Unraid Parity Explained: Single vs Dual Parity Protection
· Last verified July 2026
Parity is the single idea that makes Unraid tick, and it is also the one that trips up most first-time builders: it is not a backup, it is not a spare copy of your files, and it does not live on any one disk you can point at. This guide walks through what parity actually is, why one parity disk protects you and two protect you more, and how to decide which your build needs, written for someone comfortable swapping a hard drive but new to the maths underneath. If you want the pure definition of the operating system around it, our beginner guide to what Unraid is is the gentler starting point, and this post picks up where that one leaves off.
How Unraid parity actually works
The thing to understand first is that Unraid does not stripe your data across drives the way classic RAID 5 does. Each data disk keeps its own ordinary filesystem, so a given movie or photo lives whole on a single disk instead of being cut into fragments spread across the array. On top of those independent data disks you dedicate one drive purely to parity, which holds a running calculation across every data disk rather than any files of its own, and the manufacturer describes exactly this model in the Unraid storage management manual.
That calculation is an XOR, which is easier than it sounds: for the same block position on every disk, parity simply records whether the number of ones is odd or even. Because that single fact is enough to fill in any one missing value, a failed disk can always be reconstructed from what remains, and the same trick is what powers single-parity RAID in general, as the Wikipedia article on standard RAID levels lays out. A worked example makes it concrete. Say three data disks hold the bits 1, 0 and 1 at one position; there is an even number of ones, so parity stores 0. If the middle disk dies, Unraid lines up the survivors and the parity bit and XORs them back together: 1 XOR 1 XOR 0 gives 0, which is exactly the value that was lost. Do that across every block and the whole disk comes back.
This design has a consequence beginners tend to underappreciate: because files are not striped, only the disk you are actually reading has to spin up, and if you ever lose more disks than parity can cover you do not lose everything, since the surviving data disks are normal filesystems you can still read on their own. Parity buys you the ability to rebuild one whole disk from the others, and it does that without ever locking your files behind the array.
Why the parity drive must be your biggest disk
The XOR is computed down the full length of every data disk at once, position by position, so the parity disk needs enough room to cover the longest disk in the set. That is the whole reason for the rule you will see repeated everywhere: each parity disk must be at least as large as your biggest data disk. A 4 TB, an 8 TB and a 16 TB disk can share one Unraid array with no wasted space, but the parity has to be a 16 TB drive (or larger) so the calculation has somewhere to land across that full 16 TB range.
The practical catch shows up later, when you buy a disk bigger than your current parity. You cannot simply drop a shiny new 20 TB disk into a data slot if parity is only 16 TB; you either make the new disk the parity and demote the old parity to data, or you upgrade parity first and add the big disk afterwards. If you plan to run dual parity, remember that the size rule applies to both parity disks, not just one, which our Unraid dual-parity reference page spells out alongside the capacity math. The takeaway is simple: buy your largest drives as parity, because a parity disk can protect a smaller data disk but never a larger one.
Single parity vs dual parity: what each one survives
Single parity, one parity disk, lets the array survive any one drive failing at a time. That covers the everyday case (a disk dies, you replace it, Unraid rebuilds it from parity plus the survivors) and it is the configuration described on our Unraid single-parity reference page. What single parity cannot do is ride out a second disk failing while the first is still being rebuilt, and that gap is precisely what the second parity disk closes.
The important nuance is that the second parity disk is not a mirror or copy of the first. It is computed a different way, so instead of a plain odd-or-even count it uses a weighted calculation across the disks, the same principle RAID 6 uses to hold two independent parity values, as the standard RAID levels reference describes. Two parity values that are genuinely independent give you two separate equations, and two equations are exactly what you need to solve for two unknowns, which is why dual parity can rebuild any two disks that fail at the same moment. One parity disk, one missing value it can recover; two independent parity disks, two missing values. That is the entire intuition, and it is enough to reason about your own array without touching the algebra.
What parity costs you in usable capacity
Parity disks store the calculation, not your files, so they contribute nothing to usable space. The rule is refreshingly plain: usable capacity equals the sum of your data disks, and every parity disk is capacity you set aside for protection. An eight-bay box filled with 8 TB drives gives you 56 TB usable on single parity (seven data disks) and 48 TB usable on dual parity (six data disks), and the difference between those two numbers is the price of surviving that second failure.
Because Unraid handles mixed sizes so gracefully, the arithmetic is worth doing with your real drives rather than a tidy matched set. The Unraid Storage Calculator lets you drop in exactly the disks you own, flip between one and two parity, and read off usable capacity and the overhead in one place, and you can cross-check specific models in our NAS drive database before buying. The honest way to frame the trade is per array: one parity disk is one disk of overhead, two is two, and on a big box that second disk is a small percentage of the total for a meaningful jump in safety.
