[c-nsp] Sanity check OSPF/BGP

[Eugene Grosbein]

08.10.2020 20:00, Drew Weaver wrote:

> I have two sets of core routers due to a transition period from one set to the other.
> 
> I have noticed that when there is a connectivity disruption between the two sets of core routers and one upstream peering/edge router:
> 
> Oct  7 12:01:14 EDT: %OSPF-5-ADJCHG: Process 1, Nbr <removed> on TenGigabitEthernet2/1 from FULL to DOWN, Neighbor Down: BFD node down
> 
> <Two+ minutes of null routing traffic for no reason>
> 
> Oct  7 12:03:29 EDT: %BGP-5-ADJCHANGE: neighbor <removed> Down BGP Notification sent
> 
> What I expect to happen is:
> 
>               The route to the peering edge router's loopback interface is withdrawn when OSPF/OSPFv3 closes.
>               The core router will close the BGP session when the route to the dead peering edge router is withdrawn and will begin using one of the 5 other copies of the same route that it has.
> 
> Things I have implemented to avoid this:
> 
>               The peering edge router and the core routers peer with IP addresses that are only learnable via OSPF and aren't available in any other protocol. [It's not part of our IP space]
> 
> I guess I just need a sanity check regarding whether my assumption that it shouldn't be null routing traffic for 2+ minutes if one of our peering edge routers gets hit by a meteor is correct since we have 5 peering edge routers.

This may depend on BGP synchronization that could be disabled by default. Did you enable it?