not been successful.

Do I make two sprobe instances, say [in] and [out] each filtered by src 
and dest network?
or do I attach an incoming filter to each of the bridged interfaces and 
sprobe for each interface?
what do I use as an aggregate if any? I'm guessing [src|dst]_host so I 
can have per host resolution at the monitor.

any tips on how to configure this would be great.

Additionally, when netflowanalyser receives some data it doesn't show 
correct interface names.  When it polls back using SNMP they are 
incorrect also.  How is the interface represneted in sflow? by localname 
or local index?

thanks for any insight

Best regards
Rob









_______________________________________________
pmacct-discussion mailing list
http://www.pmacct.net/#mailinglists

Reply via email to