Data replication for resilience and availability

This layer is responsible for ensuring that the table data is replicated to more than one node and also keeps the data consistent between replicas.

The replication factor indicates how many replicas of a specific table's data should be kept—for example, if the replication factor is 3, CockroachDB keeps three copies of all the table data. The number of node failures that can be tolerated without data loss = (replication factor – 1) / 2; for example, if the replication factor is 3, then (3 – 1) / 2 = 1 node failure can be tolerated. Whenever a node goes down, CockroachDB automatically detects it and works toward making sure the data in the node that went down is replicated to other nodes, in order to honor the replication factor and also to increase survivability.

CockroachDB uses the Raft distributed consensus algorithm, which ensures a quorum of replicas agree on changes to ranges before those changes...