Routing

Hi,

 

**bleep** BGP 🙂

Here is the topology we are using:

 

CPE --> LNS1 --> Core 1 --> Cisco eBGP peer1 --> Loopback 

                               |

                               |

CPE --> LNS2 --> Core2 --> Cisco eBGP peer2 --> loopback

 

I have configured loopback interfaces on both eBGP peers to simulate Internet addresses. The addresses are different on the peers. I used 192.168.1.11/32 -12/32 - 13/32 on peer1 and 192.168.150.11/32 - 12/32 - 13/32 on peer 2.

I have configured a default route on Core1 and Core2 (Maybe this is wrong) as we need the customers entering on LNS1 to exit via Core1 and customers entering on LNS2 to exit via core2.

When I view the routing table on LNS1 I see the route to peer1 loopbacks as via the default and no BGP in the routing table (Perfect).

Likewise, when I look on LNS2 I see the routes for Peer2 loopbacks and no BGP routing table.

However, when I look at the route for peer1 on core2 I see the default as "Discard" but yet, when I look at the route for peer1 loopbacks on LNS2, it shows as via the default. I obviously cannot ping this.

It is the same the same vice versa too.

So, here are my questions:

 

1: Is this correct given that there is no iBGP currently?

2: Do I need to configure iBGP for failover if I want customers from LNS1 to exit via core2 (remembering that these core routers will have the internet routing table advertised to them so will be the same addresses)?

3: If the answer to 2 is yes, then how can I ensure that during normal operation customers from LNS1 exit via core1 and only via core2 if there is a failure?

 

It's a bit tricky trying to get an answer via google for these questions so please accept my apologies in advance

 

Thanks

Hi,


1: Is this correct given that there is no iBGP currently?

Ans. >> Can you share your configuration on core 1 and core 2 and paste the output of show route from both core routes?

2: Do I need to configure iBGP for failover if I want customers from LNS1 to exit via core2 (remembering that these core routers will have the internet routing table advertised to them so will be the same addresses)?

ANS >>> Yes, If there’s no direct connectivity between LNS1 and core 2 then you need to run iBGP between core and advertise the default route to each other.. btw, what are you running between LNS and Core? Default-routing?

3: If the answer to 2 is yes, then how can I ensure that during normal operation customers from LNS1 exit via core1 and only via core2 if there is a failure?

Ans >> Short answer “BGP Atrributes” you need to know the BGP path selection algorithm. You can change the Local preference, origin code etc to make it work.

Hi Kingsman,

 

Thank you for the response.

 

I have added iBGP between the loopbacks and configured a policy to advertise the defaults and it all worked perfectly. I got an "Established" session and the routes are no longer in a "Discard" state. Perfect. Thank you. Next up is trying to work out "IPv6" advertisement and for this I guess I will need MP-BGP. Probably a new question then.... but for now, thank you very much for the help.

 

Let me just place here what I configured in case anyone else has the same issue:

 

Core 1:

set policy-options policy-statement internal-bgp-route term 2 from protocol direct
set policy-options policy-statement internal-bgp-route term 2 then accept
set protocols bgp group internal-peers type internal
set protocols bgp group internal-peers description "Connection to HEX-CORE-02-loopback0"
set protocols bgp group internal-peers local-address xxx.xxx.xxx.xxx (Local loopback address)
set protocols bgp group internal-peers export internal-bgp-route
set protocols bgp group internal-peers peer-as xxxx (Local AS number as it is iBGP)
set protocols bgp group internal-peers neighbor xxx.xxx.xxx.xxx (Peer loopback address)

 

Core2:

set policy-options policy-statement internal-bgp-route term 2 from protocol direct
set policy-options policy-statement internal-bgp-route term 2 then accept
set protocols bgp group internal-peers type internal
set protocols bgp group internal-peers local-address xxx.xxx.xxx.xxx (Local loopback address)
set protocols bgp group internal-peers export internal-bgp-route
set protocols bgp group internal-peers peer-as xxxx (Local AS number)
set protocols bgp group internal-peers neighbor xxx.xxx.xxx.xxx (Peer loopback address)

 

Hope this helps someone