In this article, Matt Hester focuses on building the Private Cloud foundation from a storage perspective.  As I’ve pointed out in previous posts, storage is a very important part of your Private Cloud infrastructure since that is where everything is stored. 

I am including a brief snippet from Matt’s post below, but you will want to read the full article by going directly to Matt’s blog.


When you start to think of building a private cloud into your datacenters, you might naturally think you need run out and buy a SAN.  While you still may want to go buy a SAN, with Windows Server 2012 you do not necessarily need to.  In Windows Server 2012 there has been a tremendous amount of improvement done to the storage capabilities.  Specifically you now can have continuously available file shares on your cluster servers.  This is mainly due to the new SMB 3.0 protocol built-in to the server, it really has made storage a first class citizen in Windows Server 2012.  So for todays post I am going to take a look at how to build a Windows Server 2012 storage environment using Storage Pools and inexpensive Serial-Attached SCSI (SAS) disks to provide the necessary storage backend for your private cloud infrastructure.  The best part is if you assigned the storage pool to your cluster you know have a continuous available share all built on fairly inexpensive disks.  Even better than that is that Windows Server 2012 Hyper-V can leverage the clustered share now you have a virtualized environment that is always at your fingertips.

Create a storage pool is actually fairly straight forward, after you have the drives connected to the system.  However, if you are going to build the storage out on a cluster there are some things you need to do that are unique to working with a clustered environment.  Before you begin the process you will want to review the list of prerequisites located here: Deploy Clustered Storage Spaces Prerequisites.

The first thing you will need to install the multipath I/O software (MPIO).  This feature allows for multiple physical paths to physical path as a single logical path for data access.  In other words if you have a lot of drives configured in a JBOD enclosure with multiple SAS connections to each host, MPIO allows the hosts to see each physical drive path via a single redundant path.  In a cluster, MPIO can be combined with
Storage Spaces and Cluster Shared Volumes ( CSV ) to treat these physical drives as a single highly-available shared volume.

To install MPIO via the Add Roles and Features wizard in Server Manager or you can run the following PowerShell script:

Install-WindowsFeature –ComputerName servername –Name Multipath-IO

You will also want to make sure sharing is working on your servers but a bulk of your work will come from making the storage spaces for your cluster to use.  Storage Spaces offers an easy way to combine your drives into one logical volume.  You can either create the storage spaces before you create the cluster or after creating the clustering.  Either way is effective, I am going to take you through a quick look at creating the storage spaces, you can find some really detailed steps here: Deploy Storage Spaces on a Stand-Alone Server.  Overall the procedure looks like this:

  1. Create your Storage Pool
  2. Create your Virtual Disk
  3. Create your Volume


Harold Wong