Since its introduction in 9.1, [synchronous replication] (https://www.postgresql.org/docs/devel/static/warm-standby.html#SYNCHRONOUS-REPLICATION), or the ability to wait for a WAL flush confirmation from a standby before committing a transaction on the master node (depends on synchronous_commit whose default value is on), ensuring that the transaction committed has not been lost on the standby should a failover need to be done, has faced a wide adoption in many production environments for applications that need a no-data-loss scenario. Still, such application or clients may have seen something like that:
Read more...