OSPF & tunnels & IP addrs used

RouterOS Forwarding Protocols
1

hello,

i recently had an issue with an L2TP tunnel as a point-to-point interface under 4.12 OSPF:
one of the OSPF sides was not sending HELLO packets. it is an effect i remember occuring for several versions.

so far i had not spent attention to the IP addresses assigned to the L2TP interfaces - i though any address was fine as long as they are unique. is this true? in the present case one L2TP address ended in .7 the other in .8
if the OSPF code internally treats P2P interfaces as /30 networks, then .7 and .8 would be in different networks, and would be broadcast and network addresses… i am asking because when i changed the IPs of the L2TP interfaces to .1 and .2 the HELLO packets started to appear.

andy

2

what was added to OSPF networks?

for example if ppp address=1.1.1.1/32 network=2.2.2.2
then in ospf networks you have to add 2.2.2.2

3

across the entire network, i allocate unique L2TP interface addresses from 192.168.100.0/24, one after the other…
OSPF has a network entry for 192.168.100.0/24, on each router. i tend to cover large ranges with a single network statement.

the missing HELLO packets have occured last yesterday, on a 4.11 router with about five IP-in-IP tunnels.
i was able to fix it by disabling most interfaces (tunnels and also ethernet ports) and also disabling the corresponding interfaces in OSPF, then re-enabling things without a reboot. it’s a RB750G 6000km away, so one is careful…
in this case it was not related to any IP address change, it occured after manually switching some interfaces to passive (to manually control which tunnels are used for traffic).
i later on upgraded to 4.13.

andy