Disaster Recovery UCP

Ensure business continuity by learning how to back up and restore your Docker UCP deployment. This guide covers:

What to back up
CLI and UI backup methods
Step-by-step restore procedures
Key considerations and best practices

Components to Back Up

Protect the following UCP components to guarantee a smooth recovery:

Component	Description
UCP Configuration	Enterprise license, Client CA bundles, access control (teams, users, roles), certificates & keys
etcd Objects	Kubernetes declarative objects (Pods, Deployments, ReplicaSets, ConfigMaps, Secrets)
UCP Volumes	Named volumes managed by UCP
Monitoring Data	Metrics and logs collected by UCP

UCP does not back up Docker Swarm services, configs, or overlay networks. Use Docker Swarm tools for those workloads.

Backing Up UCP

You can perform backups via the CLI or the UCP web interface.

1. Backup with the CLI

Run the official UCP image in backup mode. Mount the Docker socket and a host directory to store the archive:

docker container run \
  --rm \
  --log-driver none \
  --name ucp-backup \
  --volume /var/run/docker.sock:/var/run/docker.sock \
  --volume /tmp:/backup \
  docker/ucp:3.2.5 backup \
    --file /backup/ucp-backup.tar \
    --passphrase "YourStrongPass" \
    --include-logs=false

--file specifies the output archive path
--passphrase encrypts the backup
--include-logs toggles container logs (default: true)

2. Backup via UCP Web UI

Log in as an admin.
Navigate to Admin Settings → Backup.
Click Start Backup, set a passphrase, and monitor progress.

Restoring UCP

Follow these two main steps to restore UCP from a backup:

Step 1. Uninstall Any Existing UCP

If UCP is already installed, remove it cleanly:

docker container run \
  --rm -it \
  --volume /var/run/docker.sock:/var/run/docker.sock \
  docker/ucp:3.2.5 uninstall

Step 2. Perform the Restore

Pipe the backup archive into the UCP restore command:

docker container run \
  --rm -i \
  --name ucp-restore \
  --volume /var/run/docker.sock:/var/run/docker.sock \
  --volume /tmp:/backup \
  docker/ucp:3.2.5 restore \
    --passphrase "YourStrongPass" < /backup/ucp-backup.tar

During restore you can choose to:

Reuse the same Swarm cluster
Provision a new Swarm cluster
Restore onto a standalone Docker host (UCP initializes Swarm automatically)

The image contains a list of notes about backup and restore processes related to Docker Enterprise and Swarm workloads. It highlights limitations and requirements for successful backups and restores.

Key Considerations

Topic	Details
Single Operation	Only one UCP backup or restore can run at a time.
Cluster Version	Restore targets must match the backup’s Docker Enterprise/UCP version.
Swarm Workloads	UCP backups exclude Docker Swarm services, configs, and networks.
Failed Cluster	You cannot back up a cluster that has already failed—prepare backups in advance.
Restore Destinations	Original Swarm cluster, new cluster, or standalone Docker host.

Always verify that your passphrase and backup archive are accessible before starting a restore.

The image contains notes about backup and restore processes for Docker Enterprise, highlighting limitations and requirements for swarm workloads and cluster versions.

Introduction

Docker Engine

Docker Image Management

Docker Engine - Security

Docker Engine - Networking

Docker Engine - Storage

Docker Compose

Docker Swarm

Kubernetes

Docker Engine Enterprise

Docker Trusted Registry