How was the PostgreSQL cluster organized? Replication scheme and fault tolerance?

The organization of a PostgreSQL cluster for high availability and replication typically includes the following components:

1. Primary server — accepts all data writes and changes.
2. Standby/Replica — receives data from the primary server in real-time or with a slight delay.

Replication schemes:

• Streaming Replication — streaming replication where replicas receive WAL (Write-Ahead Log) records directly from the primary server.
• Logical Replication — replication at the logical change level, allowing more flexible subscription and filtering.

For high availability, the following are usually used:

• Failover — automatic or manual switch to a replica in case of primary server failure.
• Monitoring and cluster managers (e.g., Patroni, repmgr) — ensure automation of failover and status control.

Example:

• The primary server is configured with wal_level = replica
• Replicas connect via streaming replication
• Patroni is used for automatic failover and cluster management

Thus, the cluster provides high availability and minimizes downtime during failures.