Problem solve Get help with specific problems with your technologies, process and projects.

Windows Server 2012 Hyper-V storage: How Microsoft can fix the issues

Despite improvements, the Windows Server 2012 Hyper-V storage architecture still has holes. One expert details how Microsoft can fix the issues.

Microsoft revamped the Windows Server 2012 Hyper-V storage architecture, but there is still some room for improvement.

Storage has always been a weakness for Hyper-V. The process of shrinking, expanding, migrating and backing up virtual machines (VMs) creates a lot of difficulties in Hyper-V environments. But the new live migration and storage migration enhancements in Windows Server 2012 Hyper-V, including the seamless movement of VMs and virtual hard disks between physical storage volumes without downtime, could solve many of the lingering issues.

More on Hyper-V storage

Key approaches to adding Hyper-V storage disks

New Hyper-V Replica a boon for budget-conscious IT shops

The methods proposed below would greatly reduce downtime and simplify Hyper-V management by eliminating several manual and tedious tasks. 

1. Transition between virtual hard disk formats without downtime

For legacy Virtual Hard Disk (VHD) users, it makes sense to migrate to the new VHDX format. After all, it adds greater scalability, performance and resistance to corruption. The problem is that in Windows Server 2012 Hyper-V you need to shut down the VM to make the conversion. And for the large VHDs, the conversion can take a considerable amount of time, which can lead to extended downtime.

Here's one way Microsoft could have minimized the downtime associated with converting VHDs to the VHDX format:

How should this work?

  1. In the Disk Edit wizard, administrators can select Convert when the VM is running. (Currently this option is greyed out.) In addition, it should be possible to choose the options that are available when converting a VHD to the VHDX format or vice versa.
  2. Next, the process would trigger a Hyper-V Writer-assisted snapshot, and a new virtual hard disk in the new format is placed in the same location as the source virtual hard drive while the VM is running.
  3. Next, Hyper-V would take another VM snapshot and move remaining changes to the newly built VHDX.
  4. Hyper-V would then bring the new VM online in the same state with the loss of one ping, similar to what occurs during a Windows Server 2012 Hyper-V Live Migration between nonclustered hosts or Live Storage Migration.

Benefits of this method
This capability would allow for a seamless migration to the new and beneficial VHDX format without having to schedule downtime for all your legacy Hyper-V VMs.

2. Change the virtual hard disk type without downtime

With the continued performance increases of dynamically expanding disks, it makes sense that certain workloads (e.g., test/dev servers and some production VMs), could use this type of virtual hard disk.  Migrating workloads away from fixed-sized disks could free up significant physical disk space or correct a VM mistakenly provisioned with the wrong disk type -- and, by extension, save money. 

Again, the problem is, you need to schedule downtime to switch between fixed and dynamically expanding disks and vice versa. If the goal is to gain efficiencies within your Windows Server 2012 Hyper-V storage architecture, application owners have very little incentive to approve downtime in a timely fashion. If downtime was drastically reduced, however, you could avoid this situation altogether.

How should this work?

  1. In the Disk Edit wizard, Microsoft enables the Convert option when the VM is running. (It is currently unavailable.)  At this stage, users can choose which type of disk -- dynamic or fixed -- without having to change the location of the new disk.
  2. The process performs a Hyper-V Writer-assisted snapshot. From there, a new virtual hard drive is created in the new type at the same location as the source virtual hard drive, while the VM is running.
  3. Next, another VM snapshot is taken and the remaining bits are moved to the newly built dynamic or fixed virtual hard drive
  4. The VM is brought back online in the same state with the loss of one ping, similar to what occurs during a Windows Server 2012 Hyper-V Live Migration between nonclustered hosts or Live Storage Migration.

Benefits of this method
An organization saves money because of a better utilization of existing storage resources on volumes that house VMs. And administrators can easily transition between disk types, which will reduce downtime and effort.

3. Shrink and expand VM storage on the fly

Expanding VM storage without downtime would save a considerable amount of time and greatly reduce administrative complexity over the current process. At the same time, it would also be beneficial if you could shrink VM disks.

Currently, the P2V process in System Center Virtual Machine Manager can expand, but not shrink, a VM's hard drive. So if your physical server is using 40 GB of a 300 GB hard drive, you would still need to provision a 300 GB virtual hard drive, unless you provision a dynamic disk. But if this is a production workload, you would most likely use a fixed disk for better performance. In that case, you would utilize 300 GB of your physical storage -- most likely on a SAN, which has a high cost per gigabyte. 

How should this work?

With improvements to Live Migration and Live Storage Migration, which allows for the movement of VMs to an alternate host or the migration of VM storage to another disk volume, it would seem that Microsoft should expand this technology to something similar to what is shown below.

  1. Choosing Expand under Disk Edit allows for an increase or decrease in disk size based on the limits imposed by the amount of data on the virtual hard drive or its format (i.e., VHDs are limited to 2 TB).
  2. The process triggers a Hyper-V Writer-assisted snapshot, which creates a new, properly sized virtual hard drive at the same location while the VM is running and all the data is moved.
  3. While the VM is still running, it moves the remaining changes from the VM. Next, the VM is brought online using the newly sized virtual hard drive in the same state with the loss of one ping, similar to what occurs during a Windows Server 2012 Hyper-V Live Migration between nonclustered hosts or Live Storage Migration.

Benefits of this method
Many administrators tend to overprovision disk resources to avoid the need for future expansions. The ability to expand and shrink VM disks on the fly would greatly reduce this practice. As a result, companies would better utilize their storage resources through right-sizing their virtual hard disks, which ultimately saves them money.

This method would also decrease the amount of time needed to perform these tasks manually. It would also reduce the use of third-party imaging software, partition managers and creative workarounds -- such as using the Disk Administrator console -- to shrink virtual hard disks.

In the end, resolving these Windows Server 2012 Hyper-V storage shortcomings would reduce the amount of time IT pros spend coordinating downtime, allowing them to focus on optimizing management and performance.

Dig Deeper on Microsoft Hyper-V and Virtual Server