ANALISIS PERANCANGAN CLUSTER REPLICATION PADA DATABASE POSTGRESQL MENGGUNAKAN METODE HOT STANDBY

Postgresql is a familiar database and is often used by application developers, because of the advantages of the Postgre database, namely because of the very large record capacity so that when carrying out large amounts of data CRUD activities, this database is prioritized. Postgre is usually used in the e- commerce industry and the banking world, however, due to the high transaction traffic from clients, it sometimes causes the main server to crash resulting in data being lost and transactions being delayed. The method used is Replication Hot Standby on the Postgresql cluster. In this method there are two servers, namely the master server as the main server, and the slave server as data backup/replica. In its design, the master server is designed as read/write, while the slave server is read-only, so that when the master node crashes or goes down, the standby server can be promoted to a new master using Manual Failover. But until now there are still many companies that have not used this method because it is considered data redundancy, even though by using the pg basebackup method, data on the old server can be removed and new data added so that the data can be updated. The results of this study are that by implementing a replication cluster on the Postgresql database using the hot standby method there is a decrease in database performance when replicating data from master to slave because there is no division of workload on the main server, but the advantage is that data is still available even though the master server experience a crash or down, because there is a standby server that is used as a replica that can back up data from the master server in real time.

Keyword: Postgresql, Cluster, Replication, Hot Standby, Failover