The PostgreSQL agent offers three options for a cluster configuration. In these three options, two options leverage the Write Ahead Log (WAL) file shipping features and require the installation of PostgreSQL pg_standby utility. The various options for cluster configuration are the following:
Traditional HA configuration with shared storage. In this configuration, you have a cluster with an active PostgreSQL resource, where the database directories reside on a global or a failover file system.
WAL file shipping between two PostgreSQL failover resources. In this configuration, you have two independent PostgreSQL resources in a cluster or in different clusters. One of the resources acts as a primary server and obtains the client requests. The other resource acts as a standby server applying the PostgreSQL WAL files shipped from the primary server.
WAL file shipping without shared storage. This configuration does not require shared storage. The PostgreSQL WAL file shipping replaces the shared storage. This configuration consists of three resource groups. In two single-node resource groups, one resource group contains the designated primary database resource. The other resource group contains the designated standby database resource. The third resource group contains a logical host and a Rolechanger resource. This Rolechanger resource is responsible for transforming the designated standby into an acting primary on a node outage of the designated primary.