08-05-05 07:46 AM
"luqs" <luqmankondeth@gmail.com> writes:
> I would like to know the significance of the IP address used in the GRE
> tunnel configuration.
> I thought that tunnel source XXX & tunnel destination XXXX
> are all that is required to get a tunnel to work. However, I did use IP
> address on both ends of the tunnel and kept it in the same subnet.
>
> Now, the issue is,
>
> In one setup, I changed the IP addresses to be of different subnet at
> the two ends of the tunnel for eg, at end A I would give 10.10.10.1 &
> at end B I would give 192.168.10.1.
> It didnt work.
"It didnt work" isn't a lot to go on, but my guess is you did not have
the correct static route on either side directing the traffic to the
gateway at the opposite end of the GRE tunnel.
> However, in my live network where everything (connected & static) are
> reditributed into ospf the same setup worked!
If my theory about your static routes being incorrect then it would
work with OSPF (or any routing protocol) because once you tell it
about the interfaces you want it to run the routing protocol over, it
can dynamically learn the address(es) of peers and so populate the
routing table with the necessary routes to allow traffic to flow.
> which begs the question, what is the role of the IP address assigned to
> the tunnel?
One role is as the source IP address on any packets originating
routing traffic i.e. the OSPF daemon will (depending on the
implementation) either unicast or multicast its routing updates over
the GRE tunnel with the originating tunnel IP address as its source address.
[ Post a follow-up to this message ]
| 
