NAS Drive Buying Checklist

Start with the workload, not the drive label. A NAS used for file storage, media libraries, surveillance recording, virtual machines and backups can place very different demands on its drives.

The short version:

• Choose capacity from your data growth and redundancy plan, not the raw number printed on the box.
• Prefer CMR HDDs for active RAID or ZFS storage unless the NAS and workload explicitly suit SMR.
• Match the drive’s interface, form factor and sector format to the NAS.
• Treat endurance, workload rating, noise and power as practical operating limits—not just marketing numbers.
• RAID improves availability and can protect against a drive failure. It is not a backup.
• Confirm warranty, replacement procedure and compatibility before ordering.

Use this checklist before you Browse drives.

1. Define the workload first

Write down what the NAS will actually do. A drive that is suitable for sequential media storage may be a poor choice for databases or virtual machines.

Workload	Main drive considerations
Documents, photos and general file sharing	Capacity, reliability, random access and backup strategy
Plex or other media libraries	Capacity and sequential reads; transcoding is usually a separate CPU/GPU concern
PC and Mac backups	Capacity, concurrent users and snapshot/version retention
Surveillance recording	Sustained writes, continuous duty, retention period and camera count
Virtual machines and containers	Random I/O latency, SSD suitability, endurance and memory
Databases or active application data	Low latency, consistent write behavior and endurance
Archive or rarely changed data	Capacity, cost per usable TB, power and offline backup
SSD cache	Cache endurance, workload fit, failure behavior and whether the NAS benefits from caching

A NAS may perform several of these jobs at once. If so, size and select drives for the most demanding important workload, not merely the largest one.

2. Calculate capacity from usable storage

Start with the data you need to store, then add room for growth, versions, snapshots and temporary files.

A simple planning formula is:

Required raw capacity = current data × (1 + growth allowance) + backup/versioning space

Do not confuse decimal drive capacity with the capacity shown by some operating systems. Drive manufacturers generally label capacity in decimal terabytes, while software may display a different unit convention.

Account for RAID or storage-pool overhead

Raw capacity is not the same as usable capacity. The result depends on the RAID level, filesystem, pool layout, reserved space and NAS implementation.

Useful approximations:

• RAID 1 or a two-drive mirror: usable capacity is roughly the capacity of one drive.
• RAID 10: usable capacity is roughly half of the combined raw capacity.
• Single-parity RAID or RAIDZ1: usable capacity is often approximately (number of drives - 1) × smallest drive capacity, before filesystem and pool overhead.
• Double-parity RAID or RAIDZ2: usable capacity is often approximately (number of drives - 2) × smallest drive capacity.
• JBOD or spanning: capacity may be close to the sum of the drives, but fault behavior and recoverability differ from redundant RAID.
• Mixed-size arrays: many RAID systems use the smallest drive’s capacity as the common baseline, leaving part of larger drives unused.

These are planning estimates, not guarantees. Check the NAS manufacturer’s calculator and the filesystem’s recommended free-space margin.

Worked example

Suppose you have:

• 18 TB of current data
• 30% expected growth
• 4 TB reserved for snapshots, file versions and working space
• A single-parity array using four equal-capacity drives

Planning demand:

18 TB × 1.30 + 4 TB = 27.4 TB usable

A four-drive single-parity layout would need roughly:

4 drives × 10 TB = 40 TB raw

Its approximate pre-overhead usable capacity would be:

(4 - 1) × 10 TB = 30 TB

That may be enough on paper, but the final decision must also allow for filesystem overhead, pool free space and future growth. If the array will soon exceed its comfortable operating level, buying larger drives or choosing a different layout may be more sensible than filling the pool immediately.

3. Choose HDD or SSD

HDDs

Hard disk drives usually provide more capacity per dollar for bulk NAS storage. They are a common fit for:

• Large media libraries
• Backups
• Documents and photos
• Surveillance retention
• General file sharing

Trade-offs include mechanical noise, vibration, higher access latency and power use while active.

SSDs

SSDs can reduce latency and noise and may help with workloads involving many small, random operations. Consider them for:

• Virtual machines
• Databases
• Containers
• Metadata-heavy applications
• Active project files
• A dedicated high-performance storage pool

An SSD is not automatically faster for every NAS task. A network link, NAS processor, protocol overhead or workload pattern may be the limiting factor. For sequential media streaming, a properly configured HDD array may already provide enough storage performance.

