This should get you started. Timeouts, thresholds etc. are all configurable. Change values to suit your needs.
ip sla monitor 25
type echo protocol ipIcmpEcho 20.x.x.45 <====upstream neighbor ip
timeout 10000
threshold 7
frequency 10
ip sla monitor schedule 25 life forever start-time now
track 123 rtr 25 reachability
delay down 60 <=== this delays the track from going down until down for 60 seconds
interface FastEthernet0/0 <===primary connection
description outside
ip address 172.25.0.2 255.255.255.252
interface FastEthernet0/1
description inside
ip address 172.25.0.5 255.255.255.252
interface Serial0/0/0 <====backup connection
description outside
ip address 172.26.0.13 255.255.255.252
ip access-group 120 in
ip local policy route-map FAILOVER-POLICY
ip route 0.0.0.0 0.0.0.0 172.25.0.1 track 123
ip route 0.0.0.0 0.0.0.0 172.26.0.14 119 <====119 because I'm running rip, could be anything less than 120
ip route 20.x.x.45 255.255.255.255 172.25.0.1
access-list 101 permit icmp any host 20.x.x.45 echo
access-list 120 deny icmp host 20.x.x.45 any <===makes sure ping replies don't come from secondary link
access-list 120 permit ip any any
route-map FAILOVER-POLICY permit 10
match ip address 101
set interface Null0 Serial0/0/0
set ip next-hop 172.25.0.1
So, to explain all that in plain english. An ip sla process starts and pings 20.x.x.45 every 10 seconds. As long as that ping is successful, the default route is to 172.25.0.1. If it fails, for me I have it set for 60 second delay, then the default route is 172.26.0.14. The process continues to ping until successful. At that time it reinstates the original tracked route.
Please rate helpful posts.
