I have had too many data recovery customers start our conversation with the title of this post. “I have a RAID so my data is backed up.” It’s a misconception that has cost them thousands of dollars.
RAID or redundant array of independent disks, is a group of drives that act as a single drive. There are a variety of RAID configurations. Some protect against hardware failure, some are even more prone to hardware failure than a regular single drive.
The most common RAID configurations are RAID 0, 1 and 5. There are many others, but I’ve rarely seen them used so I’m only going to talk about these three configurations.
RAID 0 takes multiple drives and stripes them to make one big drive that uses all drives to write faster. RAID 0 is super fast and is great when you need a super fast drive. How is it fast? Take a 2 drive RAID 0. If you’re writing a file with one drive, there’s one drive head to write the data. If you’re striped, there are two heads and it writes half of each file on each drive. This makes it twice as fast as a single drive. Unfortunately, this is the worst RAID from a data protection standpoint. Because all files are written in pieces on each drive, if just one drive fails, then all data could be lost,and it’s a very expensive recovery. RAID 0 means “zero data protection”.
This is often called “mirrored” and it does provide some data protection, but only protection against hardware failure. RAID 1 is made up of two identically sized drives. When you save a file, it writes a copy of the file to each drive so you have two copies. So it’s an automatic, instant backup, right? Not so fast. RAID 1 will protect you if one of your drives experiences a hardware failure. When it does, your other drive in the RAID still has all your data, which is good and saves you a trip to Mike’s. However, there are other failures that can still destroy ALL your data, even with two drives.
Because RAID 1 is mirrored, everything that happens on one drive happens on the other drive, instantly. So if you throw a file, or all your files in the trash, and empty the trash, they are deleted on both drives. Bye, bye, data. Hello, Mike’s data recovery. If your drive experiences a software failure where the drive gets confused and starts writing data over existing files, it does it on both drives. Again, bye, bye, data. RAID 1 protects you from some hardware failures but not software or human failures
This RAID needs at least three drives and provides the most data protection but it is still not a backup. You can still lose all your data. This happened to one of our customers who ran his entire business on a RAID 5 and could have lost it all because he had no backup. I’ll get to that in a moment. RAID 5 requires at least three drives. Let’s use the example of a three drive RAID 5 to explain how it works. RAID 5 writes 2/3 of your data to the first drive and simultaneously writes a 1/3 backup to each of the other 2 drives. The result is a backup of your data to the other drives. This is often used for servers because if one drive fails you can remove it, insert a new drive without turning off the server and the RAID automatically recreates the failed drive from the backup data on the other drives. Really cool.
But although your data is written in two places it suffers from the same risks as RAID 1. If you throw out your data, it’s gone. If the directory gets messed up, it can still overwrite your files in both places. Once again, bye, bye, data.
Now back to the story of that nice customer of ours. He’d been running his business on a RAID 5. He called me one Sunday in a panic. His drive wasn’t showing up. So the first thing I asked him was: “Do you have a backup?” His reply was just what I never want to hear: “I don’t need a backup, it’s a RAID.” My heart sunk. He brought it in and he had big problems. One of his drives had failed earlier and he didn’t notice it. Now a second drive had failed so he had no backup. Then when his drive didn’t show up, he restarted the RAID and the RAID brought up the old drive that had failed a while ago and had old data on it. Now the RAID was really confused and started rebuilding old data on top of new data. Bad. Bad. Bad. We recovered many but not all of the files he needed to run his business, explained RAIDs and how he needed an independent backup.
One final note about RAIDs: they’re complicated. They involve more drives and complex schemes for storing data on those drives. Because of that they’re more expensive to recover. RAID recoveries are often two to five TIMES more expensive than regular single drive recoveries. So instead of a RAID “having” a backup and protecting you, it’s going to cost you even more to recover that data.
So what’s the moral of the story with having a RAID? It’s the same moral as any hard drive: you need an independent backup. You need a copy of your data in a separate place, whether that’s in the cloud or on an external hard drive, or both.