Self-hosting

Update, back up and restore

Keep the server current and your data safe.

Update

git pull
make up

When the server starts, it applies any new database migration. Before it does, it saves a copy of the database named pre-migrate-v<N>-<time>.db in the data volume. That copy is your way back if an update goes wrong.

Back up

A backup service saves a snapshot of the database every night and mirrors the images to a folder. Point that folder at another disk or a network share, so a copy lives off the machine.

1. 01
   Mount the disk or network share on the machine, and make sure Docker is allowed to use that path.
2. 02
   Set SHOP_BACKUP_DIR in deploy/.env, for example /Volumes/backups/shop.
3. 03
   Start the stack again, then run a backup now to check it works.

   make up
   make backup

Backups run at 03:30 by default and snapshots are kept for 30 days; the newest three are always kept. If the folder is not available, the next run fails and the app shows NAS backup is out of date in Settings.

Restore

You need the repo and your backup folder.

1. 01
   Stop the stack.

   docker compose -f deploy/docker-compose.yml down

2. 02
   Restore the latest snapshot. This checks the snapshot, replaces the database and copies back any missing images.

   docker compose -f deploy/docker-compose.yml run --rm --no-deps backup restore -yes -snapshot latest

3. 03
   Start again and check.

   make up
   curl localhost:8080/healthz

Phones and tokens keep working after a restore, because they are stored in the database.

To go back after a failed update, use the pre-migrate copy instead. The steps are in the restore guide.