Active and Inactive

At any moment in time, each member of a fault tolerance group is either active or inactive. An active member directly fulfills the program’s mission—for example, broadcasting information, responding to queries, or filling requests. An inactive member provides backup capacity; inactive members wait in the background, ready to become active immediately if an active member stops functioning.

When a new member joins a group, it is initially inactive.

Each member incorporates Rendezvous fault tolerance software, which operates behind the scenes. Rendezvous fault tolerance software maintains the correct number of active members by issuing instructions, called actions, to group members. Each action instructs a member to activate, deactivate, or prepare to activate. (For details, see Fault Tolerance Callback Actions.)

Rendezvous fault tolerance monitor software can passively monitor the members of a group, so other programs can determine the number of active members without actually joining the group. For a description, see Passive Monitor.

Alternate Terminology

Some approaches to fault tolerance use different terms to describe active and inactive members. In particular, an active member is called primary, while inactive members are called secondary.

This terminology implies that only one member can be active—the primary. However, Rendezvous fault tolerance software allows any number of active members. This book uses the more general terms, active and inactive.