Managing and Troubleshooting RAID Volumes

Several things can go wrong with RAID volumes. Spanned and striped volumes are particularly vulnerable; as has already been mentioned, failure of any one disk in the volume renders the entire volume useless and data must be restored from backup. If one disk in a mirrored volume fails, you can break the mirror and use the data on the other disk as a simple volume. If one disk in a RAID-5 volume fails, the system reconstructs the missing data from the parity information and the volume is still usable, but without fault tolerance and with reduced performance until the failed disk is replaced. If more than one disk in a RAID-5 volume fails, the volume has failed and must be restored from backup after the disks have been replaced.

Besides the volume statuses already described for partitions on basic disks and simple volumes, Disk Management can display the following messages with RAID volumes:

• Resynching: Indicates that a mirrored volume is being reinitialized. This status is temporary and should change to Healthy within a few seconds.
• Data Not Redundant or Failed Redundancy: For a mirrored or RAID-5 volume, this status usually means that half of a mirrored volume was imported, or that half is unavailable, or that only part of the underlying disks of a RAID-5 volume were imported. You should import the missing disk(s) to re-create the volume. You can also break the mirror and retain the half that is functioning as a simple volume. If you have all but one of the underlying disks of a RAID-5 volume, you can re-create the RAID-5 volume by adding unallocated space of a different disk.
• Stale Data: This status is shown when you import a disk that contains a mirrored volume half, or a portion of a RAID-5 volume, with a status other than Healthy before it was moved. You can return the disk to the original PC and rescan the disk to fix the error.