never mind :)

I just noticed a message when switching NIC's ...

It said when you have the NetFlow or sFlow plugins
enabled, they force the -M

l8r!

--Blake


--- Blake <[EMAIL PROTECTED]> wrote:
> 
> Hello all!
> 
> When you have 2 NIC's ...
> 
> eth0 collecting data from a switch port
> eth1 collecting NetFlow from a router
> 
> Im thinking the -M option SHOULD be set so NTOP does
> not merge together eth0 and eth1 (or NetFlow
> device)?
> 
> Would I be correct in saying this, or would the
> nature
> of NTOP already be not to merge data between a
> regular
> interface and a logical NetFlow interface?
> 
> Thanks,
> 
> Blake
> 
> --- "Burton M. Strauss III" <[EMAIL PROTECTED]>
> wrote:
> > 
> > Assuming you only care about netFlow, you are
> > correct.  You only need to
> > monitor (real) NICs that you want to collect data
> > from.  Whether you use
> > that data only to display ntop's web pages or to
> > forward to a netFlow
> > receiver is irrelevant.  Cigarillo...
> > 
> > "why is the NetFlow-device interface created when
> > you turn on the plugin?"
> > 
> > Basically, there is a large, dynamic data
> structure
> > in ntop that holds all
> > the information about a device, collected,
> sniffed,
> > etc. from the packets.
> > ntop creates one if it's merging data or multiples
> > if it's not merging data,
> > one per "device".  So, ntop creates another one of
> > these structures to hold
> > the information it receives via netFlow packets
> when
> > it's acting as a
> > netFlow receiver.  That way, all the reporting
> logic
> > just works... it
> > doesn't know the difference.
> > 
> > The structure has to be created before the first
> > netFlow packet is received,
> > i.e. when we KNOW -- based on the persistent
> > settings for the plugin, or
> > your setting the port # -- that we'll be dealing
> > with inbound netFlow data.
> > 
> > 
> > -----Burton
> > 
> > 
> > 
> > -----Original Message-----
> > From: Blake [mailto:[EMAIL PROTECTED]]
> > Sent: Friday, September 27, 2002 2:08 PM
> > To: Burton M. Strauss III; [EMAIL PROTECTED]
> > Subject: RE: [Ntop] NetFlow Overview
> > 
> > 
> > Thanks for the reply (even though you are on
> > vacation)
> > 
> > SO here is my current config, and it appears as
> > though
> > I am receiving NetFlow packets from the router. 
> So,
> > I
> > should not enable any of the following devices
> > unless
> > I am exporting NetFlow data from that device to a
> > remote NetFlow collector ... cigar?
> > 
> > In order to receive NetFlow from a router, I just
> > need
> > to enable the plugin and specify the UDP port.
> > 
> > By the way, why is the NetFlow-device interface
> > created when you turn on the plugin?  Of course I
> > saw
> > NOTE: #4 A virtual NetFlow device is activated
> only
> > when incoming flow capture is enabled --- but I
> dont
> > understand its purpose.
> > 
> > ##############
> > CURRENT CONFIG
> > ##############
> > 
> > Interface Name NetFlow Enabled
> > eth0 No
> > eth1 No
> > NetFlow-device No
> > 
> > 
> > WARNING: as all the interfaces are disabled, no
> > flows
> > will be exported
> > 
> > Flow Statistics
> > # Pkts Rcvd.value 124
> > # Flows Rcvd.value 3,720
> > # Flow with Bad Version 0
> > Flow Senders 192.168.2.1 [124 pkts]
> > 
> > 
> > 
> > 
> > --- "Burton M. Strauss III"
> <[EMAIL PROTECTED]>
> > wrote:
> > > Pick one:
> > >
> > >   Wrongo dragon breath...
> > >   Close grasshopper, but no cigar...
> > >
> > >
> > > Separate the two activities of netFlow in your
> > mind.
> > >  THEY DO NOT OVERLAP!
> > >
> > > When ntop is acting as a receiver of netFlow
> data,
> > > it receives the packets
> > > on the specified port and stores the data in
> it's
> > > internal structures just
> > > like data collected off another network card. 
> > Hence
> > > you switch "NICs" to
> > > report on the netFlow data.
> > >
> > > When ntop is acting as a collector of netFlow
> > data,
> > > it collects information
> > > from it's network cards and sends that off to
> some
> > > netFlow receiver.  You
> > > can monitor the data ntop has received just like
> > > normal, but you can't
> > > monitor the data ntop has sent via netFlow to
> > > another device, you have to
> > > use that device to monitor it...
> > >
> > > You started ntop ... -i eth0 ...  that's why you
> > > only have the two devices,
> > > eth0 and netFlow.  Because that's all you've
> told
> > > ntop to monitor...  how
> > > about ... -i eth0,eth1 ... ???
> > >
> > > If you are only using ntop to receive netFlow
> > data,
> > > you may have a problem.
> > > IIRC, it won't run without monitoring at least
> one
> > > (real) NIC.  I might be
> > > wrong...  I suppose you could always monitor 
> the
> > > local loopback (-i lo).
> > > Once it's up?  As I've said before, the data
> > > collected by netFlow somewhere
> > > else and sent to ntop is presented via the
> netFlow
> > > pseudo-nic.
> > >
> > > You're right it won't have session and other
> > > detailed data - that's not in
> > > the flows (look at the header files, you'll see
> > > what's being "recorded").
> > >
> > >
> > > -----Burton
> > >
> > >
> > > -----Original Message-----
> > > From: [EMAIL PROTECTED]
> > > [mailto:[EMAIL PROTECTED]]On Behalf Of Blake
> > > Sent: Friday, September 27, 2002 12:09 PM
> > > To: [EMAIL PROTECTED]
> > > Subject: [Ntop] NetFlow Overview
> > >
> > >
> > > Hopefully someone can assist my ignorance in
> > regards
> > > to using NTOP as a NetFlow collector.  If not
> ...
> 
=== message truncated ===


__________________________________________________
Do you Yahoo!?
Faith Hill - Exclusive Performances, Videos & More
http://faith.yahoo.com
_______________________________________________
Ntop mailing list
[EMAIL PROTECTED]
http://listgateway.unipi.it/mailman/listinfo/ntop

Reply via email to