Help › Set up › B8
Update, back up and move your own relay to a new server
Updates for each installation way, the daily backup including the keys, restoring with a click, and what matters when you move to another server.
Preview this guide has not yet been clicked through on a real device. Button names and steps come from the program but may differ in details.
- For
- Own-relay operator
- Applies to
- Own relay
- Right in the console
- System administrator of the relay
- Control from
- Browser
What it is for
Depending on how it was installed, your own relay does not update itself, and its data belongs in a backup. This page shows how to update it for each installation way, how the backup works in the console, how to restore it, and what matters when you move to another server.
What you need
- Your own relay (B2) and a console account as system administrator.
- For Docker and Linux: access to the server's command line.
How to do it
Update
| Installed with | How to update |
|---|---|
| Docker | Run the update script in the installation folder. It loads the new image, verifies it and restarts; data/ stays. Guide: Update a Docker relay |
| Linux without Docker | With update-browserdesk-relay-linux.sh (next to the installation script). It backs up to /opt/browserdesk-relay.vorher, swaps the program files and checks /health. Guide: Install your own relay on Linux |
| Windows (MSI) | Install the new MSI over the old one. Database, relay.json and keys stay. Or in the console (see below). Guide: Install your own relay on Windows Server |
In the console: Settings → Relay update → Check for update, then Apply update. Whether the relay updates itself is set in Configuration → Self-update and mirror → Relay self-update. Leave self-update off on Linux installations that you update with the script.
Back up
- Settings → Maintenance schedule.
- Under Backup passphrase set a passphrase and Save. Without a passphrase the relay does not back up its keys.
- Add job → choose Backup (database + images) and set a schedule. Run now tests the job immediately.
- The backups are listed under Existing backups. Every database backup comes with an encrypted archive of the keys (
….nachweise.zip.enc). Download both and keep them outside the server.
Keep the value BD_KEYRING_KEY (with Docker in the .env file) separately and safely, for example in a password manager (B9).
Restore (same server)
- Settings → Maintenance schedule → Existing backups → row with the database backup (
….dump.gz) → Restore. - In the dialog Restore database? type the database name and click Restore.
- The relay restarts. Sign in again afterwards.
Move to a new server
- Install the relay on the new server the same way (B2), but do not start it yet, or stop it again right away.
- Decrypt the key archive with the backup passphrase and unpack it into the relay's data folder before the first start:
openssl enc -d -aes-256-cbc -pbkdf2 -in <name>.nachweise.zip.enc -out nachweise.zip - If
BD_KEYRING_KEYwas set, set the same value on the new server before the relay starts. - Start the relay, then restore the database backup (as above).
- Point the name (DNS) at the new server. If the devices keep the same address, you do not need to change anything on them. For a new address: C6.
- Licence: a licence is valid for one installation. Have the activation released at browserdesk.de (info@browserdesk.de) and install the licence file again on the new server (A5).
How you know it worked
- After an update, Relay status (or
/health) shows the new version. - After restoring, the console confirms that the database was restored, names the file holding the previous state, and says that the relay is restarting.
- After a move the devices sign in at the new server, and their device certificates remain valid.
Good to know
- Running sessions during an update: the viewer shows „Relay wird aktualisiert – Verbindung wird wiederhergestellt …“ (relay is being updated, reconnecting) and reconnects by itself. A direct session keeps running.
- Restoring resets everything: devices, accesses, users, tickets, settings. Whatever came later is gone. The state before lies as
…-vor-einspielen.dump.gzin the same list — that is how you undo it. - All console sign-ins end when restoring.
- Without the keys, all encrypted values are empty on a new server: a new device CA (all device certificates invalid), mail and sign-in credentials gone.
- Windows wraps the keys to the machine (without
BD_KEYRING_KEY). A move to another machine is then not possible without losing the encrypted values; there is no way to re-wrap yet. Write to info@browserdesk.de before such a move. - Linux without
BD_KEYRING_KEYstores the keys unencrypted and warns at start-up (B9). - Docker: the whole backup is the
data/folder plus the valueBD_KEYRING_KEYfrom.env. - Going back to the old version with Docker: set
RELAY_VERSIONin.envback and rundocker compose up -d, as long as the old image is still loaded.
If it does not work
| You see | Meaning and fix |
|---|---|
| The backup job reports „Nachweise (Schlüsselring, relay.json, Lizenz) NICHT gesichert — keine Backup-Passphrase …“ (keys not backed up, no backup passphrase) | No Backup passphrase is set. Set it and run the job again. |
| Restoring stops with “already exists” | Nothing was changed. The backup contains objects the relay does not know. Contact info@browserdesk.de. |
| Restoring stops after 30 seconds | A running request held a table. Nothing changed — try again. |
| After the move all devices report certificate errors | The keys did not come along, or BD_KEYRING_KEY is missing. Unpack the archive before the start and set the value. |
| The relay does not start after the move | The key ring is wrapped, but BD_KEYRING_KEY is missing. Set the same value as on the old server. |
| Licence expired on the new server | The licence is still bound to the old server. Ask for the release and install it again (A5). |
| Windows: the relay does not come back after restoring | The restart on Windows has not been tested yet. Restart the BrowserDeskRelay service. |