Backups

Open Backups in the admin sidebar. The page covers two kinds of backup:

• Panel backup: a nightly snapshot of HostCraft itself (the panel database, PowerDNS database and configuration), so you can rebuild the panel after a disaster.
• Account backups: restic repositories and scheduled jobs that back up customer files and databases.

Every backup uses restic, so snapshots are encrypted and deduplicated.

Who can use it: server admins and super admins, for any account. Account owners manage their own account's repositories and jobs from Backups in their panel (see Backups), and resellers can view their pool's snapshots.

Panel backup

The Panel backup card shows the health of the nightly snapshot that the hostcraft-panel-backup timer takes of the panel database, the PowerDNS database, /etc/hostcraft and the config snapshot store.

Badge	Meaning
Healthy	The last run succeeded on schedule.
Waiting for first run	Configured but has not run yet.
Overdue	The last success is older than expected. Check the timer on the server.
Last run failed	The error is shown on the card.
Not configured	No repository is set.

When it has run, the card shows Last run, Last success (with the snapshot ID), Size and Repository.

To configure it, set PANEL_BACKUP_REPOSITORY and PANEL_BACKUP_PASSWORD in /etc/hostcraft/panel-backup.env on the server, then run hostcraftd backup-panel once as root. To restore the panel from it, use hostcraftd restore-panel.

Caution

If the repository is a path on this server's own disk, the card warns you. Point PANEL_BACKUP_REPOSITORY at an off-server target (S3, B2 or SFTP) so a disk failure cannot take the backup with it.

Repositories

A repository is the destination a backup job writes to. Each repository belongs to one account.

The repositories table

Column	Meaning
Repository	Name and location.
Backend	S3 compatible, Backblaze B2, SFTP or Local disk.
Encryption	set when an encryption password is stored.
Retention	Days to keep and how many recent snapshots to keep.
Snapshots and Size	What is stored in it now.

Create a repository

1. Click New repository.

2. Account: the account the repository belongs to.

3. Name: for example offsite-s3.

4. Backend and Location:

   Backend	Location format	Credentials
   S3 compatible	s3.amazonaws.com/bucket/path or https://s3.example.com/bucket/path (works with Cloudflare R2, Wasabi, MinIO)	Backend credentials (KEY=value per line): AWS_ACCESS_KEY_ID=... and AWS_SECRET_ACCESS_KEY=...
   Backblaze B2	bucket:path	B2_ACCOUNT_ID=... and B2_ACCOUNT_KEY=...
   SFTP	user@host:/path	SSH private key and optional SSH port (default 22). The key is stored encrypted; the server's host key is trusted on first connection and pinned after that.
   Local disk	A path on this server, for example /var/backups/restic	None

5. Encryption password: encrypts every snapshot. Store it in a password manager. Without it the backups cannot be read.

6. Retention days (default 30) and Keep last (default 7).

7. Click Create. HostCraft initialises the encrypted restic repository at the destination.

Repository actions

Open the repository's row menu:

• Stats and retention: shows Stored, Snapshots, Files and Dedup ratio, and lets you change the Name and retention: Days, Keep last, Weekly and Monthly. Click Save.
• Test connection: checks the destination is reachable and whether the repository is initialised.
• Prune now: removes snapshots outside the retention policy and reclaims space, after you confirm.
• Delete repository: removes the repository from HostCraft. A second prompt asks whether to also delete the remote restic data. Answer Cancel to keep the data at the destination.

Scheduled jobs

A job backs up chosen paths of an account to a repository on a schedule.

Create a job

1. Click New backup job.
2. Account, Name and Repository (only that account's repositories are listed; "No repository for this account" means you need to create one first).
3. Schedule (cron): a standard five field cron expression. The default 0 2 * * * runs daily at 02:00 server time.
4. Paths (one per line): relative to the account's home directory. . (the default) is the whole home. Absolute paths must stay inside the home.
5. Include databases: dumps every database the account owns into the snapshot as .hostcraft-backup/databases/NAME.sql. With this ticked, paths become optional.
6. Exclude patterns (one per line, optional): for example cache/ or *.log.
7. Click Create job.

Job actions

The Scheduled jobs table shows each job's schedule, repository, status (enabled or paused), last run outcome and next run. Open the row menu:

• Run now: start a backup immediately.
• Run history: a drawer listing each run with its outcome, files and bytes processed, the snapshot ID, and the full error text for failed runs (restic errors and database dump failures).
• Edit: change the name, schedule, paths, databases and excludes, then Save.
• Pause schedule / Resume schedule.
• Delete job. Existing snapshots stay in the repository.

Snapshots

Snapshot activity shows the number of snapshots, the total stored and a chart of snapshots per day over the last 14 days. Recent snapshots lists each snapshot with its short ID, Repository, Account, Taken, Size and Files.

Browse and restore a snapshot

1. Click a snapshot row or Open.
2. The drawer shows its size, file count, repository, time taken and paths. An includes database dumps badge appears when databases were included.
3. Browse the folders with the breadcrumb and folder names.
4. Restore into: a folder inside the account's home. It defaults to restored-<date>-<time> in the home. Files keep their full original path beneath this folder, so live files are never overwritten.
5. Either tick the files and folders you want and click Restore selected, or click Restore all for the whole snapshot.

After the restore, copy what you need from the restore folder into place, then delete the restore folder to free disk space.

To restore a database, restore .hostcraft-backup/databases/NAME.sql and import it from the account's Databases page or the terminal.

Forget a snapshot

In the drawer, click Forget, then Forget snapshot. The snapshot is removed from the repository. Run Prune now on the repository afterwards to reclaim the space.

Common backup problems

Symptom	Fix
Test connection says unreachable	Check the location, credentials and that the server can reach the destination (firewall, bucket region, SFTP port).
A job shows failure	Open Run history and read the error. Common causes: wrong credentials, a full destination, or a database dump that failed.
Create job stays greyed out	Choose an account and a repository first.
Snapshots use more space than expected	Lower the retention numbers in Stats and retention, then Prune now.
Panel backup shows Overdue	Check systemctl status hostcraft-panel-backup.timer on the server and run hostcraftd backup-panel by hand to see the error.