Why you still need backups with a Postgres cluster

Copies between nodes keep your database running when a computer dies, but not from mistakes. Here is why a Postgres cluster still needs backups.

A Postgres cluster and a backup protect you from different things. You want both.

What the cluster protects you from

CoyoteDB keeps a full, live copy of your data on every data node. If a computer dies, its disk fails, or it loses power, another node has everything and carries on. See High availability Postgres, in plain words.

What it does not protect you from

Replication copies every change to every node, straight away. That includes the changes you did not mean to make:

  • a table dropped by mistake,
  • a bad update that overwrites the wrong rows,
  • a bug in your app that deletes data it should not,
  • data damaged by someone who should not have had access.

A second, third or fourth copy of a mistake is still a mistake. Only a copy from before it happened lets you go back.

What a backup gives you

A backup is a copy of your data at a point in time, kept apart from the live cluster. When something goes wrong, you restore from a moment before it did.

Backups in CoyoteDB

CoyoteDB does not make scheduled backups for you yet. Back up the way you would back up any PostgreSQL database: CoyoteDB runs real PostgreSQL 18, so the standard Postgres tools work with it.

Read more: What “saved on two computers” means, Node counts explained.