R1 is connected to R2. They peer to each other on their fa0/0 interface IP's (i.e. directly connected IPs). I know the bgp fast-external-failover feature should instantly shut the eBGP peer down when the physical connection fails. So I tested this by shutting down R1's fa0/0 interface
*Mar 4 19:39:07.593: %BGP-5-ADJCHANGE: neighbor 10.0.12.1 Down Interface flap
And sure enough the peer goes down right away. However on R2 I observed that this fast-external-failover feature does not work on that side once the link on R1 was shutdown. It instead waited for the hold down time to expire, as shown below. I don't understand why? This lab is built in GNS3, so I'm wondering if it's just GNS? Maybe someone has some real hardware that they can test this on?
*Mar 4 19:42:16.409: %BGP-5-ADJCHANGE: neighbor 10.0.12.2 Down BGP Notification sent
*Mar 4 19:42:16.409: %BGP-3-NOTIFICATION: sent to neighbor 10.0.12.2 4/0 (hold time expired) 0 bytes