Introduction
To become IPv6 BGP peers, any two routers need to establish a TCP session using port 179 and exchange open messages successfully. This document describes few scenarios wherein BGP peering is not established beween two Ipv6 BGP speaking routers.
Prerequisite
It is assumed that the reader has an understanding of IPv6 and BGP routing protocol.
Background Information
The two routers (R1 and R2) depicted in the topology diagram are in AS 100 and AS 200 forming an EBGP relationship. Their loopbacks are advertised via OSPFv3 routing protocol. There is another Router, R3, in between R1 and R2. The three routers are connected via Fast Ethernet interfaces and are configured with OSPFv3 routing protocol. As we wanted the BGP peering to established over the Loopback interface and not via the directly connected interfaces, Router R3 is not depicted in the topology diagram. Please see the configurations attached to the document.
Note:
Topology Diagram
Configuration
Please see attached files Router R1, Router R2 and Router R3.
Note: All configurations are tested on Cisco 3745 Router running IOS 12.4 software.
Troubleshooting Steps
Step1. AS Number
Symptom Error Message: %BGP-3-NOTIFICATION: received from neighbor 2012:ABC:123:22::2 2/2 (peer in wrong AS) 2 bytes 0064
Verify that the peer’s AS number is correct by command “show bgp ipv6 unicast summary”
Note: As seen from the output, the neighbor 2012:ABC:123:12::2 is configured in AS 123, however Router R2 is residing in AS 200.
Step2. Connectivity
- Ping the neighbor by command ping ipv6 X:X:X:X::X
- Verify that the peer is listed in the IPv6 RIB
Note: In this scenario, the peering neighbor is 2012:ABC:123:11::1, the route is learned via OSPFv3.
Step3. Update-Source interface
If a loopback interface is used to form peering, check that the command “update-source loopback” is configured under the BGP configuration mode. Without the neighbor update-source configuration command, the TCP session will use the IP address of the outgoing physical interface and the neighbor will reject the incoming TCP SYN packet as it’s not coming from a recognized BGP neighbor.
Note: The state remains in Active mode.
Step4. Ebgp Max-Hop count
If the peer is not directly connected, check that the peer ebgp-max-hop command is configured.
Step5. TCP connection
Check the TCP connection by command “show tcp brief”
Note: The state should be established.
Step6. Access-List/Routing policy
Check whether an ACL for disabling TCP port 179 is configured.
Step7. Run Debugs
Use the command "debug bgp ipv6 unicast"
References
Implementing Multiprotocol BGP for IPv6
