FM 15-02 Chapter 15 · Operations and capacity
RAID calculator
Usable capacity, fault tolerance, theoretical speed, rebuild time and URE risk for RAID 0 to 60.
Local only Runs in your browser. Nothing you enter leaves this page.
Accepts 6 x 8 TB · 4x4TB, 2x8TB · 12, 12, 10 (TB if no unit) · 960 GB · 1 TiB
- Usable
- 40 TB · 36.38 TiB
- Raw
- 48 TB · 43.66 TiB · 6 disks
- Efficiency
- 83.3 %
- Fault tolerance
- 1 disk
- Read, theoretical
- × 6 one disk
- Write, theoretical
- × 1.5 one disk · Small random writes: penalty 4 (read data, read parity, write both)
- Rebuild one disk
- 14.8 h at 150 MB/s, best case
Unrecoverable read errors during a rebuild
- Read during rebuild
- 40 TB = 3.20e+14 bits
- P(at least one URE)
- 95.9 %
With one disk gone there is no redundancy left, so every surviving bit must read cleanly. P = 1 - (1 - 1/R)b ≈ 1 - e-b/R, with b bits read and R the spec. The spec is a worst-case bound ("less than 1 in R"); real drives usually do much better, and many controllers lose only the affected stripe rather than the whole array. Treat this as an upper bound, and as a reason to prefer RAID 6 or 10 for large disks.
Capacities in TB (10¹² bytes, as on the disk label) and TiB (2⁴⁰, as most operating systems show). Filesystem overhead, hot spares and controller metadata are not subtracted. Speed factors assume identical disks, full-stripe parallelism and no cache; real arrays differ. Rebuild time assumes the whole disk is rewritten at a steady speed with no other load.