RIP or OSPF over OpenVPN

[tguldener]

Hi

I have a routing issue on my installation. I have at the moment 3 sites with OpenVPN TAP Interfaces:

(site1) tap10.1.0.2 <-> tap10.1.0.1 (site2) <-> tap10.1.0.5 <-> tap10.1.0.6 (site3)

I tried to enable OpenOSPF or RIP over this installation. At the moment rip is running. The routing table looks perfect on all 3 sites. I see the tap interfaces an the connected networks. But the routing is only working for the tap interfaces:

Site1:
10.1.0.0/30 link#13 U 3010 1500 ovpnc1
10.1.0.2 link#13 UHS 0 16384 lo0
10.1.0.4/30 10.1.0.1 UG 3 1500 ovpnc1
10.2.1.0/24 link#3 U 266 1500 re2
10.3.1.0/24 10.1.0.1 UG 0 1500 ovpnc1
10.4.1.0/24 10.1.0.1 UG 0 1500 ovpnc1

Site2:
10.1.0.0/30 link#9 U 0 1500 ovpns1
10.1.0.1 link#9 UHS 0 16384 lo0
10.1.0.4/30 link#10 U 20 1500 ovpns2
10.1.0.5 link#10 UHS 0 16384 lo0
10.2.1.0/24 10.1.0.2 UG 0 1500 ovpns1
10.3.1.0/24 link#2 U 67 1500 re1
10.4.1.0/24 10.1.0.6 UG 0 1500 ovpns2

Site3:
10.1.0.0/30 10.1.0.5 UG 0 1500 ovpnc1
10.1.0.4/30 link#13 U 30 1500 ovpnc1
10.1.0.6 link#13 UHS 0 16384 lo0
10.2.1.0/24 10.1.0.5 UG 0 1500 ovpnc1
10.3.1.0/24 10.1.0.5 UG 0 1500 ovpnc1
10.4.1.0/24 link#3 U 3753 1500 re2

the transit interfaces are reachable true all site (10.1.0.0/30, 10.1.0.4/30). The routing entries for this addresses will be directly added from the OpenVPN Server/Client.

ping source 10.2.0.2 to 10.1.0.6 is working
ping source 10.2.0.2 to 10.4.1.1 is not working
ping source 10.2.0.2 to 10.1.0.1 is working
ping source 10.2.0.2 to 10.3.1.1 is not working

All the other networks will be still routed over the default gateway. It looks for me like a bug. The routes are more specific and should have a higher priority then the default gateway.

Any idea to fix the problem?

g.
thomas