RAID

From Sysadmin

Jump to: navigation, search

A Redundant Array of Independent Disks is a broad term describing several different strategies provide additional performance or reliability to hard disk drives.

Performance and Redundancy

This table lists all RAID types that are commonly implemented.

RAID TypeModeCapacityFault Tolerance2 Disks3 Disks
ReadWriteReadWrite
RAID 0Normal100%---200%200%??
RAID 1Normal??????
Degraded?---????
RAID 5Normal?Error Detection------??
Degraded?---------??
RAID 6Normal?Error Correction------??
Degraded?Error Detection------??
Double Degraded?---------??
RAID 01Normal??????
Degraded??????
RAID 10Normal??????
Degraded??????
RAID 50Normal??????
Degraded??????
RAID 60Normal??????
Degraded??????


Capacity: The percentage of the capacity of a single drive within the array that is available for storage.

Performance: The read and write performance for the array as a total compared to each individual drive.

Simultaneous Drive Failure

Failure of two drives in an array is not as rare as it should be. Common causes include:

  • Failure of one without anyone noticing (they notice when the second one fails)
  • Failure due to a common environmental effect (eg, heat, water)
  • Failure as the drives were manufactured together an have a similar MTBF. It is now generally recommended to use dissimilar HDDs in RAID, in contrast to older advice.

Not a Backup Strategy

RAID is not a backup strategy as it is neither offsite nor offline.

A widely held belief is that RAID is a replacement for a proper backup strategy. This is commonly expressed as We don't need to backup, we use RAID 1! This is absolutely not true. RAID is not a backup strategy.

There are many failure modes that RAID will not protect you from. These include: user error, sysadmin error, failure of the RAID controller, destruction of the server, and destruction of the building holding the server.

Personal tools