The home lab ecosystem has a bunch of fascinating utilities that, if properly configured, can take your setup to the next level. That said, many of these can be double-edged swords, and if you’re not fully aware of their intricacies, you could end up with broken expectations at best, and an inoperable server at worst.

RAID, or Redundant Array of Independent (or Inexpensive, depending on who you ask) Disks is one such facility that can get complicated the deeper you dive into its rabbit hole. Unfortunately, the endless recesses of the Internet are filled with a bunch of misconceptions about RAID setups, and it’s easy to fall for them when you’re a newcomer. So, here’s a quick piece on busting the most common myths about RAID arrays.

"RAID arrays can protect your files"

Backups are more important than RAID pools

Considering that RAID configurations tend to use multiple drives to provide striped, mirrored, or parity-based storage pools, you might be tempted to think that RAID setups can keep your data safe. Well, that is true to some extent, as mirrored pools maintain extra copies of your files, while setups involving parity data can help with easier recovery. However, the main purpose of RAID configs is to provide high availability and better fault tolerance for your storage pools should a drive fail mid-operation.

While certain setups can theoretically help you recover files, RAID is not meant to be a substitute for backups. In fact, you should always (and I can’t stress this enough) prioritize backups and even snapshots over RAID. Let’s say you accidentally overwrote a bunch of files. Even hybrid RAID setups would be useless here, and you wouldn't be able to recover the data. The same holds for silent data corruption, ransomware attacks, and any unforeseen accidents that may befall your storage server, which is why you should always schedule regular backups instead of just relying on RAID pools.

"All RAID levels provide some form of redundancy"

If we’re to go technical, parity setups aren’t the same as redundant drives

Redundancy is a common word when talking about RAID setups, though only a few options provide true copies of data. In the consumer space, RAID 1 and RAID 10 setups result in data being mirrored between drives. Let’s say you’ve got two drives in RAID 1. Should drive numero uno fail, the other, redundant HDD will ensure your NAS or home server will continue operating as usual.

Parity systems, like RAID 5 and RAID 6 (or rather, RAID-Z1 and RAID-Z2, if you follow ZFS’ naming scheme) can provide redundancy in the literal sense, but they’re quite different from mirrored setups. Rather than featuring multiple copies of your files, they contain parity bits that can rebuild your array should a drive (or two, in case of RAID 6/RAID-Z2) fail. Even though you can technically access your storage pool even with a dead drive, it will be in a degraded state, and you’ll have to add another drive and resilver the pool to avoid performance penalties.

Then there’s the outlier, RAID 0, which stripes data across multiple drives instead of using parity bits or mirrored setups. While it can improve write speeds, RAID 0 is terrible in almost all consumer-oriented scenarios, as it puts your data at greater risk than a non-RAID setup. Since RAID 0 is all about splitting files across multiple drives, you’ll end up losing your data if even a single drive ascends to tech heaven. Long story short, you’ll want to be extra cautious when choosing your RAID configuration, as not all of them feature the same level of redundancy as a mirrored array.

"You’ll need dedicated RAID cards"

Software RAID has come a long way

Ask data hoarders who have been in the NAS landscape for decades about their rigs, and you’ll probably hear a thing or two about RAID cards. Back in the day, hardware RAID was the only way to go about using this facility in your server without dealing with major performance and stability issues. That said, software RAID is a lot better than it used to be, with ZFS providing killer RAID support.

In fact, I’d recommend against using hardware RAID if you’re a complete beginner. Sure, dedicated RAID controllers can offer better performance for hybrid setups involving multiple drives. But for the average home labber and NAS user, software RAID provides solid performance and compatibility without requiring you to shell out more money on an expensive HBA. Then there’s the fact that recovering from a failed RAID controller can be a nightmare unless you manage to track down a similar card with the right firmware.

"You’re safe as long as you have a high-level RAID array"

You can have a bunch of drives failing at the same time

RAID setups (minus RAID 0) may provide solid fault tolerance, but the number of drive failures your array can survive depends entirely on your configuration. RAID 1 won’t result in data loss when a single drive goes out of commission, but you shouldn’t rule out the possibility of the second HDD kicking the bucket while you try to arrange for a replacement. The situation is similar for RAID 5, and if a second drive fails during the resilvering process, you’re out of luck.

Sure, the possibility of two drives perishing at the same time is slim, but it isn’t non-zero – especially if they belong to the same batch. As such, you’ll want to keep a hot spare (or two, if you’re on RAID 6 or hybrid setups) on you at all times. While it won’t save you if two drives fail simultaneously, it should help in quicker recovery when you’ve already got one dead HDD.

Above all else, make thorough plans before committing to a RAID setup

While you can migrate certain RAID pools to higher level configurations, the migration process is a royal pain, and certain file systems like ZFS prevent you from downscaling your RAID arrays. Plus, different RAID levels occupy different capacities for redundant data, so it’s always a good idea to plan the setup well in advance. That way, you won’t end up getting stuck with the wrong RAID configuration.