Normal HotStandby operation is highly reliable. The odds of both the Primary and Secondary failing at nearly the same time are low, if they use separate reliable power supplies. But suppose that you want to reduce even the risk of power failure, or suppose that the server that failed cannot be repaired rapidly? Ideally, when a Primary fails and you replace it with a Secondary (or when a Secondary fails), you want to have a new Secondary that replaces the old Secondary so that you can continue to run with a complete pair of servers.

This situation is called the 2N + 1 Spare (or 2N + M Spares) scenario. You have N Primary servers, N Secondary servers, and at least 1 spare that can replace any Secondary that has failed or has been converted to a Primary. Spares are not dedicated to a particular server (or HSB pair of servers), and some configuration is required before the spare can replace the failed server.