What is the mechanism in replication where an automatic leader election occurs after a main server failure?

The mechanism is called leader election. In replication systems, such as distributed databases or clusters, when the primary server (leader) fails, a process of selecting a new leader from the available replicas occurs. This process ensures continuity of operation and data consistency.

In Go and related systems, algorithms like Raft or Paxos are often used, which implement a reliable leader election mechanism and state agreement among nodes.

An example of simplified leader election logic:

// Pseudo-code for leader election
func electLeader(nodes []Node) Node {
    for _, node := range nodes {
        if node.isAlive() {
            return node // the first available node becomes the leader
        }
    }
    return nil // if no nodes are available
}

In real systems, leader election is more complex and considers timeouts, votes, and node states.