RAID 6: What It Is and How It Works

Running a storage array without fully understanding RAID 6 is a bit like building a house without knowing whether the foundations can survive more than one flood. You might get away with it for years until the moment you cannot. RAID 6 is one of the most resilient and widely deployed storage configurations in professional environments across Australia, and for good reason: it is the only standard RAID level that can absorb two simultaneous drive failures and keep operating without losing a single byte of data.

This guide goes further than the manufacturer documentation. It covers the architecture, the parity mathematics, the visual layout of healthy and degraded arrays, how hot spares integrate, common failure causes, a full RAID level comparison table, and, critically, what RAID data recovery actually looks like when things go wrong. Tables, visual array diagrams, and plain English throughout.

What Is RAID 6?

RAID (Redundant Array of Independent Disks) is a method of combining multiple physical drives into a single logical volume. Different RAID levels trade off speed, usable capacity, and fault tolerance in different ways. RAID 6 achieves fault tolerance through dual parity, two mathematically independent sets of parity information are calculated and written across the drives for every stripe of data. This is the defining feature separating RAID 6 from RAID 5, which writes only a single parity stripe and can therefore only survive one drive failure.

RAID 6 was developed in the 1980s and grew in strategic importance as drive capacities increased. Larger drives mean longer rebuild times, which create extended windows of vulnerability during which a second failure could be catastrophic. Dual parity was the engineering solution to that problem.

What Does a RAID 6 Array Look Like?

RAID 6 uses both data striping and parity, the same two principles underpinning RAID 5. The critical difference is the second parity stripe. Below is how a healthy five-drive RAID 6 array looks across four stripes:

Parity 1 (shown in blue) uses an XOR operation applied across all data blocks in that stripe, a standard algorithm consistent across all RAID controller vendors. Parity 2 uses a more complex Reed-Solomon polynomial calculation (Galois Field arithmetic). The exact Parity 2 implementation varies between manufacturers, a detail that has significant implications for RAID data recovery, as discussed later in this guide.

Notice how the parity blocks rotate across different drives on each stripe. This intentional rotation distributes the write load evenly so no single physical drive becomes a dedicated parity bottleneck. Every drive in the array shares both data and parity responsibilities equally.

Usable capacity formula: (n − 2) x smallest drive size. A six-drive array with 8TB drives gives you 4 x 8TB = 32TB of usable space. Two drives worth of space is always reserved for parity. Compare that to RAID 10 on the same six drives, which gives you only 3 x 8TB = 24TB, RAID 6 delivers 33% more usable capacity for the same hardware investment.

How Does Parity Work in a RAID 6 Array?

Parity is a mathematical feature that allows a RAID controller to reconstruct missing data entirely from what remains on surviving drives. Think of it as a coded redundant representation of your data that lives within the array itself, not on a separate backup drive.

When a drive fails, the controller does not simply show gaps where data is missing. Instead, it actively recalculates every missing block on the fly using parity, presenting a complete and accurate dataset to the operating system. Users experiencing a single drive failure typically notice nothing unusual, only a performance reduction while the array operates in degraded mode.

The XOR operation behind Parity 1 is elegant in its simplicity: if Data A = 1010, Data B = 0110, Data C = 1100, then Parity 1 = 0000. If Drive B fails, the controller solves: Data B = Data A XOR Data C XOR Parity 1 = 1010 XOR 1100 XOR 0000 = 0110. Exactly the original value, reconstructed without any copy of the lost drive. Parity 2 applies a higher-order polynomial that lets the controller solve for two unknowns simultaneously, enabling the two-failure tolerance that defines RAID 6.

What Happens When One Drive Fails?

When a single drive goes offline, the array enters degraded mode. The RAID controller continues serving data normally, recalculating missing blocks from the surviving data blocks and Parity 1. From the operating system’s perspective, the volume remains fully accessible.

In the example above, Drive 2 has failed. For Stripe 1, the controller combines Data 1, Data 3, and Parity 1 to reconstruct the missing Data 2 value. For Stripe 2, it uses Data 4, Data 6, and Parity 2. Stripes where no data is missing from the failed drive require no reconstruction at all.

