Skip to main content

Restore a Server

Before restoring​

  • Server: Fresh host with Docker, SSH, and enough disk for the archive and extracted data
  • Firewall: SSH, 80/tcp, 443/tcp, and 3000/tcp open during setup
  • Clusters: Prepare trusted Swarm and registry ports
  • Data: Read What Is Backed Up and obtain separate database and volume backups

Important: Place the archive at /captain/backup.tar before running the CapRover installer. The installer detects it at startup.

Restore​

  1. Copy the archive to the new server:

    ssh root@NEW_IP 'mkdir -p /captain'
    scp backup.tar root@NEW_IP:/captain/backup.tar

  2. Run CapRover on the new server:

    sudo docker run -p 80:80 -p 443:443 -p 3000:3000 \
      -e ACCEPTED_TERMS=true \
      -v /var/run/docker.sock:/var/run/docker.sock \
      -v /captain:/captain \
      caprover/caprover

    Need Docker? Follow Getting Started first.

  3. Watch the installer logs. For clusters, follow Multi-Node Restore when prompted for node mapping.

  4. Point the original wildcard DNS to the new server when ready. For a clone running alongside the original, choose a different root domain.

  5. Restore database files, volumes, and bind mounts separately. Redeploy apps if their images are missing.

  6. Verify dashboard access, HTTPS, services, and application data before fully switching traffic.

Keep the original server and backups until verification succeeds.

Multi-Node Restore​

Prepare the new manager and each replacement worker with Docker, network access, and a plan for the old nodes' persistent data. Follow Restore a Server on the new manager.

  • For a multi-node backup, the first installer run can exit intentionally and create /captain/restoring/restore-instructions.json.

  • Inspect its oldNodesForReference and nodesMapping.

  • For each old worker, replace the placeholder newIp with its replacement node IP.

  • Leave the manager's CURRENT_NODE_DONT_CHANGE mapping alone.

  • The file also supplies the SSH user and privateKeyPath for each worker; configure access and copy the required private key securely to the indicated path (by default /captain/id_rsa).

  • Rerun the same installation command on the new manager.

  • The installer checks SSH to workers and maps saved node placement to the replacements.

  • Remove the copied private key after restoration.

  • Restore each node's volumes and bind mounts separately, verify registry access and redeploy unavailable images, then inspect placement and services.

  • Never point production DNS at the new cluster until validation is complete.

Move CapRover to Another Server​

  • Use Create a Backup on the source and back up every application's volumes, databases, and uploaded files separately.

  • Record the root domain, DNS records, registry location, custom ports, and each app's node placement.

  • Provision the destination with Docker and the necessary ports. Restore the CapRover archive, or use Multi-Node Restore for a cluster.

  • Restore application data, confirm images are accessible or redeploy them, and test the dashboard and apps.

  • Change the wildcard and dashboard DNS records to the new address when ready; check HTTPS after DNS propagates.

  • Keep the old host available until the new one has passed a restore check.