Making Virtual Server Recovery-In-Place Viable

Generically referred to as “”, this feature gives administrators the option to point a to the data residing on a disk partition (typically a appliance) so that a failed VM can be more quickly recovered. The idea is to use the data as a temporary boot area until a full data restore can be completed on to a primary storage resource.

Storage IO Mismatch

On its face, this looks like a great idea – backup storage also facilitating near instant VM recoveries. But the challenge is most disk backup appliances are tuned to handle backup workloads, not the random IO that production applications produce. The other issue is that most disk backup appliances are designed for storage capacity efficiency. They are often configured with high density (4 TB) drives that run at slow (5400 RPMs) speeds. Between these two factors, the added latency could make for very poor application response times.

Lastly, many disk backup appliances are now using some form of data deduplication. While this is a great way to efficiently store backup data on disk, if deduplication processes are running while production data is mounted on a disk backup appliance, storage IO throughput may come to a grinding halt. Especially when the above two factors are also in play.

To read the complete article, CLICK HERE

NOTE: This column was originally published in the Weekly Newsletter.

Leave a Reply