EIGRP over BGP.

[saurav.khanna]

I faced a strange issue whilke troubleshooting. What I saw was that Cisco router preffering External EIGRP routes ove External BGP routes which is quite strange.

Check the routing file which I have attached.

IOS Image: flash:c1700-ipbase-mz.123-1a.bin

The n/w setup is like 3 routers connected back to back A --- B -- C Now between A and B I have EIGRP and B and C I have BGP. A takes the external routes and sends it to B via EIGRP and C sends BGP routes to B. The issue is with the timing if Router B learns the EIGRP route first then that route is installed in the routing table and the BGP one is not installed. Now both are external routes so AD for BGP is less hence that one should get installed not sue what's happening over here. To correct this I have to flap the link b/w A and B the BGP routes are installed.

[mheusinger · Accepted solution]

Hi,

from what I can see from your attached document, the problem is somewhat more complicated. It seems to me you are redistributing EIGRP into BGP on B. This can explain the behaviour, when taking into account BGP path selection and how redistribution works.

Scenario A) EIGRP route learned first

The external EIGRP route enters the routing table and is redistributed into BGP. If the BGP updates about the same network is received, BGP path selection within BGP will select the best one. BGP path selection:

1) higher weight

2) Local Preference

3) locally originated routes

And step 3 clearly selects the redistributed path entry as best and thus does NOT replace the EIGRP entry. Note: eBGP and iBGP administrative distance are used AFTER best path is selected to "fight" for insertion into the routing table. AD plays no role during BGP path selection.

So in Scenario A) the external EIGRP route is in the routing table and remains there.

Scenario B) eBGP route learned first

Then the eBGP route is in the routing table when the EIGRP update arrives. As the AD of eBGP is better, the EIGRP rout will not be entered into the routing table and thus NOT be redistributed. Redistribution occurs only for routes entered into the routing table. As there is no second BGP path in the BGP table no path selection takes place.

So here the BGP route is in the routing table.

Works as designed

You can confirm this by looking for your BGP table entries in each scenario.

If you want BGP to always be entered into your routing table, then use a higher Local Preference for the eBGP routes received. This can be f.e. achieved through a

route-map

router bgp 65000

neighbor 1.1.1.1 remote-as 65001

neighbor 1.1.1.1 route-map SetPref in




route-map SetPref permit 10

set local-preference 120

Hope this helps! Please rate all posts.

Regards, Martin

[gauravprakash]

Thx, Martin.

It looks you got the

crux..:)

Rgds,

G

[mchoo2005]

Something does not look quite right there.

What I don't understand is, how come the same route is learned by your router from EIGRP and eBGP? If it's from EIGRP and iBGP, I can understand. What BGP AS is

CPH017027

part of? I suspect there's inappropriate mutual redistribution taking place.

1. Investigate where the route is supposed to be, make sure that it is summarised/originated by the closest BGP speaker into the appropriate BGP AS.

2. When mutually redistributing between EIGRP and BGP, make sure you filter the route from being redistributed back into BGP.

It would help more if you could provide us with a bit more information. At least: rough diagram, configs of routers in question (

CPH017027

eBGP neighbor(s) of the first router, EIGRP neighbor(s) of the first router).