OK so I have 3 routers (just mucking around so not real world)
A
B
C
1. A and C have loopback of 1.1.1.1 and 2.2.2.2
2. all three routes have point to point links A <> B and B<> C
3. A and B have interface F0/0 up linking to B
4. A(.1) <> B(.2) is the sub net 192.168.1.0 /24 and B(.2) <> C(.1) is subnet 192.168.2.0/24
5 all active physical and loop back interfaces on all three routers are in area 0
6. OSPF SPF has run and all routers can see all networks via OSPF and the command ping 1.1.1.1 source 2.2.2.2 from C works
7. on router A the commands
#int tunnel 0
#ip address 10.10.10.1 255.255.255.0
#tunnel source int f0/0
#tunnel destination 192.168.2.1
#ip ospf 1 area 0
8. on router C the commands
#int tunnel 0
#ip address 10.10.10.2 255.255.255.0
#tunnel source int f0/0
#tunnel destination 192.168.1.1
#ip ospf 1 area 0
9. the tunnel comes up, I can ping both of the 10.0.0.x addresses and a trace route shows they go via the tunnel (single hop), and they should in the routing table as via the tunnel
10. a new neighbor relation ship is formed between A and C via the tunnel interface.
However the 1.1.1.1 and 2.2.2.2 address are still seen via router C address, from either A or C. But I want traffic going from A to C to use the tunnel. Or more that when I have BGP running on A <> C as iBGP there are no routing loops and the internal ospf protocol does not need to know about it.
What simple thing am i missing?
Cheers
