Database Snapshots
stry can dump its PostgreSQL database to a disk on a schedule. A snapshot holds your users, profiles, videos, tags, chapters and settings, so you can bring an install back after a failed upgrade or a lost database volume. It uses spatie/laravel-db-snapshots.
A snapshot only contains the database, not the media files. Back up the folders or buckets of your disks too, following the 3-2-1 backup plan. A snapshot without the files it refers to can't play anything.
Schedule snapshots
Open Admin → Snapshots:
| Setting | What it does |
|---|---|
| Frequency | Never (the default), Daily, Weekly or Monthly |
| Snapshots to keep | How many snapshots are kept (1–100, default 7). Older ones are deleted after a new one is made |
| Compress | Gzip the snapshots (on by default). They get much smaller, but take longer to create and restore |
Snapshots are created at 04:30 UTC, after the other nightly maintenance tasks. Weekly snapshots run on Mondays and monthly ones on the first of the month.
The schedule service creates them, so it must be running. A changed frequency applies from the next minute, without a restart.
With the on-demand socket, the scheduler timer stops while the app is asleep, so no snapshot is made then. Run app:snapshot yourself, for example from a host cron job or a systemd timer, if the app is often idle at night.
Where snapshots are stored
Snapshots go to the disk that's the default for Snapshots under Admin → Storage. When you migrate, stry adds a snapshots disk in the snapshots folder of your exports disk, /exports/snapshots on a default install. An S3 exports disk passes on its bucket and credentials.
You can point the disk somewhere else, or make another disk the default for snapshots, like any other disk (see Storage). Storing them away from the database volume, on another drive or in a bucket, means losing one doesn't lose the other.
Each snapshot is named after the time it was made, for example 2026-10-27_04-30-00.sql.gz.
Create a snapshot now
lpod stry artisan app:snapshot
This uses the same disk, compression and number to keep as the scheduled snapshots, so it can also delete the oldest one. Take one before upgrading.
Restore a snapshot
Loading a snapshot drops every table in the database first, then loads the snapshot. Everything changed after the snapshot was made is lost.
# List the snapshots, newest first
lpod stry artisan snapshot:list
# Load one by name, or the newest with --latest. --force skips the confirmation in production
lpod stry artisan snapshot:load 2026-10-27_04-30-00 --force
lpod stry artisan snapshot:load --latest --force
# Then migrate, in case the snapshot is from an older release, and restart the workers
lpod stry artisan app:update
The snapshot:* commands read the disk named snapshots from config/db-snapshots.php. If you made a disk with another name the default for snapshots, set that name in the config before restoring.
See also
- Application Configuration: managing the storage disks
- CLI Interaction: stry's Artisan commands
- Upgrading: updating an existing install