What is RAID?
RAID stands for Redundant Array of In-line Drives/Disks and is a method of managing hard drives in servers.
Hard drives are the only moving component (except fans) inside computers and as such they are prone to mechanical failures. The failure of a hard drive would be the most common server fault and also the most costly as the hard drive is the only data store. RAID is a method of configuring hard drives in a way that prevents single point failure of the servers data store. It comes in seven flavours, though the main ones are RAID O, 1 and 5.
RAID 0 – Stripped: Where the data is split evenly over the number of hard drives. This is mainly used for performance boosting and should not be used for data protection. As the failure of one drive will destroy the array and the remaining data would be garbage without a complete array. Minimum of two drives needed.
RAID 1 – Mirrored: Where the data on the system is copied on all the drives so that in the event of a drive failing there is a complete copy on the remaining. Typically, this also allows for continued use of the system with no interruption in service. Minimum of two drives needed.
RAID 5 – Stripped with Parity: Similar to RAID 0 but with a checksum in place that allows the RAID to recover from a single drive failure by “rebuilding” the missing data using the other drives and a checksum. Minimum of three drives needed.
It is recommended that all servers have at least RAID 1 and ideally RAID 5 in place as their first backup level.