RAID 6
RAID 6 uses dual parity protection to retain data after any two member failures, provided the remaining required data and redundancy are readable and consistent.
For N equal-sized active members of capacity S, nominal data capacity is (N - 2) x S before metadata and filesystem overhead, excluding standby spares [2]. In the Linux coding scheme, P is XOR parity and Q is a distinct finite-field syndrome, not another copy of P [3]. Remaining protection is evaluated per affected stripe, not by counting errors anywhere in the array [5]. Count unavailable fragments at the same protected positions within blocks; recovery from non-overlapping damaged ranges is implementation-dependent [3]. Equal error counts and location marginals on individual surviving members do not determine how often erasures coincide in a stripe; the joint placement and remaining parity budget are also required [6].
Relations
- related
- RAID, RAID 5, Erasure coding, Reed–Solomon codes (storage), Read-modify-write, Parity RAID, Degraded array, RAID rebuild, RAID write hole, Latent sector error, Data scrubbing, LVM, Stripe unit
Advantages
- Any two missing members are within the configured device-failure tolerance under the reconstruction assumptions. [1]
Limitations
- Two known erasures are not equivalent to two arbitrary silent corruptions; dual parity alone does not guarantee recovery of the latter. [3]
- A decoder assuming one corrupt symbol can misidentify the culprit and damage a third member when two members are corrupt at corresponding byte positions; warning heuristics do not prove the single-corruption assumption. [3]
- Dirty and degraded parity can invalidate recovery assumptions; Linux MD normally refuses ordinary assembly in this state. [4]
- The level alone does not specify rebuild duration, availability policy or protection against deletion and shared failures. [5]
Citations
- SNIA Dictionary: RAID 6
- mdadm md(4): RAID4/5/6 layouts and capacity (revision e5a1d3e)
- H. Peter Anvin: The mathematics of RAID-6, sections 2 and 4 (2011 revision)
- Linux 6.12: RAID arrays, dirty/degraded assembly and corrected read errors
- Hesela: RAID 5/6 capacity and per-stripe recovery boundaries
- Hesela: identical erasure marginals with different RAID6 recovery outcomes (exact synthetic model)
Sources & caveats
- The linked placement example is exhaustive synthetic enumeration, not a measured failure probability or a hardware rebuild experiment.