On Fri, 29 Oct 2004 12:19:46 -0500, Burton wrote in message 
<[EMAIL PROTECTED]>:
> > -----Original Message-----
> > From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] Behalf
> > Of Arnt Karlsen
> > Sent: Friday, October 29, 2004 9:56 AM
> > To: [EMAIL PROTECTED]
> > Subject: Re: [Ntop] ..ntop-3.0 on a test bridge box, and
> > /etc/init.d/ntop onFedora Core 2
> >
> >
> > On Fri, 29 Oct 2004 13:40:15 +0100, Simon wrote in message
> > <[EMAIL PROTECTED]>:
> >
> > > On Friday 29 Oct 2004 12:23 pm, Arnt Karlsen wrote:
> > > >
> > > > ..with ntop-3.0 on a test bridge box, can I use the bridge
> > > > interfaces (br0, br1, etc) or do I need to use the bridge
> > > > "elementary" interfaces (eth1, eth2, eth3, etc) ?
> > >
> > > I have it here on a bridge, on eth0.
> >
> > ..br0 according to your "scrub that" msg.  ;-)
> >
> > > I vaguely recall you can use either, but there were issues with
> > > how the data is aggregated, so you're probably better of avoiding
> > > the bridge interface. Otherwise it is hard to relate it back to
> > > where the traffic is from/to, although there could be "deeper
> > > magic" in ntop I haven't discovered.
> >
> > ..in my isp case,  I threw out br0 from the start-up command line,
> > as br0 only showed the differences between its 2 eth nics.  ;-)
> >
> > > I guess it depends what you are trying to monitor, but for me it
> > > is mostly the slow Internet traffic that is of interest, and
> > > bucketing this in with the fast ethernet traffic makes no sense.
> > > If the traffic goes through it is measured once anyway.
> > > >
> > > > ..I left RH-7.3-9 for Debian and love it, this is a one-off, a
> > > > client's box, and I see "RH setup things" _has_ changed,
> > > > so I'm back to newbie status on Red Hat's and Fedora's.  ;-)
> > >
> > > The bridge here is RH9, and yes I prefer Debian as well.
> >
> > ..I have a bridge at an isp occationally running ntop-2.2 on RH7.3,
> > the bridge is ip-less and uses cbq to throttle bandwidth.  Here I
> > just put the entire start-up command line in the "start)" section of
> > /etc/(rc.d/)init.d/ntop, so my client can go "service ntop start"
> > etc whenever he feels like ntop'ing.  Here, br0 is eth0 + eth1.

..and these /etc(/rc.d)/init.d/ntop start-up hacks are no longer
neccessary to start up ntop on a bridge, just put everything in
/etc/ntop.conf and fire first the bridge and then ntop.  ;-)

> Two issues ...
> 
> As Arnt says, the br# interface only shows the packets crossing the
> bridge.
> 
> Second is that some old versions of libpcap had issues w/ bridges.
> 
> In an ntop world you usually don't want to see filtered data - you
> want to be able to filter it yourself to meet your specific needs and
> so one of the underlying interfaces is probably more likely what you
> want to monitor.  But YMMV...

.ntop-3.0 is a nice piece of work, guys.  :-)

-- 
..med vennlig hilsen = with Kind Regards from Arnt... ;-)
...with a number of polar bear hunters in his ancestry...
  Scenarios always come in sets of three: 
  best case, worst case, and just in case.

_______________________________________________
Ntop mailing list
[EMAIL PROTECTED]
http://listgateway.unipi.it/mailman/listinfo/ntop

Reply via email to