RAID 10 Explained for NAS

What Is RAID 10?

RAID 10, also written as RAID 1+0, combines two techniques:

• Mirroring: Data is duplicated across pairs of drives.
• Striping: Data is distributed across those mirrored pairs for higher throughput and parallel I/O.

A four-drive RAID 10 array contains two mirrored pairs. The NAS then stripes data across both pairs.

RAID 10 makes sense when you want:

• Strong read and write performance
• Predictable redundancy
• Fast, relatively simple rebuilds
• Good performance for virtual machines, databases, and busy multi-user workloads
• More usable capacity than a fully mirrored array, while retaining drive-level redundancy

Its main costs are capacity efficiency and expansion flexibility. You give up half of the raw drive capacity to mirroring, and adding capacity usually requires careful planning around matched pairs.

RAID 10 Capacity: You Get About Half the Raw Capacity

The basic usable-capacity formula is:

Usable capacity ≈ total raw drive capacity / 2

For drives of equal size:

Usable capacity ≈ number of drives × drive capacity / 2

The NAS filesystem, RAID metadata, and manufacturer unit conversions reduce the usable figure slightly. A NAS may also display capacity differently depending on whether it uses TB or TiB.

Worked example: four 12 TB drives

Suppose you install four 12 TB drives:

• Raw capacity: 4 × 12 TB = 48 TB
• RAID 10 usable capacity: 48 TB / 2 = 24 TB
• Capacity remaining for files: somewhat less than 24 TB after filesystem and system overhead

The array consists of two mirrored pairs:

Pair A: Drive 1 + Drive 2
Pair B: Drive 3 + Drive 4

Data is striped across Pair A and Pair B.

If you replace the four drives with four 16 TB drives, the theoretical RAID 10 capacity becomes:

4 × 16 TB / 2 = 32 TB

However, replacing only one drive does not increase capacity. The array generally cannot use additional space until the relevant mirrored pairs and the NAS implementation support an expansion operation.

How Many Drive Failures Can RAID 10 Survive?

RAID 10 can survive at least one failed drive, provided the remaining drive in that mirrored pair is healthy.

It can also survive multiple drive failures if they occur in different mirrored pairs. For example, in a four-drive array:

Pair A: Drive 1 + Drive 2
Pair B: Drive 3 + Drive 4

The array can survive:

• Drive 1 failing and Drive 3 failing
• Drive 1 failing and Drive 4 failing
• Drive 2 failing and Drive 3 failing
• Drive 2 failing and Drive 4 failing

It cannot survive both drives in the same mirrored pair failing. If Drive 1 and Drive 2 fail, Pair A is lost and the array is normally unavailable or unrecoverable through RAID alone.

This means RAID 10 does not have one fixed “number of failures tolerated.” Its tolerance depends on which drives fail.

More drives improve the probability of surviving multiple failures—but not without limits

An eight-drive RAID 10 array has four mirrored pairs. It can tolerate one failed drive in each pair, but losing both members of any one pair can destroy the array.

That is different from RAID 6, where the array is designed to tolerate any two drive failures, subject to the RAID implementation and the condition of the remaining drives.

RAID 10 Rebuilds

When a drive fails, the NAS rebuilds its mirror onto a replacement drive. In a traditional RAID 10 layout, the rebuild primarily copies the contents of the surviving member of the affected mirror rather than reconstructing every block across all drives with parity calculations.

This generally gives RAID 10 several practical advantages:

• Rebuilds are conceptually simpler than parity reconstruction.
• The surviving mirror member supplies the data for the replacement.
• The other mirrored pairs can continue serving data, depending on the NAS and workload.
• There is no parity calculation penalty during normal writes or rebuilds.

The actual rebuild time depends on factors such as:

• Drive capacity and sustained throughput
• How full the array is
• NAS processor and RAID implementation
• Current user activity
• Filesystem behavior
• Whether the NAS performs a full or usage-aware rebuild
• Drive health and error recovery

Do not treat RAID 10 as immune to rebuild risk. While rebuilding, the affected mirror has no redundancy. If its remaining drive fails before the rebuild completes, that mirrored pair is lost. A second failure in another pair may or may not be survivable, depending on the layout and which drives failed.

Large modern drives can take many hours or longer to rebuild. Keep a tested backup available throughout the process.

RAID 10 Versus RAID 5 and RAID 6

RAID 10 is usually selected for performance, rebuild behavior, and predictable mirror-based redundancy. RAID 5 and RAID 6 use parity, which improves capacity efficiency but introduces different performance and rebuild trade-offs.

Feature	RAID 10	RAID 5	RAID 6
Minimum drives	Typically 4	Typically 3	Typically 4
Usable capacity	About 50% of raw capacity	About one drive’s capacity less than raw	About two drives’ capacity less than raw
Failure tolerance	One drive per mirrored pair; multiple failures may be survivable	One drive	Any two drives, subject to implementation and remaining-drive health
Small-write behavior	Usually favorable because there is no parity update	Parity read-modify-write can add overhead	More parity work than RAID 5
Rebuild behavior	Rebuilds a failed mirror member	Reconstructs data using parity	Reconstructs data using dual parity
Expansion	Usually requires adding matched mirror pairs	Often more flexible, but NAS support varies	Often more flexible, but NAS support varies
Best fit	Performance-sensitive workloads and predictable rebuilds	Capacity-focused general storage where one-drive protection is acceptable	Larger arrays where two-drive fault tolerance is a priority

The “best” choice depends on more than benchmark performance. Drive count, drive size, workload, backup strategy, expansion plans, and the NAS operating system all matter.

