Redundant Array of Inexpensive Disks RAID

Solaris−flavored LVM is also discussed in greater detail in the case study in Chapter 27.

6.3.4 Redundant Array of Inexpensive Disks RAID

A Redundant Array of Inexpensive Disks RAID is a disk array setup, which enables the combined storage units to be used for storing duplicated mirrored data. The mirroring allows regeneration of the data in case of disk failures. There are several levels of RAID: RAID−0: Although it does not provide redundancy, striping is often referred to as a form of RAID, known as RAID−0. Striping is a technique of mapping data so that the data is interwoven among more physical disks. It offers a high data transfer rate and high IO throughput, because simultaneous data access across multiple disks can be performed. • RAID−1: Mirroring is a form of RAID known as RAID−1. Mirroring uses equal amounts of disk capacity to store the original data and its mirror. It provides redundancy of data and offers protection against data loss in the event of physical disk failure. • RAID−0+1: Striping combined with mirroring is known as RAID−0 + 1. It merges RAID−0 and RAID−1, providing redundancy and efficient access to data. • RAID−1+0: Mirroring combined with striping is known as RAID−1 +0. It merges RAID−1 and RAID−0, providing better redundancy and equally efficient access to data as RAID−0 + 1. Remember that for this RAID configuration mirroring is provided before striping, so multiple disk failures in different disk groups can still be handled. • RAID−2: Not widely implemented, RAID−2 uses bitwise striping across disks and uses additional disks to hold Hamming code check bits. • RAID−3: Uses a parity disk to provide redundancy. RAID−3 stripes data across all but one of the disks in the array, which is then used for the parity bit. • RAID−4: Represents a modified version of RAID−3 to overcome synchronization problems when data is accessed across multiple disks. By increasing the stripe unit size, a majority of IO operations can be located on a single disk without the need for synchronized simultaneous access to multiple disks. However, it still uses a separate parity disk to store redundant parity information. It is not widely implemented. • RAID−5: Represents an improved version of RAID−4, and it is practically implemented. Instead of using a separate parity disk, the parity data are also striped across all disks; the data stripes and parity stripes could be found on all disks. In case of a disk failure, the lost data can be recovered. RAID−5 provides the performance of RAID−0 + 1, but in a more economical way. • For all the options stated here, RAID−1 + 0 is probably the superior one. This is also the most expensive one, and not supported by older volume managers and disk arrays.

6.3.5 Snapshot