The Order to Buy a Home Server In: Drives Last, Not First

Buy in the order below and put the drive order at the end of it. Not because the drives matter least: they are 55 percent of the money and the only part of the build you will keep buying for as long as you own it. They come last in the sequence because every earlier decision, the capacity target, the RAID layout and the bay count, changes how many drives you need and what size they should be. Most first builds get both halves backwards: the enclosure is treated as the purchase and ordered first, and the drives are treated as an accessory added at checkout when the budget is already spent.

What order should I buy a home server in?

The buying order is a sequence of decisions, not a shopping list, and the money is committed in the reverse order to the shopping. Work down it.

1. Write down the usable capacity you need, and add a year of growth. Costs nothing, decides everything. Count what you actually hold today, in terabytes, then add what you expect to add in a year. If the honest answer is 6 TB, you are looking at a two-bay appliance and not a tower.
2. Choose the RAID layout, because it sets how much you must buy. Single parity returns (n-1) times the smallest drive, double parity returns (n-2). Getting 24 TB usable means 32 TB raw under RAID 5 and 48 TB raw under RAID 6, which is a $600 difference before you have chosen a brand. Run it in the RAID capacity calculator and read RAID levels explained if the formulas are new.
3. Pick the bay count from those two numbers, then the enclosure. Bay count is the decision, the brand is a detail. Compare enclosures in best 4-bay NAS, best 8-bay NAS and best NAS under $500.
4. Price the backup target before you price the drives. If you cannot afford a second copy of the data, you cannot afford the array. See best external drive for NAS backup.
5. Add the UPS. Around 6 percent, non-negotiable on a parity array. Sizes and waveform in best UPS for a NAS.
6. Order the drives. Now, when the count and capacity are settled, and split the order across two retailers so one production batch does not age identically in every bay. Picks and cost per terabyte in best NAS hard drives.
7. Cables, a switch if you need one, and memory. The last 2 percent, and the part that quietly causes the most trouble.

Why do the drives get 55 percent of the budget?

Because that is what they cost, and because they are the recurring purchase. Four NAS drives at $290 to $580 each is $1,160 to $2,320 against $400 to $750 for the four-bay box they go into. Move to eight bays and the ratio gets more extreme: eight drives is $2,000 to $4,000 against roughly $1,300 for the chassis, so the enclosure is a quarter to a third of the spend.

Then there is time. An enclosure runs for a decade with no attention. Drives are consumables: large-fleet annualised failure rates for modern 3.5 inch drives run about 1.0 to 1.6 percent per year and climb once drives pass roughly five years of service, so across a four-drive array over a decade you will buy drives more than once. Whatever you overspend on the box comes out of the drives twice, once now and once at replacement time. Put your own drive count and horizon into the drive failure probability calculator to see the shape of it.

A useful test before you order anything: divide the enclosure price by the drive spend. If the answer is above about 0.5 on a four-bay build, you are buying a box with money that belongs in capacity. Larger drives in fewer bays is almost always the better version of the same budget.

Why should the drives be the last thing you order?

Three specific ways ordering drives first costs money, all of them common.

• You buy the wrong count. Someone buys three drives, then discovers RAID 6 needs four, and the fourth drive arrives from a different batch at a different price anyway.
• You buy the wrong size. Every layout except Synology SHR truncates every member to the smallest drive, so a 4 TB drive bought in advance strands 8 TB on each of the 12 TB drives that join it later. That is the single most expensive mistake on this page.
• You buy the wrong drive entirely. Bay count decides whether rotational vibration sensors matter, and layout decides whether the workload rating does. Both are drive specifications you cannot evaluate until the chassis is chosen. Our guide to choosing NAS drives covers CMR against SMR, workload ratings and cost per terabyte.

Allocate the money to drives first. Spend it last. Those are not in conflict.

How do I split a $1,200, $2,500 or $5,000 budget?

Start from the percentages, then adjust. This is the planning target, not a rule that survives contact with real prices.

Shares sum to 100 percent of the hardware budget. Real parts never land exactly on them, and the two directions they drift are predictable: the enclosure share falls as the budget rises, because a four-bay box costs the same whether you fill it with 8 TB or 16 TB drives, and the backup share rises, because a larger array holds more data that needs a second copy.

What does each budget actually buy?

The same three budgets built from verified listing prices, with the share each line actually takes.

Parts in full: the $1,200 build is two Seagate IronWolf 8TB drives, a TerraMaster F2-425, a Seagate Expansion 8TB and an APC BE650G1. The $2,500 build swaps to four of the same drives, a Synology DS423, a WD Elements 20TB and a CyberPower CP1350PFCLCD. The $5,000 build runs four IronWolf Pro 16TB drives in a Synology DS925+ with two external backup targets, a TP-Link TL-SG108S-M2 and a Crucial 16GB SODIMM.

