Now R1/R3 have an ospf/eigrp peering we should consider EIGRP/OSPF admin distances.
Issuing a debug ip routing on either R1/R3 will show us the issue
(on R3)
RT: closer admin distance for 150.1.9.9, flushing 1 routes
RT: add 150.1.9.9/32 via 155.1.13.1, ospf metric [110/20]
shortly followed by:
RT: del 150.1.9.9 via 155.1.13.1, ospf metric [110/20]
RT: delete subnet route to 150.1.9.9/32
RT: updating eigrp 150.1.9.9/32 (0x0) :
via 155.1.13.1 Gi1.13 0 1048578
We see this in the RIB, continuously:
R3#sh ip route | in 150.1.9.9
O E2 150.1.9.9 [110/20] via 155.1.13.1, 00:00:04, GigabitEthernet1.13
R3#sh ip route | in 150.1.9.9
D EX 150.1.9.9 [170/131072] via 155.1.37.7, 00:00:01, GigabitEthernet1.37
Solution:
Change the AD of OSPF external routes in the OSPF process (I did this on R3) (to be higher than the DEX routes).
Off to dinner :)