If you are selecting solid-state storage, compare the relevant endurance and power data rather than assuming that any consumer SSD is suitable for continuous NAS duty. You can Browse internal SSDs separately from hard drives.

4. CMR versus SMR

CMR

Conventional Magnetic Recording writes data to distinct tracks. CMR drives are generally the safer default for active NAS arrays because their write behavior is more predictable, especially during:

• Frequent random writes
• RAID rebuilds or resilvers
• Large synchronization operations
• Multi-user file activity
• Surveillance recording combined with other workloads

SMR

Shingled Magnetic Recording overlaps tracks to increase areal density. SMR can be reasonable for some sequential, mostly-read or archive workloads, but sustained or random write activity can trigger internal data reorganization. That may cause unpredictable write performance.

Potential concerns are more significant when an SMR drive is used in:

• A busy RAID array
• A ZFS pool with frequent writes
• A surveillance system recording continuously
• A workload with repeated file changes
• A rebuild or resilver after another drive fails

SMR is not universally unusable, and CMR is not a guarantee of a perfect NAS experience. The important question is whether the drive’s recording method matches the NAS software, RAID behavior and workload. If the product page does not clearly identify CMR or SMR, check the manufacturer’s technical documentation before buying.

5. Check the interface and physical compatibility

SATA, SAS and NVMe

• SATA HDDs and SATA SSDs: Common in home and small-business NAS systems. A SATA drive generally negotiates down to the speed supported by the NAS and controller.
• SAS drives: Require SAS-capable controllers and backplanes. A typical SATA-only NAS should not be assumed to support them.
• NVMe SSDs: Use a different connector and protocol from SATA. The NAS must provide compatible M.2 slots, PCIe slots, adapters and software support.

A faster interface does not automatically make the NAS faster. The complete path includes the drive, controller, PCIe or SATA link, NAS CPU, memory, network and application.

Confirm the physical details

Before ordering, verify:

• 3.5-inch or 2.5-inch bay requirements
• Drive height and tray clearance
• SATA power and data connections
• M.2 keying and supported length for NVMe drives
• Hot-swap support
• Maximum supported drive capacity
• Whether the NAS requires a vendor compatibility list
• Whether the selected drive is supported in the intended RAID, pool or cache role

Do not infer compatibility solely from a matching connector. Some NAS systems restrict drive models, firmware, sector formats or supported capacities.

6. Evaluate endurance and workload rating

HDD endurance

For HDDs, look for the manufacturer’s stated workload rate, annualized workload limit or equivalent documentation when available. This can include reads, writes and other transferred data, depending on how the manufacturer defines it.

Also consider:

• Continuous-duty operation
• Rated error-recovery behavior
• Vibration tolerance
• RAID or multi-drive enclosure suitability
• The expected number of active hours
• Rebuild frequency and duration

Do not compare different brands’ workload figures as if they were guaranteed to use identical test methods. Treat them as useful indicators, then read the definitions.

SSD endurance

SSD endurance is commonly expressed using a TBW or DWPD figure.

• TBW: Total bytes written over the stated warranty or rating period.
• DWPD: Drive writes per day, usually based on a defined warranty period and drive capacity.

A simplified conversion is:

Approximate TBW = drive capacity × DWPD × 365 × years

The exact result depends on the manufacturer’s definition and warranty terms. NAND type, overprovisioning, write amplification, temperature and workload all affect real endurance.

For an SSD cache, estimate writes from the NAS’s actual workload if possible. A cache can receive more write traffic than the files users think they are creating because of metadata, eviction, synchronization and filesystem behavior.

7. Consider noise and vibration

NAS noise comes from both the drives and the enclosure. Drive-level factors include:

• Idle acoustics
• Read/write seek noise
• Startup noise
• Vibration
• Number of drives operating at once
• Drive spin-up behavior

Manufacturers may publish acoustic figures, but measurements can vary with test conditions. A quiet single-drive system can become noticeably louder with several mechanical drives starting, seeking or vibrating together.

Practical consequences:

• A bedroom or office NAS may justify quieter drives or SSDs.
• A rack or utility-room NAS may make noise less important.
• Vibration can affect comfort and potentially drive behavior in multi-drive enclosures.
• Do not disable scheduled spin-up or standby features without considering backup windows, drive wear, access latency and the NAS’s power-management behavior.

