OSPF neighbourship flapping between 2-WAY and EXCHANGE state

RouterOS Forwarding Protocols
1

Hi,

I am experiencing flapping of OSPF neighbours between two mentioned states. /routing ospf lsa print shows LSA’s making into ospf database but not into routing table for some reason.
I have already verified MTU on both interfaces, tried different network types and area types, device reboot, still no go.

Has anyone seen this before and found a solution?. I will provide more information when requested.
The devices are:
951-2n 6.18
1100 6.10

Any help would be appreciated.

2

Try updating ROS so both are at the same reasonably current version. Then upload the relevant configs.

Some combinations of earlier versions of ROS 6 seemed to be able to produce some odd OSPF conditions.

3

I think I forgot to mention and may be important is that the neighbourship is running over VPLS. Hence lower MTU inside VPLS network. Just to prove You may check below pings with different payloads. My both routers’ interfaces have MTU set to 1500.

ping 10.0.0.1 do-not-fragment size=1446
HOST SIZE TTL TIME STATUS
10.0.0.1 1446 64 171ms
10.0.0.1 1446 64 104ms
sent=2 received=2 packet-loss=0% min-rtt=104ms avg-rtt=137ms max-rtt=171ms

ping 10.0.0.1 do-not-fragment size=1448
HOST SIZE TTL TIME STATUS
10.0.0.1 timeout
10.0.0.1 timeout
sent=2 received=0 packet-loss=100%

ping 10.0.0.1 do-not-fragment size=1500
HOST SIZE TTL TIME STATUS
10.0.0.1 timeout
sent=1 received=0 packet-loss=100%

ping 10.0.0.1 do-not-fragment size=1540
HOST SIZE TTL TIME STATUS
packet too large and cannot be fragmented
10.0.0.2 576 64 0ms fragmentation needed and DF set
sent=1 received=0 packet-loss=100%

4

As Celtic said…there have been a number of bugs in ROS that relate to OSPF adjacency. If the issue still remains after upgrading, here are some things to verify that have to match on each neighbor for OSPF to work:

IP MTU
hello timer
dead timer
area ID
authentication type
password
stub area flag

Assuming it isn’t a bug, flapping between EXSTART and 2-WAY sounds like MTU mismatch. Here is a great article on Troubleshooting OSPF:

http://www.cisco.com/c/en/us/support/docs/ip/open-shortest-path-first-ospf/13699-29.html#req

5

Thanks for input. I’ve just realized that the connection goes via VPLS that allows only 1442B of payload which is less than in standard IP MTU 1500B set on routers’ interfaces. Seems like this may be a problem.

6

If the MTUs don’t match, you will definitely have issues. Try setting the IP MTU to match and you should get a solid adjacency