Table of Contents >> Section IV: RAID and Spanned Volumes

Chapter 14: Common Causes of RAID Failure

In most cases data is still recoverable in a RAID failure if the user takes immediate actions to write-protect the disks.

Hard Disk Failure

RAID 0 has no protection against hard disk failure. The failed disk must be physically repaired to recover data.

RAID 5 can continue to function in degraded mode after one disk has failed. The user can replace the bad disk with a new disk and perform a rebuild to regenerate data on the failed disk using parity data on the remaining disks. When the rebuild is complete, the RAID is back to normal working order.

The following can go wrong during a rebuild:
  • The user may inadvertently cause the loss of the original RAID configuration.
  • The user may inadvertently initialize the RAID. The initialization may permanently write over all data. It usually takes several hours to complete.
  • A good disk may be mistakenly replaced instead of the failed disk.
  • The rebuild may be performed with an incorrect number of disks. For example, a RAID may be configured with a "hot spare" which is mistakenly included in the rebuild.
  • The rebuild may be interrupted by an external event such as power loss, human errors, etc.
  • The rebuild may be performed with the component disks out of the original order.
  • The rebuild may be performed with a different stripe (block) size.
  • A second disk may fail before the rebuild is complete.

Copyright© 2013 QueTek™ Consulting Corporation, Houston, Texas, USA. All rights reserved.
Home