RAID Fundamentals

Suppose a server stores its database on a single 8 TB drive. If that drive fails, the entire database becomes unavailable until the data is restored elsewhere.

With multiple drives, the system can combine them into one logical storage device and distribute data according to a chosen layout. Some layouts improve performance, some add fault tolerance, and some provide both.

This technique is called RAID, or Redundant Array of Independent Disks.

RAID combines multiple drives to improve performance, availability, or both.

RAID does not eliminate failures. Each RAID level makes different trade-offs between usable capacity, performance, and fault tolerance, so understanding RAID means understanding those trade-offs, not just memorizing level numbers.

This chapter compares striping, mirroring, and parity across RAID 0, 1, 5, 6, and 10, then examines capacity, degraded operation, rebuilds, write holes, integrity limits, implementation choices, and the boundary between RAID and backups.