EIGRP Default Route AD Change

[jlizzio]

Hi Everyone,

I’m trying to come up with a solution for redundancy to our MPLS network and DMVPN seems like the obvious choice. Everything around DMVPN is working. My issue is a routing one.

I need to be able to adjust the AD of the default route being advertised to R1 from SPOKE using EIGRP. I don’t want to affect any other routes other than the default. Seems like such a simple thing to do. Maybe I’ve just been staring at this thing to long.

If I could change the AD for prefix 0.0.0.0/0 to 10 this will give us a preferred path to the Internet out our local ISP and a backup over the MPLS.

Much thanks,

-J

[chrihussey]

So I assume that Spoke and R2 are the ones advertising the default static routes via EIGRP?

If so, you could advertise the default route from the Spoke normally and give the R2 a weighted static with a greater administrative distance.

For example:

ip route 0.0.0.0 0.0.0.0 x.x.x.x 200 (Spoke would be 170)

This way R2 would not propagate the 0.0.0.0 route until the Spoke stops its advertisement.

Hope this helps.

[jlizzio]

Yes, sorry if I didn't make that clear. There is a default originate coming across through the MPLS that R2 is advertising. The spoke is advertising the default route based on an SLA. If the SLA fails then the MPLS route will pick up. Note - the SLA needs to be on the spoke as there are more than one ISP. 

Static is an option but not one I'm real excited about. 

In BGP this is a slam dunk with the ability to match on a prefix. With EIGRP it doesn't seem like there is a way to prefix match and modify the AD based on the match although it seems like there should be. 

[chrihussey · Accepted solution]

Been thinking about this. R1 is receiving the default route from Spoke and R2. I'm assuming that by the way routing is configured, R1 sees both default routes as equal cost and has both in its routing table.

Perhaps if you adjust the delay metrics on the link between Spoke and R1, the Spoke default route would be preferred over R2. Similarly, you could increase the delay metric between R1 and R2 and achieve the same result.

Regards

[jlizzio]

Hi Chrihussey,

K values... you study them, know them and wonder when will I use them. Well, today was that day. 

So what I did was increase the delay on R1 for the interface where the MPLS routes are learned from. It works and so far I haven't seen anything negative associated with it. 

Thanks to you, cofee and all the others who are helping provide solutions.

-J

[cofee]

Hi,

This is what I understand you are trying to achieve, R1 is currently receiving two default routes 1 from the spoke router and another from r2 and you want spoke router to advertise the default route to R1with an AD 0f 10.

If my understanding is correct than you can achieve this by following configuration on R1:

access-list 1 permit 0.0.0.0 255.255.255.255

!

router eigrp xx

distance 10 1.1.1.1 0.0.0.0 1

(1.1.1.1 is the source address and 0.0.0.0 is wild card mask, so you will replace 1.1.1.1 with whatever address spoke is using as next hop value when it advertise default route to r1 and 1 is the acces-list number)

!

* Below I have two outputs from the lab, with and without route manipulation:

Gateway of last resort is 35.1.1.5 to network 0.0.0.0

D*    0.0.0.0/0 [90/30720] via 35.1.1.5, 00:03:10, FastEthernet0/1
                [90/30720] via 34.1.1.4, 00:03:10, FastEthernet0/0

Gateway of last resort is 34.1.1.4 to network 0.0.0.0

D*    0.0.0.0/0 [10/30720] via 34.1.1.4, 00:01:18, FastEthernet0/0

[jlizzio]

Hi Cofee,

Thanks for your suggestions. I should probably clarify a couple of things. 

The Spoke router is redistributing the default route. It is not learned. There is an SLA monitoring Internet reachability. If the SLA fails, the route is removed and the one learned from the MPLS should be used.

The distance option with the ACL changes all learned routes and not the redistributed ones.

Maybe there is a better way to accomplish redundancy for Internet?

Thanks,

-J

[cofee]

Hi,

Distance option will only change AD for the subnets that are matched in the ACL.

[cofee]

Hi,

I see what you are saying.  Manipulation of AD using ACLs in conjunction with distance or setting metric under the route-map seem to have no affect on redistributed routes, no matter where you configure it. I believe even changing the AD of external eigrp routes won't help because both default routes are being learned by R1 as external eigrp.

You can tweak the metric  as Chrihussey  has suggested, but this would affect all other prefixes as well learned/advertised by eigrp, to overcome that you can use the distance feature for all the internal eigrp prefixes to prefer one over the other. Not sure if PBR will be beneficial in this situation without knowing the details.

I can't think of resolving this any other way, in my opinion you would need to make these tweaks to achieve the target traffic flow.  Traffic engineering features are limited when it comes to IGPs.