Skip to main content
Use backup when database state and persistent application files must move or recover together. The resulting directory contains a logical MariaDB dump, archives for the selected non-database volumes, and metadata needed for restore.
Restore replaces current state. Keep a separate copy of the current site, inspect the selected context, and stop customer writes before proceeding. After restore, run sitectl healthcheck followed by the application’s sitectl verify checks.

​
Reference

Create or restore a full-site database and file-volume backup

​
backup create

Create a compressed logical MariaDB dump plus tar archives for named Compose volumes not mounted as MariaDB data. sitectl briefly stops the running application services while leaving MariaDB available, then resumes exactly the services that were running even when backup creation fails. Full-site backup is currently supported from Linux operator hosts.

​
backup restore

Validate and stage the complete backup before stopping the site, retain recovery snapshots of current file and MariaDB volumes, then replace the site. If a commit step fails, sitectl restores every retained snapshot and restarts the stack. External Compose volumes are never modified. This destructive operation requires --yolo after the operator verifies the selected backup and is currently supported from Linux operator hosts.