8. Compare power use

Power matters for a NAS that runs continuously. Compare the manufacturer’s available figures for:

• Idle or low-power operation
• Active reads and writes
• Startup or spin-up demand
• Sleep or standby
• SSD active and idle power
• Operating temperature

A drive’s individual wattage is only part of the system total. Add the NAS chassis, fans, power supply losses, networking equipment and any attached expansion unit.

A useful annual energy estimate is:

Annual energy (kWh) = average power (watts) × 24 × 365 / 1,000

Example:

35 W × 24 × 365 / 1,000 = 306.6 kWh per year

Use your local electricity rate to estimate operating cost:

Annual electricity cost = annual kWh × electricity price per kWh

Actual power depends on workload, drive count, standby policy and environment, so treat catalog figures as comparison points rather than promises.

9. Compare cost per usable TB

Raw price per terabyte can hide the cost of redundancy.

Raw cost per TB = total drive cost / total raw capacity

For a redundant array:

Usable cost per TB = total drive cost / estimated usable capacity

Example:

• Four drives at 10 TB each
• Total drive cost: $800
• Approximate single-parity usable capacity: 30 TB before overhead

$800 / 30 TB = approximately $26.67 per usable TB

This excludes the NAS, replacement drives, backup storage, electricity and any filesystem overhead. If you need a separate backup copy, calculate its cost too. A cheaper redundant array is not necessarily a cheaper protected storage system if the backup is omitted.

10. Understand RAID implications

RAID changes how the NAS handles drive failure; it does not make the data permanently safe.

Depending on the layout, RAID can provide:

• Continued operation after one or more drive failures
• Better read performance in some workloads
• Capacity efficiency compared with keeping separate full copies

It can also introduce:

• Rebuild or resilver time
• Performance impact during recovery
• Additional stress on surviving drives
• Capacity loss to parity or mirroring
• A need for compatible replacement drives
• Risk from controller, filesystem, user or configuration failures

Use drives that are appropriate for the intended RAID or storage-pool implementation. For ZFS, consider the vdev layout, parity level, drive replacement process and the workload’s write pattern. Do not assume that adding more drives later will automatically expand an existing pool; expansion behavior depends on the NAS software and layout.

RAID is not backup

RAID does not protect against:

• Accidental deletion
• Ransomware
• File corruption replicated across the array
• Theft or fire
• NAS failure
• Incorrect synchronization
• A damaged or overwritten filesystem

Keep at least one independently stored backup of important data. Test that backup by restoring files, not merely by checking that a backup job reports success.

11. Check warranty and replacement terms

Before purchase, read the actual warranty conditions for the exact model and region. Check:

• Warranty length
• Required proof of purchase
• Advanced replacement availability, if offered
• Return shipping responsibility
• Coverage for failed sectors or degraded health
• Whether use in a NAS or continuous-duty environment is covered
• Data recovery exclusions
• Secure-drive disposal or replacement procedures

A warranty replaces a failed drive; it generally does not recover your data. Keep the backup independent of the warranty process.

Final NAS drive-selection checklist

Use this list at checkout:

• [ ] I documented the NAS workload and the most demanding application.
• [ ] I calculated current data, growth, snapshots, versions and free-space needs.
• [ ] I calculated usable capacity after RAID, mirroring or parity.
• [ ] I included the cost and capacity of a separate backup.
• [ ] I confirmed the drive is CMR or SMR and that the recording method suits the workload.
• [ ] I selected HDD, SATA SSD or NVMe based on latency, capacity and workload—not interface speed alone.
• [ ] I verified form factor, height, connector, sector format and NAS compatibility.
• [ ] I checked the maximum supported drive capacity and pool expansion rules.
• [ ] I reviewed workload rate, endurance or TBW/DWPD figures where relevant.
• [ ] I considered seek noise, vibration, startup noise and the NAS location.
• [ ] I estimated power use and annual energy cost.
• [ ] I compared cost per usable TB, not only cost per raw TB.
• [ ] I checked warranty coverage and replacement terms in my region.
• [ ] I have a tested backup that does not depend on the RAID array.
• [ ] I will keep replacement drives and rebuild considerations in mind before filling the pool.

When exact compatibility, noise, power or endurance data is not published, do not guess. Check the drive datasheet, the NAS compatibility list and the storage-pool documentation before committing to a purchase.