RAID capacity calculator — usable space, fault tolerance & rebuild time
Enter your drive count, drive size and RAID level. You get the usable capacity, how many drives the array survives, and an estimated rebuild window — then one more step that the other calculators don't have: check whether your part numbers are SMR, because that spec, not the RAID level, is what stretches a rebuild from hours to days.
Is your part number SMR? Type a drive model — the array math above assumes CMR, and that's the assumption a rebuild can break.
How the numbers are computed
| RAID | Usable | Drives lost you can survive | Minimum drives | Best for |
|---|---|---|---|---|
| 0 | n × size | 0 | 2 | Speed, no redundancy (media, scratch space) |
| 1 | size | n/2 − 1 | 2 | Small critical volumes, mirroring |
| 5 | (n−1) × size | 1 | 3 | 3–5 drive home arrays, moderate rebuild risk |
| 6 | (n−2) × size | 2 | 4 | Larger arrays, 16 TB+ drives, the safe default |
| 10 | n/2 × size | 1 per mirror pair | 4 | Write-heavy workloads, mixed SSD/HDD |
Rebuild estimates assume 80–150 MB/s sustained (typical NAS controller, 5.4–7.2 kRPM drives) with a fast scenario at 250 MB/s. Real time depends on controller bandwidth, drive count in the pool and whether the drives are CMR or SMR — the SMR check below is the step most calculators skip.
RAID 5 vs RAID 6 — when the second parity disk is worth it
The cost of RAID 6 is one full parity disk: on a 4×16 TB build you lose 16 TB of usable space versus RAID 5. What you buy in exchange is the ability to survive a second failure while the first drive's rebuild is still running — and that window, not steady-state reliability, is where arrays actually die. Rebuilding a 16 TB drive takes many hours (use the estimate above); in that window every other drive in the pool is running hot, and large drives also carry a higher uncorrectable-read-error rate. The practical rules:
- 3–5 drives: RAID 5 is defensible; RAID 6 on 4 drives costs 25% of capacity and is the cheapest the level can be, so many 4-bay boxes run RAID 5 or 10 instead.
- 6+ drives, or any drive 16 TB and above: RAID 6 (or ZFS RAIDZ2) is the safer default — the rebuild window is long enough that a second failure is a real probability, not a tail risk.
- Write-heavy or VM workloads: consider RAID 10 — it rebuilds per-mirror (faster, less sustained read load) and trades the spare capacity for fault isolation.
- Whatever you choose: keep a 3-2-1 off-box backup. A rebuild is not a backup — it restores an array, not the mistakes, ransomware or config errors that also live on it.
If you're on FygoOS, the same math applies under ZFS: its RAIDZ1/RAIDZ2 map to RAID 5/6, and the FygoOS hard-drive guide covers the SMR angle specific to ZFS rebuilds. The pool layout guide shows how the levels map to partitioning.
FAQ
How is RAID usable capacity calculated?
RAID 0: n × size (no redundancy). RAID 1: size (mirroring, needs 2 or 4 drives). RAID 5: (n − 1) × size. RAID 6: (n − 2) × size. RAID 10: (n / 2) × size. We show the raw number; real-world usable space is a little lower after the file system reserves overhead.
How long does a RAID rebuild take?
A rebuild rewrites the lost data across the surviving drives, so the time scales with the drive size, not the failed drive alone. Our estimate uses 80–150 MB/s sustained rebuild rates (typical for a 5.4–7.2 kRPM NAS drive on a consumer controller) and a fast 250 MB/s scenario. A 20 TB drive in a RAID 5/6 can mean many hours to a day-plus at the slow end — which is exactly why the recording technology (CMR vs SMR) matters.
Why does SMR affect RAID rebuild time?
SMR (shingled) drives throttle when their cache fills, which happens during the sustained writes of a long rebuild. A drive that rebuilds at 150 MB/s as CMR can crawl below 50 MB/s once in shingle mode, stretching a 6-hour job past a day. Use the part-number checker on this page (or the full SMR/CMR checker) before you build the array.
RAID 5 or RAID 6?
RAID 6 is the safer default for most NAS builds once you have 6+ drives or large (16 TB+) drives, because the window where a second drive fails while the first is rebuilding is the main way arrays die. Our RAID 5 vs RAID 6 comparison page shows the capacity trade-off and the rebuild math in detail.