How is usable storage capacity calculated in RAID 5 and RAID 6?

RAID 5 allocates the capacity equivalent of exactly one disk for distributed parity, resulting in usable capacity calculated as (N - 1) × disk size. RAID 6 utilizes dual parity, reserving the capacity of two disks, resulting in (N - 2) × disk size.

Why is formatted drive capacity lower than advertised storage?

Hard drive manufacturers measure capacity using decimal notation (1 TB = 1,000,000,000,000 bytes). Operating systems compute capacity using binary notation (1 TiB = 1,099,511,627,776 bytes). This creates an inherent 9.09% reduction in reported capacity.

What is the difference between ZFS RAIDZ1, RAIDZ2, and RAIDZ3?

ZFS RAIDZ configurations assign dynamic parity across vdevs. RAIDZ1 tolerates 1 disk failure per vdev, RAIDZ2 tolerates 2 disk failures, and RAIDZ3 tolerates up to 3 disk failures per vdev.

RAID & ZFS Storage Capacity Calculator

01 Configure Array

Enter your drives, then choose a layout.

Number of Drives

Capacity per Drive

Unit

Only count drives that are actually part of the array here. Standby drives go in Hot Spares below.

Hot Spare Drives (optional)

Hot spares sit outside the array on standby and add zero usable capacity — they just swap in automatically after a drive fails.

Example: a 3-drive RAID 1 with one spare = 2 in "Number of Drives" above, 1 here.

Array Layout

Format / Filesystem

Display Units

Drive vendors use decimal (1 TB = 10¹² bytes); Windows/most OSes report binary (1 TiB = 2⁴⁰ bytes). Same data, different units — switch to see the number your OS would show.

02 Usable Capacity

Live results based on your configuration.

Usable Capacity (after redundancy & filesystem overhead) —

Raw Total Capacity

—

Usable After RAID

—

Redundancy Overhead

—

Filesystem Overhead

—

Total Physical Drives (Array + Spares)

—

Storage Efficiency

—

03 Array Visualizer

Data Parity / Mirror Hot Spare

Understanding RAID & ZFS Array Capacity

Designing storage arrays requires accurate capacity calculations that factor in parity distribution, redundancy overhead, spare allocations, and base-10 to base-2 binary conversion losses. This calculator evaluates standard hardware and software RAID levels along with OpenZFS pool layouts to provide operational metrics before deployment.

Standard RAID Level Overview

RAID 0 (Striping): Combines all drives into a single logical volume with 100% capacity efficiency. Offers no fault tolerance; a single drive failure causes complete data loss.
RAID 1 (Mirroring): Duplicates identical data across paired drives. Provides 50% usable capacity with single-drive fault tolerance per mirror set.
RAID 5 (Block-Level Parity): Distributes single-parity blocks across a minimum of 3 drives. Usable storage equals (N - 1) × drive size, sustaining 1 drive failure.
RAID 6 (Dual Parity): Uses dual distributed parity across a minimum of 4 drives. Usable storage equals (N - 2) × drive size, surviving up to 2 simultaneous drive failures.
RAID 10 (Striped Mirrors): Combines RAID 1 mirroring with RAID 0 striping across 4 or more drives. Provides high throughput and 50% storage capacity.
RAID 50 & RAID 60 (Nested Parity): Combines block-level parity vdevs into striped sets, optimizing rebuild speed and fault tolerance for enterprise high-density arrays.

OpenZFS Pool Architecture (RAIDZ)

OpenZFS bypasses traditional hardware controllers, implementing soft allocation via Virtual Devices (vdevs).

RAIDZ1: Single parity overhead per vdev (similar to RAID 5, requires 3+ drives).
RAIDZ2: Double parity overhead per vdev (similar to RAID 6, requires 4+ drives).
RAIDZ3: Triple parity overhead per vdev (tolerates 3 disk failures per vdev, requires 5+ drives).

Decimal (TB) vs. Binary (TiB) Overhead Conversion

Storage vendors market drives using base-10 metrics (1 TB = 10¹² bytes), whereas host operating systems calculate storage in base-2 binary tebibytes (1 TiB = 2⁴⁰ bytes = 1,099,511,627,776 bytes). Converting raw vendor capacity to usable binary space uses the equation:

Capacity (TiB) = [ Raw Capacity (TB) × 1,000,000,000,000 ] / 1,099,511,627,776