An OceanBase cluster consists of multiple OBServers that are deployed across zones.
The cluster name and cluster ID are the unique identifiers of an OceanBase cluster. You can create, restart, and take clusters offline. In addition, you can scale clusters by adding or removing the zones and OBServers of the clusters. OceanBase clusters are elastic. You can scale out or scale in OceanBase clusters. OceanBase clusters use a triplicate architecture to support zone-disaster recovery. Based on this, you can deploy five replicas in three data centers to implement geo-disaster recovery. Scale-in and scale-out can also be used to replace and repair devices. When a hardware fault occurs on a server in a cluster, you can add a new server to the cluster and remove the faulty server from the cluster. After this process, the number of cluster nodes remains unchanged. The RootService of the cluster automatically detects new nodes and supplements replicas.