HP Surestore Disk Array 12h - And FC60 Advanced User's Manual page 55

RAID 3 works well for single-task applications using large block I/Os. It is not a good
choice for transaction processing systems because the dedicated parity drive is a
performance bottleneck. Whenever data is written to a data disk, a write must also be
performed to the parity drive. On write operations, the parity disk can be written to four
times as often as any other disk module in the group.
RAID 5
RAID 5 uses parity to achieve data redundancy and disk striping to enhance performance.
Data and parity information is distributed across all the disks in the RAID 5 LUN. A RAID 5
LUN consists of three or more disks. For highest availability, the disks in a RAID 5 LUN
must be in different enclosures.
If a disk fails or becomes inaccessible, the disk array can dynamically reconstruct all user
data from the data and parity information on the remaining disks. When a failed disk is
replaced, the disk array automatically rebuilds the contents of the failed disk on the new
disk. The rebuilt LUN contains an exact replica of the information it would have contained
had the disk not failed.
Until a failed disk is replaced (or a rebuild on a global hot spare is completed), the LUN
operates in degraded mode. The LUN must now use the data and parity on the remaining
disks to recreate the content of the failed disk, which reduces performance. In addition,
while in degraded mode, the LUN is susceptible to the failure of the second disk. If a
second disk in the LUN fails while in degraded mode, parity can no longer be used and all
data on the LUN becomes inaccessible.
Figure 22 illustrates the distribution of user and parity data in a five-disk RAID 5 LUN. The
the stripe segment size is 8 blocks, and the stripe size is 40 blocks (8 blocks times 5 disks).
The disk block addresses in the stripe proceed sequentially from the first disk to the
second, third, fourth, and fifth, then back to the first, and so on.
Disk Array High Availability Features
55