New CoreOS style appliances now come with a /backup URL endpoint from UpGuard v2.18, which you can use to retrieve snapshots of your database for restoring from in the case of a total system failure.

Background

Postgres runs as a microservice on the container-style architecture on CoreOS. UpGuard uses the continuous archiving method of backup by ensuring that PostgreSQL maintains a write ahead log (WAL) in the pg_xlog sub-directory inside of the $PG_DATA directory. The existence of these WAL files help to bring the system back to a current state from a file system backup in conjunction with a replay of the backed-up WAL files.

Normally, should the database file directory be destroyed or corrupted, restarting the CoreOS appliance will automatically restore the database from the latest file system level backup along with the WAL files on booting up.

Downloading Nightly Backup Files

We have provided a set of Windows and Unix/Linux scripts for running a scheduled download of the nightly backups for your use in the respective sections below. The Linux script can be run with a crontab task as specified in the comments, or if you’re on Windows, a Scheduled Task can be run for the Powershell script.

If you wish to do so manually, the endpoint for accessing the backups can be found on your appliance URL as follows:

https://<appliance IP or hostname>/backups

Linux Download Script

Recovery Process

In the event that you wish to perform a Point-in-Time recovery of your database due to an incident, please contact the UpGuard support team at support@upguard.com to schedule a remote session with us to better assist you.

Tags: appliance