Introduction

Every time the CloudSoda software is updated, it performs a local backup of its Postgres database. This is to be used in the event that the upgrade needs to be rolled back. However, when not running the hosted version of CloudSoda, backing up CloudSoda is the responsibility of the customer.

The following sections provide more information based on the various methods on which CloudSoda runs.

VMware

When CloudSoda is being run in an on-prem Virtual Machine (VM), it is recommended that CloudSoda is backed up in the same manner that other VMs are being backed up.  The key directory where data persists is /soda. Only backing that directory up can facilitate a disaster recovery redeployment, but your organization's recovery point objective (RPO) must tolerate a longer outage than if using a snapshot-style backup method, which backs up and restores full system states.

Cloud

When running CloudSoda in one of the cloud providers, configuring a disk snapshot is the recommended approach. That snapshot interval should be in line with the your organization's RPO.

Metal

If CloudSoda is running on a bare-metal server, much like the VMware backup, the /soda folder must be backed up.  Backing up that folder should be in line with your organization's RPO.