On 30/12/2017 04:26, Fakity Fakfak wrote: > [...] > The two raspberry pi is connected using an ethernet cable through each of > the Pi's eth0 port. Other devices are connected using usb-to-eth cable.
So you have two "host" (RPi-A and RPi-B), each one with:
- one physical network interface (eth0)
- _three_ USB-network-adapters (providing eth1, eth2, eth3);
> My network infrastructure is similar to a Pi Stack Switch as seen on this
> github post. ( https://github.com/OpenWinCon/pi-switch ) but only using two
> (2) raspberry pi
So you have an UTP ethernet cable directly connecting both RPis (let's
assume, eth0<=>eth0).
On 29/12/2017 14:29, Fakity Fakfak via discuss wrote:
> [...]
> I wanted the topology to work in such a way that Host A can ping
> every host in the network (B,C,D,E,F)
So, basically you want to build a sort of "distributed switch" with 8
ethernet ports (eth0, eth1, eth2, eth3 from both RPiA and RPiB) with two
ports (eth0, eth0) used by your platform and the other 6 ports available
for external equipments.
On 29/12/2017 14:29, Fakity Fakfak via discuss wrote:
> [...]
> I want to use one pox controller for my network topology.
So you also want to use a single openflow controller (pox) to control all
the infrastructure.
If the above is correct (is it correct? If not, please, reply with
details....), the solution seems quite simple :-)
1 => install the openvswitch switch component to _BOTH_ RPis (so to have
the OVS components ready to run, in both RPis);
2 => configure a bridge on _BOTH_ RPIs (so to have a virtual-switch (a
bridge) running on both RPis);
3 => add eth0, eth1, eth2, eth3 to the above bridges (so to have related
physical ports ready to be managed by OVS);
This should be enough to have your infrastructure look like a single 8-port
ethernet switch (providing that you directly connect the two RPis each
other with one direct network connection).
In order to have "pox" controlling _BOTH_ your OVS-instance (RPIs), you
simply have to configure the controller (for BOTH RPis).
You can also "push" things a bit more, by _REMOVING_ the direct cable
between RPis and replacing it with a common Layer-3 link. In other words,
you can simulate having two _DISTINCT_ OVS instances, interconnected VIA
THE INTERNET (and not via a direct link).
In such a case, once given proper IP addresses to at least one interface
for each RPis (so that they can reach each other. Aka: they can PING each
other and get a reply), you can simply remove the eth0 interface and
configure an additional "tunnel" interface (type=GRE or type=VXLAN) in each
of the two bridges (for each RPis) and... everything will work again :-)
With tunnels (GRE as well as VXLAN) you might encounter issues with MTU...
but... I'm going much beyond your original question, so let's stop here :-)
Hope this help!
Enjoy!
Bye,
DV
--
Damiano Verzulli
e-mail: dami...@verzulli.it
---
possible?ok:while(!possible){open_mindedness++}
---
"Technical people tend to fall into two categories: Specialists
and Generalists. The Specialist learns more and more about a
narrower and narrower field, until he eventually, in the limit,
knows everything about nothing. The Generalist learns less and
less about a wider and wider field, until eventually he knows
nothing about everything." - William Stucke - AfrISPA
http://elists.isoc.org/mailman/private/pubsoft/2007-December/001935.html
signature.asc
Description: OpenPGP digital signature
_______________________________________________ discuss mailing list disc...@openvswitch.org https://mail.openvswitch.org/mailman/listinfo/ovs-discuss
