It is built as mirrored pairs with striping across them, so whether a second failure is survivable depends entirely on which disk it is. Lose one from each pair and you are fine. Lose both from one pair and the whole volume goes.
Disks pair off into mirrors, and data is striped across the pairs. You give up half the raw capacity and get quick writes and failure behaviour that is easy to reason about.
RAID 10 pairs disks into mirrors, then stripes across those pairs. Four disks become two mirrored pairs; the stripe spreads data across both pairs for speed while each pair keeps a complete second copy for safety.
The order matters and is frequently misunderstood. It is not striped-then-mirrored, which would be RAID 01 and behaves quite differently under failure. In RAID 10, losing a disk costs you nothing because its partner holds the same data.
The consequence is that RAID 10 can survive multiple simultaneous failures — provided they land in different pairs. Lose one disk from each of two pairs and the array carries on. Lose both halves of one pair and the whole volume goes, however healthy the others are.
Both halves of one pair failed. The only way to lose a RAID 10, and it happens because mirrored pairs are usually identical disks bought together, installed together and worn identically.
The first failure went unnoticed. A degraded pair looks like a healthy array from outside. When its partner goes months later, the volume drops without warning.
A rebuild ran onto the wrong member. Rebuilding in the wrong direction within a pair overwrites the good copy. Recoverable, but considerably harder than it needed to be.
The controller lost its configuration. The disks are fine and the map describing which pair is which has gone. Derivable, but it has to be derived before anything reads.
At least one, and possibly half of them — but only if the failures land in different pairs. Two failures within one pair destroys the volume regardless of how many disks remain.
Different rather than safer. RAID 10 rebuilds far faster because it copies one disk rather than recomputing parity, which shortens the risky window. RAID 6 tolerates any two failures wherever they land.
Then both were in the same pair. Send every member labelled by bay — the surviving pairs are intact and the failed pair is frequently partially recoverable, which may be enough.
Yes. Pair membership and stripe geometry are derived from the disks themselves, which is standard work on any array whose controller has gone.
From £500 +VAT, fixed in writing after the free diagnostic, regardless of member count.
Five to ten working days. Every member is imaged before reconstruction, and that imaging is most of the elapsed time.