Recently, as I was perusing some forums, I ran across a post outlining a user’s Hyper-V setup. In his post, the person outlined his existing Hyper-V setup in which he stated that his company uses RAID 0 for the Hyper-V system because it gives the best performance. However, use of RAID 0 carries with it massive risks. Despite the fact that “0” is a “RAID” level and the acronym RAID includes the work “redundant” when it comes to RAID 0, nothing could be further from the truth. With RAID 0, all data is striped across all of the disks that make up the RAID set. Because multiple disks are used, there is, indeed, a performance increase. However, there is also exactly zero data protection in place. The loss of a single drive will result in the loss of all data on the RAID 0 set.
In the title, I indicated that you should almost never use RAID 0 in production. I try my very best to avoid definitives, but I can’t think of too many situations that would force me to assume significantly increased risk for a bit of additional performance.
Hi Scott, Generally, I would agree with your recommendation but I wouldn’t exclude the use of RAID 0 in some specific situations where performance beats redundancy. For instance, you might find a scenario where hosted VMs are identical and do not store data, and the system configuration never changes. A backup image of the VM and Hyper-V may be available to restore the whole setup if a drive fails! The company may live with a recovery process of a couple of hours but may need the best performance of their VMs when they are running! I can’t recall any specific situation but I guess these kinds of situations may exist.