RAID 10 Expansion Flexibility

RAID 10 expansion is less straightforward than simply inserting one larger drive.

Because RAID 10 is built from mirrored pairs, many implementations expand by adding another complete pair. For example:

Four drives: 2 mirrored pairs
Six drives: 3 mirrored pairs
Eight drives: 4 mirrored pairs

A six-drive array can therefore provide approximately:

6 × drive capacity / 2 = 3 × drive capacity

But support varies by NAS platform. Some systems can add a mirrored pair to an existing storage pool. Others may require rebuilding the pool, creating a new volume, or replacing every drive with a larger model before the extra space becomes available.

Before choosing RAID 10, verify:

• Whether the NAS supports RAID 10 specifically
• Whether it can expand RAID 10 by adding a matched pair
• Whether expansion requires drives of the same or larger capacity
• Whether the new drives must be installed as a pair
• Whether the storage pool or volume must be taken offline
• Whether the NAS can migrate from another RAID level to RAID 10
• Whether the process preserves data, and what backup is required first

Do not buy a four-bay NAS assuming you can later add a fifth drive to increase RAID 10 capacity. A fifth drive may serve as a spare, remain unused, or require a different storage arrangement. Confirm the platform’s documented expansion behavior before purchasing drives.

Drive Selection for a RAID 10 NAS

RAID 10 does not require every drive to be identical, but mismatched drives can complicate capacity and performance.

For a new array, aim for:

• Drives with the same nominal capacity
• NAS-appropriate models when the workload and vendor guidance call for them
• Consistent sector format and interface requirements
• Similar performance characteristics within each mirrored pair
• Enough endurance for the workload
• A replacement plan for the same or larger capacity

The usable capacity of a mirror is normally limited by the smaller drive. If you combine different capacities, the NAS may leave some space unused or use only the common capacity.

Also consider whether your NAS supports hot-swapping, drive health monitoring, scheduled scrubs, and notifications. These features do not prevent failures, but they reduce the chance that a failed drive goes unnoticed.

RAID 10 Is Not Backup

RAID 10 is redundancy, not backup.

It helps keep a NAS operating when a drive fails, but it does not protect against:

• Accidental deletion
• Ransomware or malware
• File corruption replicated across mirrors
• Theft or fire
• A failed NAS enclosure or controller
• Improper administration
• Damage caused by power events
• Multiple failures that destroy a mirrored pair

A deletion is normally replicated to both sides of a mirror. If ransomware encrypts files, the encrypted files can also be replicated. RAID cannot restore an earlier version unless you have separate snapshots or backups.

A practical NAS protection plan should include:

1. RAID 10 for local drive redundancy
2. Snapshots or versioning where supported
3. A separate backup target
4. An off-site or cloud copy for important data
5. Periodic restore tests

Use the 3-2-1 principle as a starting point: keep at least three copies of important data, on at least two types of storage or locations, with at least one copy off-site.

When RAID 10 Makes Sense

Choose RAID 10 when:

• You have at least four drive bays.
• You can accept losing about half of raw capacity to mirroring.
• Performance and low write overhead matter more than maximum capacity efficiency.
• The NAS will run virtual machines, databases, active project files, or other I/O-intensive workloads.
• You want mirror-based rebuild behavior rather than parity reconstruction.
• Your expansion plan can accommodate additional matched drive pairs.
• You maintain backups separate from the NAS.

RAID 10 can also be a strong choice for a Plex NAS when the system is serving multiple users, handling downloads, managing metadata, or running other applications at the same time. For a basic single-user Plex library, however, RAID 10 may provide more performance and redundancy than the workload needs. Media files still need a backup strategy, and RAID does not replace one.

RAID 10 is less attractive when:

• Maximum usable capacity is the priority.
• The NAS has only three drive bays.
• You need protection from any two drive failures regardless of which drives fail.
• Your NAS cannot expand RAID 10 in the way you plan.
• The workload is mostly sequential media storage and does not justify the lost capacity.
• You expect to mix drive sizes without checking the platform’s rules.

SSD cache does not automatically make RAID 10 unnecessary. Cache can help some workloads, but it adds cost and may introduce additional configuration and endurance considerations. Select the RAID layout for the storage pool’s reliability and capacity requirements first, then evaluate cache based on measured workload behavior.

RAID Decision Table

Your priority	Usually consider	Why
Strong performance and simple mirror-based rebuilds	RAID 10	Uses mirrored pairs and avoids parity updates
Maximum usable capacity with one-drive protection	RAID 5	Sacrifices roughly one drive’s capacity to parity
Protection from any two drive failures	RAID 6	Sacrifices roughly two drives’ capacity to dual parity
Four or more drives, but you need a simple mirror	RAID 10	Provides predictable pair-based redundancy
A small two-drive NAS	RAID 1	Mirrors the two drives; RAID 10 normally needs four
No redundancy, easily replaceable or backed-up data	Single disk or separate backup design	Lowest capacity overhead, but no local drive fault tolerance
High-value data	RAID plus independent backups	RAID keeps services running; backups restore lost or damaged data
Future expansion by adding one drive at a time	Check RAID 5/6 and NAS-specific options	RAID 10 commonly expands by matched pairs, if supported

Before buying, compare the NAS bays, supported RAID levels, expansion rules, drive capacities, network connection, backup destination, and expected workload. You can Browse NAS or review available NAS and storage servers after deciding whether RAID 10’s capacity and expansion trade-offs fit your plan.