Read the share column rather than the totals. In all three, drives plus backup target is somewhere between 70 and 77 percent of the build. The enclosure share falls from 20% to 16% as the budget grows, which is the correct direction: bigger budgets buy capacity and protection, not a fancier box.

Note the layout choice on the $5,000 build. Four 16 TB drives under single parity would give 48 TB usable, and it runs SHR-2 for 32 TB instead. Sixteen terabyte members put a single-parity rebuild into the multi-day range, and the second parity drive is the correct purchase at that capacity. The full argument is in RAID 5 against RAID 6.

How much should I set aside for a backup target?

Twelve percent, and treat it as a hard floor rather than the last line to be cut. Parity survives a drive dying and nothing else. It does not survive a deleted folder, a ransomware run, a controller writing corrupted blocks, a theft or a fire, all of which take every member of the array at once because they are all in the same box.

Size the target against the data you actually hold, not against the array's capacity. A 24 TB array holding 9 TB of files needs a backup target above 9 TB with room to grow, not one above 24 TB. That is why the $2,500 build above pairs a 24 TB array with a 20 TB external and is not being reckless. Work out the real figure in the 3-2-1 sizing calculator and read the 3-2-1 backup rule guide for what the three copies should be.

A pair of large external drives, rotated so one is always at another address, is the cheapest honest version of this. Two WD Elements 18TB drives is a fraction of what a second NAS costs and provides the same offsite copy. Also budget a padded case for the drive that travels, because the offsite copy that gets dropped is not a copy.

What should I deliberately delay buying?

Every item below is genuinely useful, and every one of them is a worse use of the first budget than drives are. Delay is not the same as never.

Real network ceilings for reference: gigabit about 113 MB per second, 2.5GbE about 280 MB per second, 5GbE about 560 MB per second, 10GbE about 1,100 MB per second. One 7200 rpm 3.5 inch drive sustains 150 to 260 MB per second on its own, so a single drive already saturates gigabit.

The networking line deserves its own note, because it is where enthusiasm costs the most. Moving from gigabit to 2.5GbE is cheap and worthwhile: a five-port 2.5G switch is around $35, an USB 2.5GbE adapter around $26, and it roughly doubles what a single client sees. Moving to 10GbE means a 10Gb PCIe card in the client, a NAS with a 10GbE port and a switch that carries it, and four spinning drives will not fill that pipe on anything but a large sequential copy. Compare the two properly in 2.5GbE against 10GbE and pick a switch in best 2.5GbE switch for a NAS.

Flash is the other one. An SSD as a cache or app volume is a good purchase for the right workload and a poor one as bulk storage, because SATA SSD capacity costs several times what a hard drive does per terabyte. The full trade is in HDD against SSD for a NAS, and the picks are in best SSD for a NAS.

What if I cannot afford the whole thing at once?

Then buy it in this order, and stop wherever the money runs out. Each stage is a working system rather than a half-finished one.

• Stage one, about $600. A two-bay enclosure and two drives in RAID 1, or a mini PC and one drive. You have a file server. Picks in best budget NAS and best mini PC for a home server.
• Stage two, add about $280. An external drive as the backup target and a small UPS. Now you have protection rather than just storage. This stage is more important than more capacity.
• Stage three, add drives. Fill the remaining bays, or swap the existing drives up one capacity if the box is full. Capacity is the easiest thing to add later and the hardest to plan for.
• Stage four, everything in the delay table. By now you know which of them you actually need, which is the point of having waited.

The one thing not to stage is parity level. Changing from single to double parity later usually means rebuilding the array, which means copying everything off and back on. Decide that at the start.

What is the most common way this budget goes wrong?

An enclosure with more bays than the budget can fill. It happens because bay count is the headline specification and because "room to grow" sounds prudent, and it is the most expensive non-mistake in the category: an eight-bay chassis holding four drives is a four-bay NAS that cost roughly $600 more, and that $600 would have bought two more drives or moved the whole set up a capacity tier.

Second place goes to the opposite error, buying small drives to fill every bay today. Four 4 TB drives cost about $876.00 for 12 TB usable under single parity. Four 8 TB drives cost about $1,159.96 for 24 TB usable, $284 more for double the space, in the same bays, at the same power. Small drives never win on cost per terabyte, and they leave you with no bays left when you need more room.

Third is skipping the backup because the array has parity. That one is not a budgeting error, it is a category error, and it is the only one on this page that loses data. Before you build any of this, decide what you would do if the whole box vanished tonight. If the answer is not a sentence you can say out loud, start with the 3-2-1 backup rule and build the array afterwards.

Related reading

• Beginner home server setup, intermediate and expert parts lists at each budget
• Best NAS for home and best NAS for Plex for the enclosure shortlist by workload
• NAS or DIY home server for whether to buy the box at all
• Power draw calculator for the running cost of the build you just priced
• How to migrate to a bigger NAS for what stage three actually involves