Master-slave replication schematic:
Most operations are performed on the master server, and a small number of operations are performed on the slave server. The slave server automatically requests the master server to backup the master server information. The master server dumps the rdb , that is, sends the information to the slave server in the form of a snapshot. Information is being saved on the master server, and is saved to each slave server in the form of aof logs.
redis cluster:
When the master server goes down, one slave server becomes the master server, and other slave servers back up information from this server