Master‑Slave Replication in Redis: Full and Incremental Sync Explained

Redis Master‑Slave Replication Overview

Redis is a critical middleware in large‑scale architectures. Master‑slave replication copies data from a master node to one or more slave nodes, similar to MySQL replication, and consists of full‑copy and incremental‑copy phases.

By default each Redis instance acts as a master; a master may have multiple slaves, while a slave is linked to a single master.

Key Benefits

Data backup – slaves can take over when the master fails, enabling rapid failover.

Read/write separation – the master handles writes, slaves serve reads, improving load capacity.

High availability – master‑slave replication is the foundation for Redis Sentinel and cluster solutions.

Full‑Copy Replication Process

The initial synchronization follows seven steps:

Slave sends PSYNC to the master and establishes a long‑lived socket connection.

Master receives the command and executes BGSAVE to generate an RDB snapshot.

Master transmits the RDB file to the slave.

Slave clears its data and loads the received RDB.

Master streams any write commands that occurred during snapshot creation from its replication buffer to the slave.

Slave executes the received write commands.

Master continuously pushes subsequent write commands over the socket to keep master and slave consistent.

Incremental Replication Process

If the connection is lost and the master continues to receive writes, reconnection triggers incremental sync, which follows six steps:

Slave detects the broken connection.

Master retains recent write commands in its replication buffer.

Slave re‑establishes a socket connection to the master.

Slave issues PSYNC with an offset to indicate the last processed command.

Master checks whether the slave’s offset lies within the buffer; if so, it streams the missing commands, otherwise it falls back to a full copy.

Master continues to push new write commands over the socket to maintain consistency.