A backup captures the durable state needed to reconstitute the deployment:
the application database, the durable crypto (the master key and secret-box key
that wrap your data-encryption keys), and appliance configuration. Ephemeral
infrastructure secrets are regenerated on restore, which is what lets one bundle
restore cleanly onto a different box.
What’s protected
- Application database — tenants, org structure, rooms, file metadata, audit log, policies.
- Durable key material — the master key and secret-box key (envelope-wrapped), so restored data remains decryptable with your KMS/keys.
- Appliance configuration — the settings needed to bring the platform back up.
Create a backup
- Admin UI
- Console
Admin → Backup & Restore → Create backup. The bundle is produced on the appliance and listed for download.
enclave-backup-<timestamp>.tar.age. On a cluster, backups run on the primary (a replica reports that the primary’s schedule already covers it).
Schedule automatic backups
- Admin UI
- Console
Admin → Backup & Restore → Schedule — enable and pick a cadence.
Verify a backup
Always confirm a backup is intact and restorable before you rely on it:- Admin UI
- Console
Admin → Backup & Restore → the backup’s Verify action.
Copy backups off-box
Keep copies off the appliance so a lost box doesn’t lose the backups. Configure an off-box target (SCP or a mounted directory) and test reachability:- Admin UI
- Console
Admin → Backup & Restore → configure the off-box destination, then Test the connection.
Restore
Restoring overwrites the target appliance’s data with the backup’s contents, then reboots.1
Start from a clean target
Deploy a fresh appliance (or factory-reset an existing one). Complete just enough of first-run to reach the admin UI.
2
Upload the backup + identity
Admin → Backup & Restore → Restore. Upload the
.tar.age bundle and, when prompted, the age identity (private key) used to encrypt it. The identity is used only for this restore and is shredded immediately afterward — it is never stored at rest.3
Confirm and reboot
Confirm the overwrite. The appliance restores the database + durable keys, regenerates ephemeral infrastructure secrets, and reboots. A pre-restore safety snapshot is taken first, so a mistaken restore is recoverable.
4
Reconnect storage and KMS
Point the restored appliance at the same object storage and KMS as the original so it reconnects to your encrypted file content and key material.
On a cluster, restore is a standalone operation. Restore onto a fresh
single appliance and then re-form the cluster from it, or roll a node back to
standalone first. Do not restore into a live cluster member — see
High availability & clustering.
Recovery model at a glance
Next steps
Key Management
The KMS the restored appliance must reconnect to for file decryption.
Storage
The object storage that holds your encrypted file content.