I came across a setup where a router with traffic coming into an interface belonging to ipv4 vrf A for example exits an interface that belongs to the global routing instance. For this destination X, the route on vrf A points to a next hop in vrf default learned via bgp.
From what i understand, the only way to leak global/default routes into a vrf is by entering a static route on the vrf with an egress interface belonging to the global routing table. But from the above scenario, vrf A learned the route via ebgp and there are no static vrf A entries for destination X.
Checking further, the next hop Y route entry in the global routing table is more specific given its mask. The next hop W for destination Y shows a static vrf A entry pointing to an egress interface belonging to the default/global routing table.
If i leak a network say 22.214.171.124/24 via a static vrf A statement with an egress interface in the global routing table say gi0/1, does that mean that all routes vrf A learns with a next hop that is within the 126.96.36.199/24 range will be considered in routing decisions?
You are abso right that the leaking works by pointing static entry but sometimes if you don't want to leak the route(some countires forces not to leak route for security), during that time you need to create a PE-CE relationship which could also work on dynamic protocol.
Thanks Shivlu for the response. I believe i figured it out; after leaking the interface belonging to the global routing table to the vrf, peering to another router via that interface was established using the vrf.