Migrate Your Physical Workloads to IBM Cloud VPC Using P2V Solutions

3 min read

The RackWare Management Module (RMM) migration appliance simplifies the job of migrating your workload from a physical server to an IBM Cloud VPC virtual server instance (VSI).

The RackWare Management Module (RMM) is at the heart of the P2V migration solution that we use to orchestrate and resolve many challenges today in moving your workloads to IBM Cloud VPC. RMM is hardware agnostic for x86 workloads that run in your IBM classic infrastructure source environment:

P2V diagram

How the solution works

Provision the migration appliance

All you do to provision the RackWare Management Module (RMM) is access the provisioner tile from the catalog offering and launch RMM after providing a few details. Once it launches the RMM virtual server instance (VSI), IBM helps you obtain a valid license by running the discovery tool script from the RMM server. With the initial three licenses free for your consumption — sponsored by IBM and valid for three months — you are quickly up and running, with your source physical compute nodes identified in the RMM GUI under the Waves summary of the Replication option.

Use the following option with the discovery tool to obtain the promotional license for the RMM server:

#/opt/ibmtool/tool/discoverTool -p
Once RackWare is licensed, access the RMM GUI that looks like this.

Once RackWare is licensed, access the RMM GUI that looks like this.

Set up source and target environments in RMM

Prior to migration, you need to provision the target VSI with the same OS major version as the source server. The CPU and memory can be different from the source, but the disk space needs to be the same or greater than the source.

Next, you need to create what RackWare calls waves, which are essentially a list of source servers that you want to migrate. This can be done through the discovery tool with the “-d” option:

#/opt/ibmtool/tool/discoverTool -d

The script will find all the bare metal servers in your IBM Classic infrastructure account, except for servers that are running VMware or a custom OS like Citrix, Vyatta and AT&T.

Once you have identified your source server and added the target VSI post discovery, you can easily configure them to migrate your server data over to VSI in a few steps using the RMM GUI. This includes the ability to right-size the target volumes or partitions. What right-sizing means is that it allows achieving filesystem syncs by enabling custom sizes for the volumes or partitions on the target, based on volume or partition data size on source:

Here, a centos8 source from the classic environment and target at VPC have been synced.

Here, a centos8 source from the classic environment and target at VPC have been synced.

RackWare requires SSH Secure Access as a prerequisite to connect to source and target servers in the migration flow, which means copying the RMM public SSH key to both source and target to enable key-based SSH access. Once it has established access at both ends of migration pair, we are all set to trigger a migration job.

RMM: Under the hood

There are three major phases that RackWare RMM performs:

  1. Discovering and analyzing the source configurations (CPU cores, RAM, disks, partitions)
  2. Prepping the target VSI
  3. Copying the data from source to target

First, RMM discovers and analyzes the source server while the source server is running without any disruption to the production environment. Then it deploys a Micro-Kernel onto the target VSI that inserts itself in the Bootloader options of the platform OS and boots from it. The Micro-Kernel can be considered as a LiveCD for Windows or Linux platform that is being migrated. It essentially houses the required drivers that will be needed to correctly boot the compute node on IBM Cloud VPC.

Once the target VSI boots up from the Micro-Kernel, it starts communicating with the RMM server that does the remaining part of prepping the target server disks as per the configuration that the RMM discovered from source server. This is followed by taking LVM (Linux) and VSS (Windows) snapshots of partitions or volumes on the production source host before starting the filesystems sync on target.

Finally, RMM prepares the target system to boot back to the OS on completion of data migration and releases the target for the end user access and verification, as a last step.

Cutover

Perform a final sync when you are ready to switch over to the target server. This requires you to perform a final sync for applications that essentially syncs only the delta in data post for the last successful file-systems sync. It is recommended to quiesce the applications on the source host before the final delta sync.

RMM comes with loaded software to perform the migration without any worry about complexity in the Hardware Abstraction Layer, as the Micro-Kernel addresses a rich variety of drivers that are needed to make the compute node boot seamlessly in IBM Cloud VPC. It does all this while not forcing downtime in any phase of migration and being hardware agnostic for x86 workloads.

Get started

Be the first to hear about news, product updates, and innovation from IBM Cloud