Embedded Postgres Backup ​

You can enable a simple, automated daily backup of an embedded Postgres database.

For deployments using an embedded (containerized) Postgres database it's recommended to have a backup/restore strategy. An automated daily backup can be enabled in the GUI config under the "Database" section with the "Embedded Postgres Backup" option. Alternatively, when using a Helm-based install it will be enabled by default, and can be disabled by adding the following to your local values.yaml file:

paramify:
  pgbackup:
    enabled: false

paramify:
  pgbackup:
    enabled: false

When enabled, a daily cron job will be deployed to generate backups, upload them to your object store, and prune old versions over time. It can also be executed manually via kubectl exec to list available backups and restore a selected backup.

Backup ​

Once setup as a daily cron job it will execute with the default backup command, which does the following:

• Executes standard pg_dump, generating a dated backup file like postgres_2024-01-01_12-30.dump
• Uploads the backup file to the configured object store in a backup folder
• Prunes old backups, retaining a specified number of daily, weekly, and monthly backups

To manually execute a backup use the following:

# Startup the convenience container to run manual backup/restore commands
kubectl -n paramify scale sts pgbackup --replicas=1

# A backup can ge generated manually by executing the "backup" command
kubectl -n paramify exec -it pgbackup-0 backup

# The convenience container can be shutdown when it's not needed
kubectl -n paramify scale sts pgbackup --replicas=0

# Startup the convenience container to run manual backup/restore commands
kubectl -n paramify scale sts pgbackup --replicas=1

# A backup can ge generated manually by executing the "backup" command
kubectl -n paramify exec -it pgbackup-0 backup

# The convenience container can be shutdown when it's not needed
kubectl -n paramify scale sts pgbackup --replicas=0

Restore ​

When the container is started with restore (no backup specified) it will output the list of available backup files from the object store.

For example (replacing -n paramify with the appropriate namespace, if different):

# Use "restore" to list available backups
kubectl -n paramify exec -it pgbackup-0 restore

# Use "restore" to list available backups
kubectl -n paramify exec -it pgbackup-0 restore

Once the desired backup is identified, the job can be executed with restore <backup> to do the following:

• Download the specified backup file
• Execute pg_restore of the backup file to restore the database to that version

# First, shutdown the Paramify application by scaling app to 0 pods
kubectl -n paramify scale --replicas=0 deploy paramify

# Then use "restore <backup>" to restore a selected backup
kubectl -n paramify exec -it pgbackup-0 restore postgres_2024-01-01_12-30.dump

# Finally, restart the Paramify application by scaling app back to 1 pod
kubectl -n paramify scale --replicas=1 deploy paramify

# First, shutdown the Paramify application by scaling app to 0 pods
kubectl -n paramify scale --replicas=0 deploy paramify

# Then use "restore <backup>" to restore a selected backup
kubectl -n paramify exec -it pgbackup-0 restore postgres_2024-01-01_12-30.dump

# Finally, restart the Paramify application by scaling app back to 1 pod
kubectl -n paramify scale --replicas=1 deploy paramify

NOTE: It's recommended to shutdown the application before performing the restore, but required to restart the application after the restore is complete.