RAID Data Recovery FAQ

Frequently Asked Questions Concerning Mirrored,

Striped, & Other RAID Volumes

A: The term RAID is short for a “Redundant Array of Inexpensive Disks” (or a “Redundant Array of Independent Disks”). A RAID is a collection of disk drives that are grouped together and accessed based on a pre-set configuration to provide data mirroring, striping or redundancy, or some combination thereof.

A: Hardware RAID utilizes a special physical controller that maintains information about the hard drives in the volume. Software RAID (for example, in Windows 2003 Server) is an implementation of RAID wherein the normal drive controller is used, and software manages the drives in the array.

A: RAID levels represent different configurations providing varying amounts of speed and fault tolerance. RAID levels currently defined are RAID 0, 1, 2, 3, 4, 5, 6, 10, 50 (or RAID 5+0), 100 (or RAID 10+0), and RAID 0+1. The most widely used are RAID levels 0, 1, and 5.

A: In a RAID level 0 (or disk striping) configuration, data is written evenly across 2 or more drives. This utilizes all of the storage capacity of each drive in the array, but it provides zero fault tolerance. If one drive fails, all of the data within the volume is lost. With RAID level 1 (or disk mirroring), data is written to a primary and a “mirror” drive at the same time. In theory, if one drive fails, the other drive can be installed as the primary drive with no loss of data. This implementation requires a minimum of 2 drives of equal capacity. If one drive has less capacity, the controller will default both drives to the lower capacity.

A: Yes – under what is referred to as RAID 0+1. This configuration combines disk striping (RAID 0) across 2 or more drives and simultaneously “mirrors” the data to a duplicate set of drives (as with RAID 1). This provides fault tolerance, but at a fairly high cost in terms of overhead. A RAID 0+1 volume requires a minimum of 4 hard drives.

A: In a RAID 5 configuration, additional data is written to the disk that should allow the volume to be rebuilt in the event that a single drive fails. In the event that a single drive does fail, the volume continues to operate in a “degraded” state (no fault tolerance). Once the failed drive is replaced with a new hard drive (of the same or higher capacity), the “parity data” is used to rebuild the contents of the failed drive on the new one.

A: “Just a Bunch Of Disks”, or JBOD, is not really a RAID configuration. The term refers to hard drives that are concatenated (combined or “spanned”) to create a large volume. A JBOD configuration provides no real performance benefits or fault tolerance, but it does allow for total usage of all the available space on each disk.

A: This is determined by the RAID level being used. For example, RAID level 0 (disk striping) and RAID level 1 (disk mirroring) each require at least 2 hard disk drives. A RAID 5 volume, on the other hand, requires a minimum of 3 drives of equal capacity to configure.

A: In most cases data can still be recovered, but it will depend on how the volume was re-formatted. A re-format through the operating system (Windows, for example) will re-establish the logical volume quickly, but the previous data will still physically reside on the disk. A controller-level format may only overwrite the array information and leave the original data intact (a relatively quick format), or it may overwrite every block on every drive in the array (a considerably longer, low-level process). It depends entirely on the manufacturer of the controller and the format utility used.

A: The safest approach to data recovery with a RAID volume (or with any media) is to capture every storage block on each device individually. The resulting drive “images” are then used to help rebuild the original array structure and recover the necessary files and folders. This approach limits continued interaction with the media and helps to preserve the integrity of the original device. One of the dangers in using data recovery software is that it forces the read / write heads to travel repeatedly over areas of the original media which, if physically damaged, could become further damaged and possibly unrecoverable.

A: If one drive fails in a RAID 5 configuration, the volume still operates – but in a degraded state (it no longer writes parity information). The important data should be backed up immediately and verified to be usable before any rebuild operation is started. When it comes to critical data, anything that is used to read or write to the original volume represents a risk. Is the hardware operating properly? Are all other drives in the volume functioning correctly? If you are the least bit unsure, a rebuild should not be performed.

A: In many cases, the answer is yes. It usually requires that data be recovered from each failed hard drive individually before attempting to address the rest of the volume.

A: We only need the original drives that comprised the array. In certain circumstances, it may be beneficial to have the original controller that was being used. However, sending the controller may not be possible or practical, such as when the RAID controller is incorporated into the system motherboard. It doesn’t hurt to include the controller with your shipment, but it isn’t necessary to do so for us to recover the data.