HSRP is an awesome tool, and coupled with interface tracking, it can be the means for near-total redundancy. There are situations, however, where HSRP is not enough. The example I will show here is one of my favorite interview questions, because usually only someone with real-world experience in complex networks has seen it.
Figure 13-4 shows a deceptively simple HSRP setup. Two locations, New York and Los Angeles, are connected via two T1s. The routers on either side are connected via the F0/1 interfaces, and HSRP is implemented with interface tracking on the F0/0 interfaces. The idea here is that if either of the primary routers should fail, the secondary routers will take over for them. Additionally, should the primary T1 link fail, the secondary link will take over, because interface tracking is enabled.
Here are the primary Ethernet configurations for each router:
interface f0/0 ip address 10.10.10.2 255.255.255.0 standby ip 10.10.10.1 standby preempt standby track Serial0/0 10
Figure 13-4. Two-link failover scenario using HSRP
interface f0/0 ip address 10.10.10.3 255.255.255.0 standby ip 10.10.10.1 standby priority 95 standby preempt standby track Serial0/0 10
interface f0/0 ip address 18.104.22.168 255.255.255.0 standby ip 22.214.171.124 standby preempt standby track Serial0/0 10
interface f0/0 ip address 126.96.36.199 255.255.255.0 ...