Highly available routers can be created and managed using the same router command set that was discussed in the previous chapter. The L3 agents are responsible for configuring the routers in a VRRP group, and the routers are left to elect the master router and implement their respective keepalived configuration.

While HA routers provide a level of redundancy over their standalone counterparts, they are not without their drawbacks. A single node hosting a master router is still a bottleneck for traffic traversing this router. In addition, if the network used for dedicated VRRP traffic between routers experiences a loss of connectivity, the routers can become split-brained, causing two or more routers to become master routers and potentially causing ARP and MAC flapping issues in the network. Connection tracking between routers has not been implemented as of the Kilo release, which means that connections to and from instances may be severed during a failover event. In addition to the previously mentioned limitations, there are also outstanding bugs in relation to router failover and L2 population that affect both LinuxBridge- and Open vSwitch-based Juno and Kilo environments and should be addressed in a future release of OpenStack.

In the next chapter, we will look at distributed virtual routers: a method of distributing routers across compute nodes that are responsible for routing traffic for their respective instances.