hesela.devRAID 1
RAID 1 stores a full copy of each logical block on multiple storage members.
Copies: A single N-way mirror with N equal-sized members of capacity S provides ideal data capacity S, before overhead. Two-way mirroring uses half the aggregate member capacity, but a three-way mirror uses one third. [1] [2]
Survival: One current, readable full copy can retain the data after other members fail. Array assembly, consistency and the availability of its controller and paths still matter. A stale or incomplete member is not interchangeable with a synchronized copy. [2] [3]
Recovery: Mirrored writes also propagate unwanted overwrites. RAID 1 does not by itself preserve a historical recovery point. Conventional paired RAID 10 combines multiple mirror groups with striping, rather than keeping a full array copy on every member. [4] [5]
Reads: Linux MD balances reads among members, but its documented single sequential stream does not gain the striping benefit of RAID 0. Concurrent streams or random reads can use multiple members; mirrored writes update the active copies. [3]
Synchronization: By default, mdadm starts an initial resync when creating a RAID 1 array. With optional write-behind enabled, Linux MD can report write completion before designated write-mostly members have caught up. Member count alone does not prove that every copy is current. [6] [3]
Relations
- related
- RAID, Mirroring (storage), Mirror set, RAID 0, RAID 10, Degraded array, Backup, MD mismatch count