Infrastructure estimates
How much RAID capacity do four 4 TB disks provide?
Tools to use
Multiplying drive size by count can differ substantially from usable array capacity. Redundancy and spares dedicate part of the same physical capacity to other purposes. Comparing RAID configurations helps separate purchased capacity, modeled array capacity, and filesystem free space.
Four 4 TB drives in RAID 6 model as 8 TB
total disks: 4
disk size: 4 TB
RAID level: 6
spares: 0Raw capacity is 16 TB, modeled usable capacity 8 TB, and redundancy 8 TB. The formula is (4−2)×4 TB. 'Two drives' describes the capacity equivalent of parity, not two particular drives dedicated to holding all parity.
Filesystems and metadata can consume additional space beyond this model. TB/TiB display differences matter too, so an operating system showing a number other than eight is not by itself evidence of lost capacity. Convert to the same byte basis before comparing units.
The same four drives give different capacities by level
With four 4 TB drives and no spares, the model gives RAID 0:16 TB, RAID 5:12 TB, RAID 6:8 TB, RAID 10:8 TB. RAID 0 has no redundancy capacity, so maximum space is not automatically the best choice. Equal 8 TB results for RAID 6 and 10 do not imply identical architecture or performance.
This tool models RAID 1 as all active drives mirroring the same data, so four 4 TB drives give 4 TB. That is not two separate two-drive mirror groups. Identify the mirroring and grouping structure rather than inferring capacity from the level name alone.
A spare does not add normal usable capacity
Five 4 TB drives with one spare leave four active drives, so RAID 6 remains 8 TB usable. Raw 20 TB splits into 8 usable, 8 redundancy, and 4 spare. With no spare, all five participate and the model gives 12 TB. Roles change the result even with the same purchase.
Keeping four total drives and selecting one spare leaves only three active drives, which fails the RAID 6 minimum. Spares are part of the total, not added on top. Write down total, active, and spare counts separately before input.
Minimum and parity constraints belong to the calculation
The model requires at least two active drives for RAID 0/1, three for RAID 5, four for RAID 6, and an even count of at least four for RAID 10. Five active RAID 10 drives are rejected; five total with one spare leave four active.
Devices and software may support other layouts or proprietary expansion schemes. The calculator's rules are not a compatibility catalog for all products. Confirm supported counts, drive requirements, and expansion procedures in the target product.
For mixed sizes, inspect the smallest-drive assumption
The engine models active drives at the smallest capacity. The regular UI currently takes one common drive size, not a detailed mixed-drive list. Conceptually, 4, 4, 8, and 8 TB under that model still give 8 TB usable in RAID 6; unused portions of larger drives are not assumed usable automatically.
Proprietary mixed-size layouts or multiple combined regions are different models. If a product's calculator differs, compare assumptions first. A smaller result here does not prove that the device must waste that exact difference.
Separate RAID capacity from backup capacity
RAID is not automatically a history of deleted files. Incorrect updates and deletions can be mirrored as valid writes. Redundancy capacity is distinct from a backup that restores an earlier state. Estimate backup retention and increments separately.
This calculation does not determine IOPS, rebuild time, failure probabilities, or application performance. Combine the capacity breakdown with backup requirements, specifications, and measurements. Report a qualified result such as '8 TB modeled for RAID 6, four 4 TB drives, no spare, before metadata overhead.'
Equal usable capacity does not imply equal behavior
Four 4TB drives give 8TB in both RAID 6 and RAID 10 under this model, but equal capacity does not imply equal failure behavior. In RAID 10, which mirror members fail matters; half the raw capacity being redundant does not guarantee survival of any two failures. The tool does not simulate failure combinations or rebuild states.
A spare being present is different from a rebuild being complete. Replacement delay, degraded performance and additional failures are outside the capacity formula. Counting a spare as usable capacity overestimates storage. Read total, spare and active counts separately, and verify that adding a spare has not reduced active drives below the RAID minimum.
The 8TB result precedes filesystem, reservation and product-specific overhead. A product displaying TiB introduces a representation difference as well. Align units first, then investigate remaining differences in actual usable space. Separating the RAID formula, unit notation and allocated filesystem capacity makes the apparent loss explainable.