BPC, Articles and Glossary

RAID 1 - Mirroring - Fault Tolerance

Definition: RAID 1 mirroring is an arrangement of hard disks that creates an exact copy (or mirror) of a set of data on two or more disks. This is useful when read performance or reliability are more important than data storage capacity.

RAID I, for example, writes two copies of the data simultaneously on two separate drives. This is called fault tolerant because if one of the mirrored drives suffers a mechanical failure (e.g. spindle failure) or does not respond, the remaining drive will continue to function. The RAID 1 configuration is performed either by a hardware RAID controller… or performed in software. It is suited to applications requiring high fault tolerance at a low cost and where a duplicated set of data is more secure than using parity. RAID 1 is popular for accounting and other financial data. It is also commonly used for small database systems, enterprise servers, and home PCs where a fairly inexpensive fault tolerance is required.

A RAID device maintains a mirror of all the data in a partition... on another partition. This second partition is usually on another drive (note: in Linux systems the second partition can be on the same drive). There is a small performance hit to be expected when configuring your hard disks in a RAID 1 partition as the data has to be written to every disk in the RAID array.

Wikipedia

More information on RAID

How to install RAID 1 on a Linux system

RAID 1 can be used together with RAID 0 (RAID 0 + 1) providing the advantages of both striping and mirroring. An interesting concept is that RAID 0+1 is not the same as RAID 1+0 and this technical article explains the difference and makes a case for why RAID 1+0 is better than RAID 0+1.