Choosing the right recovery tool for Hyper-V
 Microsoft® Hyper-V may be a new release, but it’s certainly a burgeoning platform. There are already a lot of great tools for Hyper-V recoverability on the market, many with great track records for providing the functionality that matters most. To choose the right tool just ask around, but above all test those claims thoroughly.
Stress testing is particularly important. Many workloads are left in inconsistent or non-bootable states because either their OS or application state wasn’t properly protected. Think of it like a partially written transaction that can’t be replayed at recovery because the order of operations wasn't properly preserved – write-order integrity and state preservation matters. The same rules apply to application workloads under Hyper-V.
Hyper-V does a great job of preserving state through its use of native VSS functionality on the host and extending it to the guest workloads. It's not real-time protection, but if you have a database workload running and trigger a backup from the host, it will pass the VSS request to the guest and then on to the application. This guarantees that the database can preserve its state to disk, followed by the guest saving its state correctly, thus providing a good point-in-time to backup. But what if we want better recoverability than a snapshot every hour? Here’s where Microsoft’s true strength enters - third-party partners; but you need to do your homework.
Many of the tools on the market can’t adequately protect many workloads from the host layer, due to architectural limitations. This market is still too new and many vendors are too inexperienced so they’re competing solely on price. Many vendors provide volume pricing for virtual guest workloads that ease the total cost of ownership issues. Just make sure that you fully test and know exactly what you’re getting for your money. Don’t be afraid to challenge vendor claims either, you’re the customer, remember?
Nicholas Schoonover,
Senior Solutions Architect, Double-Take Software
|
|
Speed and Adaptability
Two criteria count above all others when looking at disaster recovery solutions in a Hyper-V environment: solution adaptability and migration time. You’ll find out right away if you need to bother doing functional testing for each and every vendor. After all, if the product can’t perform on your systems and your load comfortably under stress, how can you ever trust it in production when you’ll need it the most?
Certain solutions do not support P2V (physical to virtual) migration in real time. You may need to lock out users for several hours (or even days) to complete migration following an incident. There are vendors that make tools that don’t require you to stop your applications and still provide a real-time crash consistent data image for recoverability. Additional features such as baseline synchronization help reduce migration time.
Vendors make all sorts of claims and it’s your responsibility to test them. Just because a product supports IP doesn’t mean that it’s going to provide crash-consistency and scale. Thorough testing is never more important than for site-to-site solutions. You should also take into consideration that Microsoft made some significant changes to clustering in Windows Server 2008 that allows geoclustering. Some third-party vendors don’t even require SAN technology at all to implement MSCS. |
|
