IBM Cloud Docs
Classic bare metal to bare metal or virtual server on VPC by using a private network

Classic bare metal to bare metal or virtual server on VPC by using a private network

The RackWare Management Module (RMM) migration solution provides a seamless migration from a classic physical bare metal server to a bare metal or virtual server instance on VPC environment. Use the intuitive GUI to move the OS, application, and data from VMware ESXi in IBM Cloud classic to IBM Cloud VPC virtual server instances. The RackWare Management Module (RMM) solution simplifies the overall migration process of moving the operating system, applications, and data.

The migration can occur either over the public or private interface of the compute resource. The only requirement is that all three components are reachable from one and another. This guide focuses on the private path, which requires setting up a transit gateway for a communication channel between classic and VPC over the private interface.

Objectives

  • Prepare your source and target servers for migration.
  • Learn how to use the RMM solution.
  • Successfully complete a classic bare metal to VPC bare metal migration for private networks.

Architecture diagram

Physical to virtual migration private IP diagram.
Figure 1. Migrating classic bare metal to bare metal on VPC over private interface

Physical to virtual migration private IP diagram.
Figure 2. Migrating classic bare metal to virtual server instance on VPC over private interface

This architecture diagram is also applicable for classic bare metal to VPC bare metal.

Before you begin

  • Check for correct IBM Cloud VPC user permissions. Be sure that your user account has sufficient permissions to create and manage VPC resources. See the list of required permissions for VPC.
  • Understand the differences between classic and VPC infrastructures (the following list is not exhaustive):
    • VPC does not support shared volumes or file-based volumes
    • VPC does not support snapshot or replication
    • GPU is not supported in VPC
  • For more information, see Comparing IBM Cloud classic and VPC infrastructure environments.
  • Plan for your target server. For more information, see Planning for Bare Metal Servers on VPC.

Supported Operating Systems

  • For classic bare metal to VPC bare metal

    • Ubuntu 20.04

  • For bare metal to VPC virtual server instance

    • CentOS 7.8, 7.9

    • RHEL 7.2, 7.3, 7.4, 8.1

    • Ubuntu 18.04, 20.04

    • Debian 9.x, 10.x

    • Windows 2012, 2012R2, 2016, 2019

The Convert LVM feature is only supported for RHEL 7.x and RHEL 8.x

Order RMM

The RMM tool is available in the IBM Cloud catalog. After you order, a virtual server with RMM software is installed into your VPC of choice. The RMM server has a public IP address for reachability and a default login.

If public IP address is not attached to RMM server, then its 'Reserved IP' address can be used to access RMM server with bastion host.

  1. Order the RMM server from the IBM Cloud catalog tile.
  2. After you order, log in to the RMM server.
  3. In the RMM server, change the default password, create users, and create an SSH key with the correct name.

Set up and provision a VPC, bare metal server, or virtual server instance

Target bare metal server can be provisioned only manually. You can set up the target virtual server manually or with the RMM auto-provision feature.

Option 1: Manual

The RMM solution handles only the OS, application, and data movement. It does not to set up a VPC on the target side. Therefore, you need to set up the IBM Cloud VPC infrastructure. At a bare minimum, you need to set up a VPC, subnets, and bare metal or virtual server instances. For more information, see the Virtual Private Cloud (VPC) documentation.

  1. Create a VPC.

  2. Create subnets.

  3. Order a bare metal or virtual server instance.

    • SSH key
    • Operating system name (Linux or Windows and their respective version)
    • Security groups
    • Secondary volume (optional)

    Encrypted volumes are not supported.

Option 2: Auto-provision

Setting up a cloud user

  1. Log in to the RackWare web console.
  2. In the RackWare web console, go to Configuration > Clouduser.
  3. When you add a cloud user, enter a name and select IBM Cloud VPC for the Cloud Provider. Select the region where you want to auto-provision the virtual server instance, and enter your IBM Cloud API key.
  4. Click Add.

Creating a wave and replication

A wave contains a single host or multiple hosts that are migrated. For this migration, you need to create one or more waves, provide information about the hosts in the wave, and then start the wave.

