RAID 10 Calculator
RAID 10 stripes across mirrored pairs, combining RAID 0's speed with RAID 1's simple redundancy — the standard choice for databases and VM datastores.
The RAID 10 formula
Usable Capacity = (N / 2) × Disk Size (50% efficiency)
Disks are paired into mirrors, then striped like RAID 0. Capacity overhead is a flat 50% no matter how large the array gets — unlike RAID 5/6, this never improves with scale, which makes RAID 10 the most expensive redundant level per usable terabyte.
Worked example
8 disks × 2 TB = 8 TB usable out of 16 TB raw, 50% storage efficiency, tolerates 1 failure per mirrored pair (up to 4 total across the array, as long as no pair loses both members). Read speed scales with total disk count; write speed is roughly half that, since every write updates both sides of a mirror.
Why RAID 10 is worth the capacity cost
Rebuilds are the fastest and simplest of any redundant RAID level — replacing a failed disk only requires copying from its surviving mirror partner, not reading the entire rest of the array like parity RAID does. Combined with strong random I/O performance, this makes RAID 10 the standard choice for transactional databases and VM datastores, where both rebuild risk and I/O latency matter more than raw capacity efficiency.
Why is RAID 10 more expensive than RAID 6 per usable TB?
RAID 10's 50% overhead is fixed regardless of disk count, while RAID 6's overhead (2 disks' worth) shrinks as a percentage the more disks you add — a 20-disk RAID 6 array is 90% efficient, while a 20-disk RAID 10 array is still exactly 50%.
Can RAID 10 survive 2 disk failures?
Sometimes — it depends which disks fail. If the two failed disks are in different mirrored pairs, the array survives fine. If both disks in the same mirrored pair fail, that pair's data is lost even though the rest of the array is intact.
Need a different RAID level? Use the full RAID Calculator covering RAID 0, 1, 5, 6, 10, 50, 60 and ZFS RAIDZ1/2/3, or read the RAID levels comparison guide.