224TB of Disks, 200TB of Storage: Planning a Parity Array's Usable Capacity

A 224TB home server is not a flex about disk count, it is a statement about how much of that number you can actually fill. Parity is the part everyone underestimates: on a parity-based array, some of your drives hold no data at all, and the drives you do fill have to satisfy a rule about size that decides your upgrade path years in advance.

Getting the arithmetic right before you buy is worth an hour, because the alternative is discovering at 90 percent full that the next drive you can add will not increase usable space the way you assumed.

How usable capacity is actually calculated

Two rules sit underneath that table, and both catch people out. A parity drive must be at least as large as the largest data drive, and it contributes nothing to capacity. So a mixed pile of old drives gives you the sum of the data drives minus the parity, with every drive limited by the smallest common denominator of the parity rule.

Planning an array you will not regret

1. Step 1 Start from usable terabytes, not raw

   Decide the number you need to store, add 20 percent headroom for growth and snapshots, then work backwards to drive count. Buying raw capacity is how people end up with 240TB that can only hold 200.

2. Step 2 Decide the failure count before the drive count

   One parity survives one drive failing during a rebuild. Two parity survives a second failure while the array is already degraded, which is exactly when rebuild stress is highest.

3. Step 3 Buy the largest drives you can afford, fewer of them

   Fewer, larger drives mean fewer failure points and lower idle power for the same capacity. The trade is a longer rebuild per drive when one does fail.

4. Step 4 Check that the drives are not SMR

   Shingled drives behave badly in parity arrays, especially during a rebuild. The model number tells you; marketing pages often do not.

5. Step 5 Budget the rebuild time, not just the disk

   A rebuild reads every drive in the array. Plan on hours per terabyte, during which the array is unprotected and should not be doing anything else.

6. Step 6 Keep the irreplaceable set outside the array

   An array is a convenience layer, not a backup. The files you cannot re-download belong on a second medium and, ideally, in a second building.

Capacity traps worth knowing before you buy

• Raw and usable are different numbers, and vendors quote the first.
• Filesystem overhead and parity both eat into usable space, on top of the drive's decimal-versus-binary difference.
• A parity drive smaller than your largest data drive is not allowed; plan the parity last, not first.
• Adding a drive larger than your parity drive forces you to replace the parity too.
• An array filled past about 80 percent slows down and gets harder to rebuild.
• Snapshots, versioning and a recycle bin all consume capacity you budgeted for data.

Can I add a bigger drive later?

You can add a data drive larger than the existing ones, but usable space grows only by the size of the new drive, and your parity drive must be at least as large as it. That is why parity size decides the ceiling of the whole array.

Is one parity enough at home?

For replaceable media, usually yes. For a library you could not rebuild, two parity is the sane choice, because the dangerous window is the rebuild itself rather than normal operation.

Does the array count as a backup?

No. It survives a disk failure and does nothing about deletion, ransomware, a power surge or a fire. A backup is a second copy you have tested, on a medium that fails independently.

What about power consumption at this size?

It scales with the number of spinning drives, not with capacity, which is another argument for fewer large disks. Spin-down helps only if the array is idle for long stretches, and parity arrays rarely are.