MB Networks
Drives · RAID · Usable Space
How much usable storage will your array give you?
Supports RAID 0/1/5/6/10, ZFS RAIDZ, Unraid, and Synology SHR — with equal or mixed drive sizes.
Usable capacity
9 TB
Fault tolerance
1 drive
Efficiency
75%
Raw total
12 TB
Min drives
3 for this level
For Unraid and SHR with mixed drive sizes, use the Mixed drives tab — those modes calculate correctly per-drive.
Quick add
Why usable capacity is never the number on the box
Four 8TB drives is 32TB raw — but the moment you put them in a redundant array, some of that capacity becomes parity or mirror overhead, not storage. How much you lose depends entirely on which RAID level or filesystem you pick, and the difference between levels is bigger than most spec sheets make it look. This calculator runs the actual math for ten different schemes so you can compare before you buy drives, not after.
Equal-size RAID vs. Unraid/SHR: the real difference
Traditional RAID (0/1/5/6/10) and ZFS RAIDZ both require every drive to be treated as the size of the smallest drive in the array — mix a 4TB drive into a set of 8TB drives and every drive effectively becomes 4TB. Unraid and Synology's SHR work differently: parity is sized to the largest drive, and usable capacity is the sum of every other drive. That's why Unraid/SHR exist — they let you add drives of different sizes over time (a common real-world pattern: buy whatever's cheapest per TB each time you expand) without wasting capacity the way equal-drive RAID does. The tradeoff is they're specific to Unraid and Synology's own software, not a RAID standard your controller or motherboard understands natively.
Capacity efficiency by scheme (4-drive example)
| Scheme | Min drives | Fault tolerance | Usable (4×8TB equal) | Efficiency |
|---|---|---|---|---|
| RAID 0 | 1 | None — any failure loses everything | 32TB | 100% |
| RAID 5 / RAIDZ1 | 3 | 1 drive | 24TB | 75% |
| RAID 6 / RAIDZ2 / SHR-2 | 4 | 2 drives | 16TB | 50% |
| RAID 10 | 4 (even) | Up to half, depends which drives | 16TB | 50% |
| Unraid (1 parity) / SHR | 2 | 1 drive | 24TB* | 75%* |
*Unraid/SHR efficiency shown here assumes equal drive sizes for a fair comparison — their real advantage only shows up with mixed sizes, which this table can't represent. Use the Mixed Drives tab above to model your actual drives.
Worked example
Six drives — two 12TB, four 8TB — in Unraid with single parity: usable capacity is the sum of every drive except the largest, so 12 + 8 + 8 + 8 + 8 = 44TB usable, with the remaining 12TB drive serving as parity. Put the same six drives into standard RAID 5 or RAIDZ1 and every drive gets clamped to the smallest (8TB): usable capacity is (6−1) × 8TB = 40TB. The 4TB difference is exactly the cost of RAID 5 not being able to use the two larger drives' extra capacity — multiply that gap by how often you actually buy mismatched drive sizes over an array's lifetime.
How much usable space do I actually get from RAID 5?
(n−1) × smallest drive size. Four 8TB drives in RAID 5: (4−1) × 8TB = 24TB usable, 75% efficiency. The percentage improves as you add drives — a 3-drive RAID 5 array is 67% efficient, a 10-drive array is 90% — because the one parity drive's cost is spread across more capacity.
What's the difference between RAID 6 and RAIDZ2?
Functionally similar — both tolerate two drive failures with the same usable-capacity math, (n−2) × drive size. RAID 6 is a hardware/software RAID standard; RAIDZ2 is ZFS's implementation of the same double-parity concept, with ZFS's additional checksumming and self-healing on top. If you're already on ZFS, RAIDZ2 is the equivalent choice; RAID 6 is for everything else.
Why does Unraid give more usable space than RAID 5 with mismatched drives?
Because Unraid sizes parity to your largest drive and sums every other drive as usable, while standard RAID clamps every drive down to the size of the smallest one. If your drives are all the same size, the two approaches land close to the same efficiency; the gap only opens up once you mix drive sizes, which is the normal way home/small-business arrays grow over time.
Can I mix drive sizes in RAID 5 or RAID 6?
You can physically install different sizes, but standard RAID treats every drive as the size of the smallest one in the array — the extra capacity on larger drives is simply wasted, not usable. If mixed drive sizes are a hard requirement, Unraid or Synology's SHR are built for exactly that case; equal-size RAID is not.
How many drives can fail before I lose data?
Depends on the scheme: 0 for RAID 0, 1 for RAID 5/RAIDZ1/Unraid-1-parity/SHR, 2 for RAID 6/RAIDZ2/SHR-2/Unraid-2-parity. RAID 10 is the exception — it can survive up to half the drives failing, but only if no two failed drives are the same mirrored pair; lose both drives in one mirror and the array is gone regardless of how many other drives are still healthy.
Is RAID 10 or RAID 6 better for a 4-drive array?
Both give the same 50% usable-capacity efficiency at 4 drives, but they fail differently. RAID 10 rebuilds faster (it's a direct mirror copy, not parity computation) and handles write-heavy workloads better, but a specific pair of drive failures kills the whole array. RAID 6 always survives any two drive failures and is more efficient as you add more drives, but rebuilds slower because of the parity math. For a database or VM host doing heavy writes, RAID 10; for bulk file storage, RAID 6.
What does "efficiency" mean in the results above?
Usable capacity as a percentage of raw (total) capacity. RAID 0 is 100% efficient because it has no redundancy at all; every other scheme trades some percentage of raw capacity for fault tolerance. Higher efficiency at the same fault tolerance is generally better, but drive count changes the tradeoff — the same RAID level gets more efficient as you add drives.
Related tools