Home | Contents | Help   Raidhelp     Automatic Volume Rebuilding One of the most vital aspects of RAID functionality is reconstruction also referred to as rebuilding. This is the primary objective of storing parity information across an array. The rebuilding of the missing data is achieved by collecting the parity information stored on the working drives and subtracting the existing data from the total.  The missing data is then ready for writing to a new hard drive to complete the working array once more. See Understanding Parity for a further details. Depending on the sophistication of the controller the rebuilding process may have a number of options which may be preset for automatic operation, or available for user selection prior to the manual rebuild.  One excellent advantage that many hardware controllers offer is the ability to specify an automatic rebuild of the array in the event of failure. This procedure requires no user intervention and the controller will shut down all I/O requests and remove the target address from its entry as a usable drive. The controller establishes contact with the designated spare drive, bring in fully on-line, and begins to calculate the missing data and write it to the new member. This ability to either have rebuilds automated or manual initiated, can be vital in mission-critical environments, or situations where constant monitoring is not practical.  Chronology When a drive fails, or is reported as bad or missing, the controller will first check for a spare stand-by drive. If the controller supports an option of having a Local spare drive and a Global spare drive, it will first search for a Local spare drive. If it finds one it will initiate an automatic rebuild of the array. If it does not find a Local spare drive, it will search for a Global spare drive. Again, if it finds a suitable replacement it will commence the rebuilding of the logical drive to incorporate the new drive member.  If the controller does not differentiate between Local and Global drives every spare drive will be seen as a Global one and the controller will simply use the first working drive it finds. As soon as the controller has verified the operation and working order of the new drive, it can begin the rebuild process.  Setting the Reconstruction Priorities A further option sometimes given in conjunction with an automatic rebuild function is the ability to select a priority for the rebuild. When rebuilding a logical drive or volume the controller commits a certain amount of its resources to the process. The controller may allow you to specify whether the rebuilding process is given a Low, Normal, or High priority. If set Low the controller rebuilds the array in the background but immediately stops the rebuild to service all I/O requests it receives. When set to Normal, the controller commits an equal share of its resources to both tasks. When set to High the controller will commit resources to rebuilding the array regardless of the number of I/O requests. Each priority has definite advantages in suitable situations. It provides for the data from a failed disk to be automatically written to its replacement disk on-line and in a manner transparent to users. Reconstruction is a sequence of regeneration tasks in which the replacement disk is the recipient of all the data from the failed disk. The sequence, which can be lengthy, commences with a disk failure and ends when its replacement is rebuilt. RAID sub-systems offer a pathway to data redundancy, data regeneration, and reconstruction after hardware failure. RAID applies to the actual concept of offering data protection whilst EDAP specifies how this can be achieved together with strict criteria on the tolerances the actual physical equipment must offer. Related topics: Components: RAID Controllers Defining RAID Levels Understanding Parity   Raidhelp   RAIDhelp© Copyright 1999-2004 Antony Kershaw