  1. In the RackWare web console, go to Replication > Waves.
  2. When you create a wave, select Target Type as Autoprovision.
  3. Enter source and target details.

If the source server has a Linux® OS and it has a boot volume greater than 100 GB, then leave the ‘Provision disk’ textbox empty and select the Convert to LVM option. This selection creates an extra disk of the required size and converts all eligible volumes to LVM on the target. For Windows, no action is required. It is taken care automatically by RMM.

  1. After you enter your source and target information, you need to provide your IBM Cloud VPC information.
  2. From the edit option in the Actions menu of your source, select the IBM Cloud VPC Options tab, enter the relevant information, and click Modify.
  3. Run the replication.

Ensure that your VPC, subnet, and other necessary cloud components are set up before you add a cloud user in RMM.

Assigning the environment to wave

  1. In the RackWare web console, go to Replication > Waves.
  2. Select the wave that needs to be migrated.
  3. On the Wave Detail page, select the Autoprovision option as Not configured.
  4. Select your cloud user for the Environment, enter the region where the virtual server instance needs to be provisioned, and apply the changes.

Auto-provision feature is not available if target is VPC bare metal. The user needs to select Existing System as the target type in this case.

Prepare source and target servers

There are a few things that need to be done on the source and target server for the migration to work. The RMM server needs to SSH into the servers; thus, the RMM public SSH keys need to be copied onto both the source and target servers. If the source server has both public and private interfaces, host routes need to be added to direct the communication between the servers over the transit gateway path. Complete the following steps to prepare your relevant servers.

Linux systems

  1. Copy the RMM SSH public key to both the source and target servers.
  2. If your compute resource has both public and private IP addresses, the host level route needs to be added for it to work properly. Run the following command on your classic compute resources for your operating system:
ip route add <destination_network> via <Gateway_address> dev <private_ethernet_interface>

Windows systems

  1. Copy the RMM SSH public key to both the source and target servers.
  2. You need to download the SSH key utility. You can download it from the RMM server.
  3. The user is SYSTEM, and you need to key in the RMM SSH key to authenticate for both the source and target servers.
  4. If your compute resource has both public and private IP addresses, the host level route needs to be added for it to work properly. Run the following command on your classic compute resources for your operating system:
route ADD <destination_network> MASK <subnet_mask> <gateway_ip> <metric_cost>

If you use the auto-provision feature, you don't need to set up a target. Only the friendly name for the target server is required.

Set up RMM waves

You can migrate the servers over one by one or do simultaneous migrations. If you are doing multiple, simultaneous migrations, download the CSV template from the RMM server and complete the appropriate fields.

  1. Log in to the RMM server.
  2. Create a Wave and define the Wave name.
  3. If multiple hosts are to be migrated, download the template, complete the appropriate fields, and then upload the template.
  4. Select the Wave name to enter source and target information.
  5. Select the "+" sign.
  6. Add source IP address or FQDN and add source username.
  7. Target Type = Existing system
  8. Sync Type = Direct sync
  9. Add a target IP address or FQDN.
  10. Add a target-friendly name, and add a target username.
  11. Start the migration.

The username field for Linux environments is root. The username field for Windows environments is SYSTEM.

Within the discovery script, a helper script is provided to help with the discovery of the classic bare metals and creating the waves for steps 2 and 3. The script asks for your classic username and API to discover your classic bare metals.

/opt/IBM/discoverTool -d

For more information about the discovery tool, click here.

Order a IBM Cloud® Transit Gateway

IBM Cloud Transit Gateway provides connectivity between classic and your VPC infrastructure. When you connect the two entities, be aware that the classic infrastructure is on the 10.0.0.0 network, which means that the VPC network needs to be on different network; otherwise, the networks overlap and cause communication issues.

  1. Order IBM Cloud Transit Gateway
    • Use a local routing option
  2. Add connections
    • Classic infrastructure
    • VPC
      • Select region
      • Select VPC

Validate your migration

Before you decommission the source server, you need to validate the target server. The following list is not exhaustive, but suggests some of the items to validate:

  • Application
  • Licensing
  • Reachability (host-level configuration changes)
  • Remove RMM SSH key