RAID Is Not a Backup: What Each Level Actually Protects Against

RAID is the acronym for Redundant Array of Independent Disks. RAID increases availability in the event of drive crash, not recoverability. RAID has nothing to do with backup.

RAID 1, RAID 5, and RAID 6 can recover from one or two disk failures and rebuild the disk, but that's all.

If any other failure happens, like ransomware, accidental deletion, or a power failure, RAID 1, 5, or 6 won't protect you.

Drive failures and rebuild risk

RAID 1, RAID 5, and RAID 6 can recover from a single disk failure - and RAID 6 can recover from two disk failures.

However, disk failures are not the only risks.

Failed RAID implementations, like those running on a controller, can crash and result in comprehensive failure. A full disk failure, hard disk failures, or a controller failure can happen, which RAID alone cannot recover.

Even if there is enough redundancy to let your RAID rebuild after one or two disk failures, you are still exposed to unknown data loss because of:

• Disk crashes
• Filesystem corruption
• Controller failure

That's why RAID is not a substitute for backup.

Rebuild time

During RAID rebuild, each RAID reads every sector on every other drive. It's a very safe state, but it takes a lot of time.

With 1 TB disks and a sequential write rate around 115 MB/s, a RAID 6 rebuild is approximately 2.5 hours, according to ActualTech Media. With 4 TB disks, this simple formula says a rebuild would take 10 hours.

But rebuild time can be much longer: a multi-terabyte hard drive can take many hours to days to rebuild, writes ServNet UK.

If another drive fails during rebuild you can easily lose all your data.

Rebuild time depends entirely on drive capacity. Arriving at any estimate involves:

• Drive capacity
• The RAID rebuild rate, which is throttled by ongoing workload
• The controller's processing power

What RAID does not survive

On top of lack of guarantees for specific failure modes, RAID cannot:

• Protect against filesystem or controller corruption
• Detect or correct silent data corruption.
• Protect against ransomware, accidental deletion, or site-level disaster

The traditional way to secure your data, even when it's on RAID arrays, is with actual backup.

Backup includes the whole range of techniques and mechanisms for keeping copies of your files aside from, independent of, and ideally in a different geographic location, as-is, which you can access in case of trouble. A snapshot, an incremental backup, a media vaulting process, all those are just backup.

If your files get encrypted, corrupted, or deleted, you need a backup to restore them from.

Break this down

Not all RAID arrays are the same:

• RAID 1 is a mirror or 2- drive RAID

- - - -

Each RAID has different degrees of redundancy and fault tolerance. RAID 1, 5, and 6 are the most commonly used. RAID 10 is also popular. But none of them is a backup - they're all forms of RAID, which at best can detect and recover from disk failure.

There's another important distinction. RAID protects you against individual drive failures. But that's only a fraction of the risks.

The truth is that RAID will not prevent catastrophic data loss as a whole. RAID is focused on keeping data available in the event of disk failure. But you also need to recover from ransomware, accidental deletion, logical corruption, or site-level disaster, because RAID doesn't do any of that.

Read more: ]

RAID is not a backup

RAID does not provide protection against these things:

• Accidental deletion
• Ransomware
• Catastrophic failure from fire, flood, or fall
• Silent data corruption
• Controller failure
• Filesystem corruption
• Data error

For that, you need real backup. Backup restores data, RAID gives you availability in the even of a disk failure.

Mirror math and deletion

RAID 1 is a mirror: all the data is written to two disks at once. But even in RAID 1, the failed disk will be replicated, its data overwritten by the healthy one's.

If a file is deleted or corrupted from the RAID set, it will be replicated and destroyed on the mirror.

This is why RAID alone doesn't prevent accidental deletion. For that, you need a backup solution that keeps a separate, protected copy.

Controller, filesystem, and corruption boundaries

RAID is about the disks, not the filesystem. If the filesystem gets corrupted, all your RAID has done is provide some protection against drive failure.

This happens more often than one would think. User errors. Program bugs. Ransomware can overwrite your data. A backing collection like a database can lose its index.

And RAID as such is about file blocks, bits, stripes, parity. It doesn't make your data error-free, just more available.

That's why it's important to implement RAID alongside: [TO VERIFY: a primary-source citation on the procurement, configuration, sizing, maintenance, and monitoring in some specific scenario of a filesytem-level corruption detector]

Snapshots and separate copies

Snapshots are point-in-time backups that let you recover the state of a disk at a certain time. They're great for quickly recovering from user error, but they don't provide full backup protection.

Separate copies are the key to real backup. This means copies kept separate from the primary storage, with enough copies that you can survive a site-level disaster.

By keeping separate copies, you can recover from ransomware, accidental deletion, or even a full-site disaster. This is the key to real backup.

Conclusion

While RAID is a great way to keep your data available in the event of a single drive's failure, it's not a substitute for backup.

For real data protection, you need a comprehensive backup solution that includes separate copies, multiple backups, and the ability to recover from all types of data loss.

RAID is high-availability technology for storage hardware that protects against drive failure, but it doesn't prevent human error, ransomware, or site-level disasters. For that, you need a separate, dedicated backup solution.