Database Replication

Normally we have more read than write ratio. So the number of slave database is normally higher than master database

Advantages

• Better performance: distribute read and write, allows queries to be processed parallely
• Failover: if master dies, we promote a slave up to be master. There is no dataloss
• High availability: even if the database is offline, you can still access the data in other databases
  • If a slave is offline, a new slave will be launch up to replace unhealthy slave
  • If master is offline, a slave is promoted to become master and a new slave is launched to replace it.

Problem

There could be stale data when user reading from slave database. This is because the data has not yet been synced to the slave database.

• Consider: Database Replica lag solution