How do you perform a P2V with little downtime for large workloads?


Since I got my big bad Dell 8500 with 32GB of RAM I have excess capacity and decided to P2V my Windows Home Server.  After cleaning up the disk space on the WHS I ended up needing to “only” convert a 1.1 TB physical machine.  This reminded me of my time leading a project to migrate 500 physical workloads from an enterprise data center to a hosted shared environment.  If you’ve done virtualization for any period of time you’ve probably taken advantage of Insert your favorite Hypervisor Vendor tool here to do basic P2V’s.  Replicating this much data within a single data center can be a challenge but, how about migrating 500 workloads from one data center to another and needing limited down time?

My team used 2 methods

  • SAN to SAN replication
  • Software replication using Novell Platespin Migrate

SAN to SAN replication is nice but expensive. The provider I was doing the work for and the customer used disparate SAN technologies.  This isn’t a big deal today as most SAN vendors give your replication tools to migrate data from one SAN to another.  The SAN replication is licensed by the amount of data being replicated.  This creates a financial barrier to doing an all SAN approach.  Another challenge is that the customer wasn’t leveraging SAN to do boot from disk.  So we needed a solution for the local OS.  This is where PlateSpin came into play.

PlateSpin Migrate is a pretty slick product.  Platespin was an independent company that got purchased by Novell.  Novell of course got purchased by NetIQ.  Products like Platespin is why there are free tools from hypervisor vendors.  The primary features we leveraged from PlateSpin were Scheduling and Replication.  Platespin is like VMware Converter merged with backup software.  It allows you to create replication jobs similar to how you’d create a backup job.  So a perfect use case would be my 1.1 TB migration I’m attempting with my home server.  I could configure a job to replicate the local OS volumes to my hypervisor.  Once an initial replication is done and it’s time to cut the physical server over to the virtual, I can configure a job that sync’s the delta between the last replication and current data set.  This meant that most workloads only required little downtime to migrate and clean up the virtual machine. An added bonus is that Platspin allowed you to disable common hardware specific services that break the OS in a virtualized environment.

Of course when you migrating 500 workloads there are going to be some workloads that aren’t suitable to being virtualized.  Platespin gives you the ability to do Physical to Physical migrations as well as Virtual to Physical migrations or Virtual to Virtual.

My main complaint with Platespin was cost and stability.  My team spent at least 100 hours on the phone with Novell support working through random issues that were sometimes Platespin related and sometimes VMware or hardware vendor specific.

What tools have you used to do large scale P2V migrations?

Published by Keith Townsend

Now I'm @CTOAdvisor

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

%d bloggers like this: