- bgp next-hop-self, you obviously need this if the remote ibgp peer isn't reachable via an IGP. Otherwise, how is the traffic supposed to get routed to a hop that it doesn't know of.
- and when redistributing the static vrf route into BGP VPNv4 table, when you want to verify via Pings, you need to redistribute connected interface too because they are the source of your pings unless you source your pings from the loopback. Because if you don't redistribute connected interfaces, the vrf's are not aware of how to get back to the source of your pings.
Grrr... small, but vital things that slipped my mind.
Noted and on to the next lab.
