RAID Calculator - Capacity and failure simulation

Compare RAID 0, 1, 5, 6 and 10 by usable capacity, then click disks to simulate failures.

RAID capacity calculator

Private by design: every number is computed in your browser.

1 to 32 disks in the array. Hot spares are counted separately.

Standby disks (0 to 8). They add to the installed total but never to usable space.

Capacity comparison

Usable space is theoretical, before formatting and filesystem metadata.

8 TB · usable8 TB installed
LevelUsableEfficiencyDisk failures tolerated
8 TB100%None — any disk loss stops the array.
2 TB25%Up to 3 disks — a single survivor keeps the data.
6 TB75%1 disk.
4 TB50%Up to 2 disks.
4 TB50%At least 1 disk; more when failures land in different mirrors.

Failure simulation

Select a level, then click disks to fail and recover them. Spares sit aside: they do not count as instant tolerance while an array rebuilds.

Simulated level: RAID 5 (single parity)

Click a disk again to mark it healthy.

Array state: Healthy — no failed disks.

GB and TB are decimal (1 TB = 1,000 GB); GiB and TiB are binary (1 TiB = 1,024 GiB). 2 TB ≈ 1.82 TiB. Formatting and filesystem metadata take a further cut, so every figure here is a theoretical maximum.

RAID is not a backup: it keeps the array running through disk failures, but it does not protect against deletion, ransomware, fire or theft. Keep independent copies of important data.

This page estimates capacity only. It never forecasts transfer speed, rebuild duration or failure probability.

Compare RAID levels by usable capacity

  1. Type how many active disks the array will use (1 to 32) and the capacity of each disk — for example 4 disks of 2 TB. Hot spares stay in a separate field so they never inflate the usable total.
  2. Worked example: 4 x 2 TB gives 8 TB on RAID 0, 2 TB on RAID 1 (one complete mirror), 6 TB on RAID 5 and 4 TB on both RAID 6 and RAID 10. These follow directly from (N - 1) x S and (N - 2) x S with N = 4 and S = 2 TB.
  3. Open the individual disk sizes when your disks differ: the conventional simulation caps every disk at the smallest one, so 2 + 2 + 4 + 4 TB behaves like 4 x 2 TB with 4 TB left unused.
  4. Pick a level to simulate, then click disks to fail them. RAID 10 treats pairs honestly: two failures inside the same mirror keep the array degraded, while losing both members of any pair stops it.
  5. Read TB as decimal terabytes and TiB as binary tebibytes: 2 TB formats to about 1.82 TiB, and the operating system plus filesystem metadata take a further cut, so treat every figure as a theoretical maximum.
  6. Copy the text summary or a link that reopens the same setup. Everything runs locally — no disk inventory ever leaves your browser.

RAID capacity questions

Why does RAID 1 with four disks still show 2 TB?

This page models RAID 1 as one complete mirror across all N disks, which yields exactly one disk of usable space (S). It is not modelled as independent pairs — that layout is RAID 10, which with 4 x 2 TB gives 4 TB instead.

Why do mixed disks waste space?

Conventional RAID stripes data evenly, so every disk can only contribute up to the smallest one. With 2 + 2 + 4 + 4 TB, each disk counts as 2 TB and 4 TB stays unused. Some vendor-specific layouts reuse that extra space, which is why the page labels the result conventional.

What is the difference between TB and TiB?

TB is decimal (1 TB = 1,000,000,000,000 bytes) and TiB is binary (1 TiB = 1,099,511,627,776 bytes), so the same array reads smaller in TiB: 8 TB is about 7.28 TiB. Drive makers advertise TB while operating systems usually display TiB.

Why are there no speed or rebuild-time forecasts?

Throughput and rebuild duration depend on the controller, drive speeds, workload and array state — none of which a capacity formula can see. Any page that prints MB/s or hours from disk sizes alone is inventing numbers, so this one refuses to.

Does RAID replace backups, and what do hot spares do?

No: RAID survives disk failures but not deletion, ransomware, controller faults, fire or theft. A hot spare is a standby disk that can start rebuilding automatically, but it is not instant tolerance — during the rebuild the array is still exposed, which is why the simulator keeps spares aside.