ok - I am REALLY myphed by this concept. According to the lab docs I am working with the Cost attribute of a route redistiributed from BGP into EIGRP keeps the router from accepting EIGRP routes from a back door?
For the life of me I could not get R6 to use the route via sw1 for network 172.16.5.0/24 (network path r5-sw2-sw1) that was all native to EIGRP 3. I changed the redistribution paramters on R6 bgp such that bandwith 1k and delay 100 to prefer the route via Sw2.
With the connection between sw1 and sw2 active as the back door, r6 still chose the bgp path through r4.
use this scenario -
R6- pe
R4 - p (net 155.1.67.0 /24 is in vpn_a)
R5 - pe (net 155.1.58.0/24 and 172.16.5.0/24 are in vpn_a)
Sw1 - ce
Sw2 ce
the topology is:
r6 - r4 - r5 - sw2 - sw1 - r6
r6 - r4 - vpn peers
r4 - r5 - vpn peers
r4 configured as route reflector for r6 and r5
r6 has vpn_a configured facing sw1 - they are eigrp neighbors, r6 is configured for ipv4 vpn_a - switch is unaware of vrf
r5 have vpn_a configured facing sw2 - they are eigrp neighbors, r5 is configured for ipv4 vpn_a - switch is unaware of vrf
sw1 and sw2 are eigrp neighbors
r6 is mutually redistributing eigrp vpn_a and bgp ipv4 vpn_a
r5 is mutually redistributing eigrp vpn_a into bgp ipv4 vpn_a
Rack1R5#srr (this configuration is mirrored on R6 except that vpn_a is 155.1.67.0)
router eigrp 3
no auto-summary
!
address-family ipv4 vrf vpn_a
redistribute bgp 654 metric 10000 100 255 1 1500
network 155.1.58.5 0.0.0.0
network 172.16.5.0 0.0.0.255
no auto-summary
autonomous-system 3
exit-address-family
router bgp 654
no bgp default ipv4-unicast
bgp log-neighbor-changes
neighbor 150.1.4.4 remote-as 654
neighbor 150.1.4.4 update-source Loopback0
!
address-family vpnv4
neighbor 150.1.4.4 activate
neighbor 150.1.4.4 send-community extended
exit-address-family
!
address-family ipv4 vrf vpn_a
redistribute eigrp 3
no synchronization
network 1.1.1.5 mask 255.255.255.255
exit-address-family
!
Sw1 is clearly chosing the path via sw2:
Rack1SW1#sh ip route 155.1.58.0
Routing entry for 155.1.58.0/24
Known via "eigrp 3", distance 90, metric 28416, type internal
Redistributing via eigrp 3
Last update from 155.1.12.8 on FastEthernet0/13, 00:02:20 ago
Routing Descriptor Blocks:
* 155.1.12.8, from 155.1.12.8, 00:02:20 ago, via FastEthernet0/13
Route metric is 28416, traffic share count is 1
Total delay is 110 microseconds, minimum bandwidth is 100000 Kbit
Reliability 255/255, minimum MTU 1504 bytes
Loading 1/255, Hops 1
R6 is still prefering the path via bgp:
Rack1R6#sho ip route vr vpn_a 155.1.58.0
Routing entry for 155.1.58.0/24
Known via "bgp 654", distance 200, metric 0, type internal
Redistributing via eigrp 3
Advertised by bgp 654 (self originated)
eigrp 3 metric 1 2000 1 1 1
Last update from 150.1.5.5 00:28:55 ago
Routing Descriptor Blocks:
* 150.1.5.5 (Default-IP-Routing-Table), from 150.1.4.4, 00:28:55 ago
Route metric is 0, traffic share count is 1
AS Hops 0
It isn't until I shut down the backdoor, removing the bgp path that r6 actually picks up the EIGRP route via Sw1.
Rack1R6(config-subif)#do sho ip route vr vpn_a 155.1.58.0
Routing entry for 155.1.58.0/24
Known via "eigrp 3", distance 90, metric 30976, type internal
Redistributing via bgp 654, eigrp 3
Advertised by bgp 654
Last update from 155.1.67.7 on FastEthernet0/0.67, 00:00:09 ago
Routing Descriptor Blocks:
* 155.1.67.7, from 155.1.67.7, 00:00:09 ago, via FastEthernet0/0.67
Route metric is 30976, traffic share count is 1
Total delay is 210 microseconds, minimum bandwidth is 100000 Kbit
Reliability 255/255, minimum MTU 1500 bytes
Loading 1/255, Hops 2
IS THIS THE RESULT OF THE COST ATTRIBUTE?
I DON'T GET IT. WHY WOULD R6 PERFER A BGP ROUTE WITH AN AD OF 200 OVER AN EIGRP ROUTE OF AD 90?