Cluster Principles
An etcd cluster consists of a management node (the leader) and multiple compute nodes (followers). The leader synchronizes logs to followers to ensure that each member in this cluster has a copy of data. The leader and followers must be on the same network and can communicate with each other.
When installing and starting the etcd service, each node needs to know the information (generally the IP addresses and ports) about other nodes in the cluster. An etcd cluster can be started in two modes: static configuration and service discovery. This document uses the static configuration mode, which is applicable to the offline environment where the number and IP addresses of cluster nodes are known.
Figure 1 etcd cluster principles
Parent topic: Cluster Principles and Networking