What happens during a rebuild, and why size matters
When a data disk fails and you drop in a replacement, Unraid rebuilds it by reading every other disk in the array, the remaining data disks and the parity disk together, and recomputing the missing disk block by block. That means a rebuild is not a quiet background job: it drives all your disks at full tilt for as long as it takes, and on large modern drives that window is long. Our RAID rebuild time guide walks through the arithmetic, but the short version is that a single big disk can keep the array in a degraded, working-hard state for the better part of a day or more.
Two things can turn that long window into data loss on a single-parity array. The first is a second disk failing outright while the rebuild is running, which single parity simply cannot absorb. The second is subtler: an unrecoverable read error, or URE, the statistical read fault that every large consumer disk hits every few terabytes, showing up on one of the surviving disks exactly when parity needs every block to be readable. On a single-parity rebuild, either event stops the recovery cold, and both get more likely the larger and more numerous your disks are, because there is simply more surface to read cleanly. That is the real argument for dual parity: it does not make disks fail less, it keeps the array recoverable when the worst moment (a rebuild) collides with a second problem. The bigger and busier your array, the more that insurance is worth.
Parity is not a backup: checks, sync and the write penalty
Parity keeps the array online through a disk failure, but it protects against exactly one thing: a disk dying. It does nothing about a file you deleted by mistake, a ransomware run that encrypts your shares, a power event that takes several disks at once, or the whole box walking out of the house. Redundancy and backup are different jobs, which is the point our RAID is not a backup guide hammers home, so keep separate 3-2-1 copies no matter how many parity disks you run.
To make sure parity is actually correct rather than merely present, Unraid runs a parity check, a scheduled pass that reads the whole array and confirms the parity math still matches the data, repairing it if you let it. It is the same family of maintenance as a filesystem scrub, and if you want the difference between the two spelled out our guide to scrubbing and bit rot covers it; a monthly check is a sensible default for most home arrays. Writing to the array carries a cost as well, because updating a file also means updating parity, which normally means reading the old data and old parity before writing the new values, so array writes are slower than a plain copy to one disk.
For the more technical crowd: that write penalty is not fixed. Unraid offers a reconstruct-write mode, usually called turbo write, that spins up every disk during a write so it can compute new parity directly from the other disks instead of doing the slower read-modify-write cycle. It trades the low idle power that makes Unraid attractive for meaningfully faster array writes, so it is a per-system choice rather than a default, and most home users leave it off and let an SSD cache pool absorb the write speed instead. That single knob is the only deep dive here; everything else in this post is beginner territory.
When one parity is enough, and when to run two
There is no universal answer, only a weighing of how likely a second failure is against how much a rebuild costs you, and four factors decide it. Read them together rather than in isolation, because a small young array with good backups can sit happily on single parity while a large aging one really should not.
- Drive count. More disks mean more chances for any one to fail, and a longer rebuild that reads all of them, so the case for dual parity climbs steadily past roughly six drives.
- Drive size. Larger disks take longer to rebuild and expose more surface to a URE during that rebuild, which is why 12 TB and up is the point most builders switch to two parity disks.
- Drive age and health. Disks that are older, running hot, or from the same batch tend to fail near each other, and a second parity disk is what carries you through overlapping failures.
- Backup posture. If parity is your only line of defence, buy the extra safety; if everything important is also backed up elsewhere, single parity plus a solid restore plan can be the rational choice even on a bigger array.
The clean rule that falls out of those four: run single parity on small, young, well-backed-up arrays, and step up to dual parity once you pass six drives, reach 12 TB disks, or simply value the array enough that a day-long degraded rebuild scares you. You can start on one parity and add a second later without rebuilding the array, so the decision is not permanent.
What an Unraid parity build needs
Recommendation
For most home builds the honest default is single parity to start: it protects against the failure you will actually see, keeps a disk of capacity for data, and takes nothing off the table since you can add a second parity disk later without rebuilding. Move to dual parity once your array crosses roughly six drives, once your disks reach 12 TB and up, or whenever the thought of a day-long degraded rebuild with no safety net makes you uneasy, because that is exactly the window where a second failure or a URE does the damage. Buy your biggest drives as parity so they can cover every data disk, run a monthly parity check to keep the math honest, and treat all of it as redundancy rather than backup: model the exact capacity in the Unraid Storage Calculator before you buy, then keep your 3-2-1 copies regardless of how many parity disks you land on.
Related articles
What Is Unraid? A Complete Beginner's Guide to the NAS OS
Unraid vs Synology 2026: Build It or Buy the Appliance?
RAID Rebuild Time: Why a Big Disk Keeps the Array Degraded for a Day