Degraded mode is not a crisis, but it is a warning requiring immediate action. Every read touching data that would have resided on the failed drive now requires a parity recalculation, increasing I/O load on surviving drives and creating the conditions for a second failure. Experienced administrators treat the first failure as an alarm, not the emergency itself, but the signal that one is coming if action is not taken promptly.

What Happens When Two Drives Fail?

This is where RAID 6 decisively outperforms RAID 5 and every other single-parity configuration. With two drives failed, the controller draws on both parity stripes simultaneously, solving a system of two equations to reconstruct two missing data values per stripe.

In the example above, Drives 2 and 4 have both failed. For Stripe 1, the controller uses Data 1, Data 3, and Parity 2 to reconstruct both missing values. For Stripe 2, it uses Data 4, Data 6, and Parity 1. The Galois Field arithmetic embedded in the controller firmware makes this possible, the controller solves the two-unknown system with precision, and the operating system continues to see clean, complete data.

The moment a third drive fails, the system has more unknowns than parity equations can solve, and the array cannot self-recover. This is the point where professional RAID data recovery becomes the only viable path forward, and where the expertise of the recovery team matters enormously.

How Does a Hot Spare Integrate with RAID 6?

A hot spare is a drive installed in the enclosure that carries no data during normal operation. It simply waits. The moment the RAID controller detects a drive failure, it immediately begins rebuilding the lost data onto the hot spare, automatically, without human intervention.

Once a failure triggers the hot spare:

The failed drive can then be physically removed and replaced. The replacement becomes either the new hot spare or is incorporated into the array directly. The original data is now fully protected again, the entire recovery happened without any administrator manually initiating a rebuild.

Why does the hot spare matter so much?

Modern high-capacity drives (16TB, 20TB, 24TB) can take 24 to 72 hours to rebuild on a busy production array, sometimes longer. During that entire window, the array is in degraded mode and cannot survive a second failure. Drives from the same manufacturing batch often share similar wear curves, making a contemporaneous second failure more probable than intuition suggests. A hot spare eliminates that window entirely. For arrays with 8 or more drives, at least one hot spare is not optional, it is essential.

RAID Level Comparison: How RAID 6 Stacks Up

No RAID level is universally superior. Each involves deliberate engineering trade-offs. Here is a complete comparison of the configurations most commonly deployed in Australian business environments:

RAID 6 Advantages and Limitations, An Honest Assessment

What Causes RAID 6 Arrays to Fail?

Despite its resilience, RAID 6 is not immune to failure. Understanding failure modes is essential both for prevention and for knowing when professional recovery is needed.

Choosing Hardware for a Reliable RAID 6 Deployment

Hardware selection has a direct bearing on both ongoing performance and recoverability when things go wrong. These are the components that matter most:

If your RAID 6 array underpins business-critical operations, establishing a pre-arranged relationship with a recovery specialist before a failure occurs is one of the most pragmatic investments you can make. Recovery Squad offers priority service agreements for Australian businesses that need guaranteed response times and pre-negotiated recovery terms, because when a production array fails, every hour of downtime carries a real financial cost.

Summary

RAID 6 is one of the most thoughtfully engineered trade-offs in enterprise storage technology. Its dual-parity architecture delivers a level of hardware fault tolerance that no comparable single-parity configuration can match, all while maintaining the strong read performance and capacity efficiency that make it practical for real-world deployments at scale.

The trade-offs are genuine: write performance is slower, rebuild times on modern high-capacity drives can stretch across multiple days, and the proprietary Parity 2 algorithm adds real complexity to the recovery process. But for organisations that understand these trade-offs and plan around them, with hot spares, battery-backed cache, regular health monitoring, and a professional RAID data recovery partner on standby, RAID 6 remains one of the most dependable storage foundations available for business-critical data.

If your RAID 6 array has failed, is running in degraded mode, or you have questions about your current configuration, Recovery Squad is available for a free, no-obligation assessment. Our engineers carry hands-on experience with every major RAID controller platform and can advise on recovery options, configuration review, and preventive measures before the next failure arrives, not after.