High Capacity RAID Systems.
RAID configurations are used to protect against the failure of hard disks, storing data redundantly across a small group of disks. When one disk fails a spare disk is brought into play to rebuild the system either from a mirror copy in RAID 1 or from the remaining data and parity in RAID 5 or RAID 6. A RAID1 or RAID5 configured system cannot cope with more than one hard disk failing during the rebuild. A RAID 6 array, thanks to its dual parity can cope with two concurrent disk failures.
RAID Rebuild Times
The problem with RAID systems is that they take an inordinate amount of time to rebuild in a failure situation, which involves operational disruption for an increasing number of businesses. This disruption is a consequence of the large capacity of the disks being used and the fact that the read/write speeds have not increased in relation to the growth in capacity. With the array in use the rebuild operation must also compete with normal I/O operations and with a large capacity disk fail the rebuild will be extremely slow to recover.
Typically an array using 2TB Hard Disk Drives will have a rebuild time of 6-8 hours and upward.
We must also take into account that under rebuild a whole hard disk drives’ worth of data needs to be read off the remaining hard disk drives in the RAID array. This is a heavy workload that will place stress on the remaining hard disk drives and increase the probability of a second hard disk drive failure.
Larger numbers of smaller capacity enterprise class hard disk drive units in a nested RAID configuration can militate against disruption during a failure situation. In considering an implementation involving a larger number of disks this has to be offset against the growing complexity of parity and access times.