What the RAID calculator does
Choose a RAID level, enter the number of drives and their capacity, and the calculator shows the usable capacity, the space lost to redundancy, and how many drive failures the array survives. It is the first check when specifying a NAS, a home server, a video-editing workstation or a small business file server, and it runs in your browser.
The levels compared
| Level | Usable capacity | Survives | Min drives | Character |
|---|---|---|---|---|
| RAID 0 (stripe) | n × size | 0 failures | 2 | Fastest; any single failure loses everything |
| RAID 1 (mirror) | size (of the smallest) | n − 1 failures | 2 | Simple, fast reads, 50% efficiency with 2 drives |
| RAID 5 (single parity) | (n − 1) × size | 1 failure | 3 | Good balance for 3–6 drives; slow rebuilds on large disks |
| RAID 6 (double parity) | (n − 2) × size | 2 failures | 4 | Safer for large arrays; the usual choice for 6+ drives |
| RAID 10 (mirror + stripe) | n ÷ 2 × size | 1 per mirror pair | 4 (even) | Fast writes and rebuilds; 50% efficiency |
6 × 8 TB drives:
RAID 0 48 TB usable RAID 1 8 TB RAID 5 40 TB
RAID 6 32 TB usable RAID 10 24 TBChoosing a level
- RAID 0 only for scratch space you can recreate — render caches, temp video storage. It multiplies the chance of loss by the number of drives.
- RAID 1 for two-bay NAS units and boot drives: simplest, and a single surviving drive is readable on its own.
- RAID 5 for 3–5 drives up to a few TB each. With 8 TB+ drives, the hours-long rebuild after a failure is when a second drive is most likely to fail — which loses everything.
- RAID 6 for 5+ large drives; the second parity covers the rebuild window.
- RAID 10 for databases and VMs where write performance matters and 50% overhead is acceptable.
- ZFS RAID-Z1/Z2/Z3 and Synology SHR are equivalents with the same capacity maths (Z2 = RAID 6) plus checksumming and, for SHR, mixed drive sizes.
Mixed drive sizes
Conventional RAID treats every drive as the size of the smallest. Three drives of 4, 4 and 8 TB in RAID 5 give (3 − 1) × 4 = 8 TB usable, wasting 4 TB. Enter the smallest drive's size in the calculator for standard RAID; for SHR, Unraid or Drobo-style pools that use the extra space, the vendor's own calculator gives the real figure.
RAID is not a backup
Redundancy protects against drive failure and nothing else. Accidental deletion, ransomware, a failed controller, a power surge, fire or theft all take the whole array with them, mirrored perfectly. Treat the usable capacity as primary storage and keep a separate backup — the 3-2-1 rule: three copies, on two kinds of media, one off-site. The space “lost” to parity is buying uptime, not safety.
Rebuild and failure realities
- Rebuilding an 8 TB drive at 100–150 MB/s takes 15–24 hours of full-array stress.
- Unrecoverable read error rates on consumer drives (1 in 10¹⁴ bits) make hitting an error during a large RAID 5 rebuild statistically likely; RAID 6 tolerates it.
- Drives bought together fail together — same batch, same age, same environment. Stagger purchases or mix vendors.
- Use NAS-rated drives (CMR, not SMR) — SMR drives can drop out of arrays under rebuild load.
- Keep a spare drive on the shelf, or configure a hot spare, so rebuilds start immediately.