cancel
Showing results for 
Search instead for 
Did you mean: 
cancel
542
Views
0
Helpful
4
Replies

EIGRP and default route

WILLIAM STEGMAN
Level 4
Level 4

I'm trying to influence a default route sent out by 2 different hub routers at 2 different data centers.  I want branches to prefer one default route over another.  This is in a DMVPN environment where each branch has 2 tunnel interfaces, one tunnel interface connects to the 2 different hub routers via MPLS at 2 separate data centers, and the other tunnel interface connects to 2 different hub routers via the Internet at the h2 different data centers (dual router single domain).  I have no problem influencing the routes that come out of a single data center via 2 different tunnel interfaces (mpls dmvpn router and Intnernet dmvpn router), however when I'm advertising a default route out of both data centers across the same tunnel interface, the path leads to the less desirable data center and its Internet connection.  I tried increasing the AD distance (200) using the source of the less preferred site, but it only modifies the networks advertised out of that data center that are not redistributed.  The MPLS DMVPN router at the less preferred data center learns about 0.0.0.0/0 via the core and EIGRP, but the core is using a static route that is redistributed into EIGRP. I've tried modifying the delay and bandwidth on the tunnel interface at the less preferred site as well.  I'm not sure how the feasible distance is being calculated and why one site is preferred over another that has greater bandwidth and higher delay.  

This is from a branch.  10.118.4.1 is where I want default traffic to go first.  

show ip eigrp top all

P 0.0.0.0/0, 1 successors, FD is 3343646720, serno 8411
via 10.118.4.2 (3343646720/1966080), Tunnel20
via 10.118.4.1 (3349544960/10485760), Tunnel20
via 10.118.0.1 (65608744960/10485760), Tunnel10

This is from the same branch; you can see the distance for non default networks has been modifed.

sh ip route

D*EX 0.0.0.0/0 [170/26122240] via 10.118.4.2, 00:04:28, Tunnel20
10.0.0.0/8 is variably subnetted, 16 subnets, 5 masks
D 10.0.0.0/24 [200/26122240] via 10.118.4.2, 00:04:28, Tunnel20

This is the less preferred MPLS DMVPN gateway

Tunnel20 is up, line protocol is up
Hardware is Tunnel
Description: description GRE/IPSEC Tunnel via MPLS to Remotes
Internet address is 10.118.4.2/22
MTU 9972 bytes, BW 100000 Kbit/sec, DLY 500500 usec,

This is the more preferred MPLS DMVPN gateway

Tunnel20 is up, line protocol is up
Hardware is Tunnel
Description: description Tunnel via MPLS to Remotes
Internet address is 10.118.4.1/22
MTU 9972 bytes, BW 1000000 Kbit/sec, DLY 50000 usec,

thank you,

Bill

update - I see the issue with the metric.  I hadn't considered the port channel interfaces connecting the different MPLS VPN routers to the core, they had different bandwidth #s.  So I'm good there, but am still curious why the admin distance command didn't work for the 0.0.0.0/0 network.

1 Accepted Solution

Accepted Solutions

Jon Marshall
Hall of Fame
Hall of Fame

Bill

As far as I am aware you cannot modify the AD of specific external routes.

So you can modify the AD of specific internal routes or change the AD for all internal routes.

However for external EIGRP routes you can only modify the AD of all external routes.

Don't ask me why :-)

Jon

View solution in original post

4 Replies 4

Jon Marshall
Hall of Fame
Hall of Fame

Bill

As far as I am aware you cannot modify the AD of specific external routes.

So you can modify the AD of specific internal routes or change the AD for all internal routes.

However for external EIGRP routes you can only modify the AD of all external routes.

Don't ask me why :-)

Jon

Interesting, I was using the config below, which does appear to follow your statement.  It works fine for all internal routes, it just has no impact on external routes. 

router eigrp IWAN
!
address-family ipv4 unicast autonomous-system 5
!
topology base
distance 200 10.118.4.2 0.0.0.0

thanks 

Yes, I can't find it documented and don't know the reason for it but I suspect that was the issue.

Jon

Hello

On the spoke router.

Have you tried just changing the delay on the tunnel or physical interface?

Edit -  didnt read you post correcty.

FYI
Name Eigrp compositic metrics are computed by a Metric-Rib-Scale and by default this is set to 128

Therefore  if the FD is divided by this RIB -Scale thats the value that is entered in the routing table

3343646720 / 128 = 26122240

P 0.0.0.0/0, 1 successors, FD is 3343646720, serno 8411
via 10.118.4.2 (3343646720/1966080), Tunnel20
via 10.118.4.1 (3349544960/10485760), Tunnel20
via 10.118.0.1 (65608744960/10485760), Tunnel10

D*EX 0.0.0.0/0 [170/26122240] via 10.118.4.2, 00:04:28, Tunnel20




res

Paul


Please rate and mark as an accepted solution if you have found any of the information provided useful.
This then could assist others on these forums to find a valuable answer and broadens the community’s global network.

Kind Regards
Paul
Review Cisco Networking for a $25 gift card