NVIDIA UFM Enterprise User Manual
NVIDIA UFM Enterprise User Manual
Breadcrumbs

Migration

UFM Versions Manager provides migration capabilities to move UFM installations to new hardware for both Standalone (SA) and High Availability (HA) environments.

Migration Types:

  • Standalone Migration - Move UFM from one server to another (source unchanged)

  • HA Migration (RMA) - Replace standby node in HA cluster

Standalone Migration

Requirements

Source system:

  • UFM SA installation

  • SSH access to target

Target system:

  • Same or newer OS version

  • Podman rootless is installed and configured as instructed in the podman rootless install steps.

  • Sufficient disk space

Network:

  • SSH connectivity between source and target

Migrate SA to Remote Host

During the migration, UFM validates some config files' schema; in case of failure, the operation is stopped


  1. Prepare the target host with OS, Docker/Podman, and SSH.

  2. Preview migration. Run on source:

    ufm_versions_mgr restore --migrate-sa --target-ip 192.168.1.30 --dry-run


  3. Perform migration. Run on source: 

    ufm_versions_mgr restore --migrate-sa --target-ip 192.168.1.30

  4.  SSH to target and start UFM. Run:

    ssh root@192.168.1.30
systemctl start ufm-enterprise


  5. Verify UFM running on target. Run:

    systemctl status ufm-enterprise

Command Options

Option

Description

--migrate-sa

Standalone migration mode

--target-ip IP

Target host IP address

--dry-run

Preview migration

--verbose

Enable detailed output

  • The source UFM stays running and unchanged.

  • The target UFM is installed but does not start automatically.

High Availability Migration

Requirements

  • HA version 2.1.0 or higher

  • Tool run from master node

  • Hacluster password

  • New node with OS,Podman rootless, and SSH configured

Migrate HA Standby Node

  1. Prepare new node with OS, Docker/Podman, network configuration, and SSH.

  2. Preview migration. Run on master: 

    ufm_versions_mgr restore --migrate-ha \
    --target-ip 192.168.1.20 \
    --hacluster-password mypassword \
    --dry-run

  3. Perform migration. Run on master: 

    ufm_versions_mgr restore --migrate-ha \
    --target-ip 192.168.1.20 \
    --hacluster-password mypassword

  4.  Verify cluster status. Run: 

    ufm_ha_cluster status
     




Command Options

Option

Description

--migrate-ha

HA migration mode

--target-ip IP

Target node primary IP

--target-secondary-ip IP

Target node secondary IP (optional)

--hacluster-password PWD

Hacluster user password

--dry-run

Preview migration

--verbose

Enable detailed output

  • The migration must be performed from the master node.

  • The HA cluster is taken offline during the migration.

  • SSH trust is configured automatically.

Dual Network Configuration

For HA with management and backup networks. Run: 

ufm_versions_mgr restore --migrate-ha \
    --target-ip 192.168.1.20 \
    --target-secondary-ip 192.168.2.20 \
    --hacluster-password mypassword

Migration Process Details

SA Migration Steps

  1. Create temporary backup on source

  2. Establish SSH trust with target

  3. Check target (cleanup if UFM exists)

  4. Stream Docker images to target (parallel, compressed)

  5. Transfer configuration files

  6. Install UFM on target

  7. Cleanup temporary backup

HA Migration Steps

  1. Validate HA environment and version

  2. Check/cleanup target node

  3. Stop HA cluster

  4. Backup images

  5. Stream images to target

  6. Copy service file

  7. Execute HA RMA

  8. Restart cluster

Last updated: