This recipe assumes that you have read the previous recipe, Hot physical backups with Barman, and successfully installed Barman on the malcolm server, backing up the Postgres databases running on angus and bon. We will use the same nomenclature in the examples of this recipe.

A recovery procedure is a reaction to a failure. In database terms, this could be related to an unintentional human error (for example, DROP operation of a table), an attack (think of Little Bobby Tables), a hardware failure (for example, a broken hard drive), or—less likely—a natural disaster.

Even though you might be tempted to think that you are immune to disasters or failures (I wish you were), you are advised to perform regular tests and simulations of recovery. If you have a team of engineers, I suggest that you schedule a simulation every 6 months (at least) and regularly test your backups through the safest way of checking their content—performing a recovery.

You don't want to be somebody who...