Why does my array show less space than the disks add up to?

Two reasons stack. Redundancy takes whole disks, and the operating system reports base-2 units against the base-10 units the disks were sold in, which is another ten percent at tera.

Is RAID 5 still acceptable?

On large modern disks it is questionable. The rebuild runs for many hours reading every remaining disk, and a second failure in that window loses the array. RAID 6 exists for exactly this reason.

Technologystorage

RAID Capacity Calculator

Usable capacity, fault tolerance and rebuild time for the common RAID levels, in both the base-10 units the disks are sold in and the base-2 units the operating system reports. Rebuild time is the figure that decides whether single parity is still reasonable at modern disk sizes.

Also called: raid calculator, usable storage calculator.

Usable capacity

38 TB usable from 6 x 8 TB in RAID 5, 48 TB raw. 8 TB goes to redundancy, and the array survives 1 disk failure. The operating system will report 34.6 TiB, not 38, because drives are sold in base-10 units and reported in base-2 ones. Rebuilding one disk takes about 22.2 hours at 100 MB/s, and longer on a busy array. A second failure inside that window loses a single-parity array, which is the argument for RAID 6 on disks this size. RAID is not a backup in any case: it protects against a disk failing, not against deletion or corruption.

Raw capacity

Usable in TiB, as the OS reports it

34.56

Given to redundancy

Storage efficiency

83.3%

Disk failures tolerated

Rebuild time

22.2

On redundancy

8 TB goes to redundancy, and the array survives 1 disk failure. The operating system will report 34.6 TiB, not 38, because drives are sold in base-10 units and reported in base-2 ones.

On rebuilds

Rebuilding one disk takes about 22.2 hours at 100 MB/s, and longer on a busy array. A second failure inside that window loses a single-parity array, which is the argument for RAID 6 on disks this size. RAID is not a backup in any case: it protects against a disk failing, not against deletion or corruption.

How this is calculated

The same disks at other levels

Level	Usable TB	Failures tolerated
RAID 0	45.6	0
RAID 1	7.6	5
RAID 5	38	1
RAID 6	30.4	2
RAID 10	22.8	1

Method and background

How this is calculated

Each level trades capacity for redundancy in a fixed way: RAID 5 gives up one disk, RAID 6 two, RAID 1 and 10 half. The interesting number is not the capacity but the rebuild. A large array reading every remaining disk end to end to reconstruct a failed one runs for many hours, and during that window a second failure destroys a single-parity array. That window is why RAID 6 became the default on large disks and why RAID 5 on eight terabyte drives is a decision that needs justifying. RAID is not a backup either way: it protects against a disk failing, not against deletion, corruption or ransomware.

usable capacity is the disks less those consumed by parity or mirroring, which is a property of the level and not of the disks

n: Disks
p: Disks given to redundancy
s: Size of each disk

Worked examples

Each of these is asserted on every build. If a change to the engine ever moved one of these answers, the build would fail before the page could print it.

six 8 TB disks in RAID 5

Number of disks: 6
Size of each disk: 8 TB
RAID level: RAID 5, single parity
Filesystem overhead: 5%
Rebuild speed: 100 MB/s

Usable capacity38

5 usable disks x 8 TB less 5% overhead

Open this example

the same disks in RAID 6

Number of disks: 6
Size of each disk: 8 TB
RAID level: RAID 6, double parity
Filesystem overhead: 5%
Rebuild speed: 100 MB/s

Usable capacity30.4

boundary: one more disk of parity buys the second failure

Open this example

Written about this

Your 1 TB drive is not missing 69 GB · 1 min · A new 1 TB drive reports 931 GB and nothing is wrong. The manufacturer and the operating system are counting in different bases, and both are telling the truth.

Method and limits

What it assumes

Disks of equal size, since a mixed array is limited by the smallest member.

What it deliberately does not model

Rebuild speed varies with load, and a busy array rebuilds far more slowly than an idle one.
RAID protects against disk failure only. It is not a backup.

Formula version 1.0.0 · definition 1.0.0 · United States · Report a problem with this calculator

Frequently asked questions

Why does my array show less space than the disks add up to?: Two reasons stack. Redundancy takes whole disks, and the operating system reports base-2 units against the base-10 units the disks were sold in, which is another ten percent at tera.
Is RAID 5 still acceptable?: On large modern disks it is questionable. The rebuild runs for many hours reading every remaining disk, and a second failure in that window loses the array. RAID 6 exists for exactly this reason.