Why you shouldn't use RAID

RAID is used for 3 reasons

• Uptime  - If you need it to be always up all the time, then you get the RAID so you can hot swap the disk and incur no downtime.

• Speed - More disks minus the parity disk = more speed of data transfer for multiple simultaneous users. (mainly for read - write speeds will be slower than read speeds because of parity calculations.)

• Space - If you need more space than a single disk can hold then you get a RAID to increase the storage.  

RAID is overrated for a home user

If you don't really need these 3 features, a single disk is sufficient for storage.

For the vast majority of home use, a single disk is sufficient, and much cheaper.  You can then just buy a 2nd disk for a backup.  You can even buy a 3rd disk for less than a RAID NAS unit with 2 disks and you'll have 2 backups.  If your primary disk fails, you can quickly swap in one of the backup disks.

Each additional disk adds to the probability of failure.  The more disks you have, the more likely you'll encounter a disk failure.

For easier calculations, let's say that the probability of failure of a single disk is 10%. The probability of failure for 2 disks would be calculated from the probability that no disks fail.

90% x 90% = 81% of no disk failures.

100% -81% = 19% of any disk failing in a 2 disk system.

For 3 disks:

90%x90%x90% = 72.9% of no disk failures

100%-72.9% = 27.1% for any 3 disks failing.

When you calculate the failure rates for smaller probabilities such as 0.01%, you can aprroximate the failure rates by just multiplying by the total number of disks. More disks = increased failure rates. 10 disk = approximaely 10x the failure probability of a single disk.

RAID has more parts to it. In essence, more parts means more points of failure. It needs to be managed. It's not a "set it, and forget it" type of appliance. Unless you need the Space, Speed, or Uptime, it's unnecessary for a typical home user or small business.

RAID also needs to be watched. 

In a server environment. the sysdamins are paying attention to their servers and replace failed disks immediately.  They usually have a spare and order new ones right away.  This is what makes RAID "reliable" in server environments.  Someone's paying attention to maintenance.

I've encountered numerous home users and small businesses that buy RAID and neglect backups because they mistake RAID for backups.  They don't pay attention to the failed disk or will muddle through with the failed disk because, "it's still running." They're then caught by surprise when the next disk fail and they have no backups.  RAID only works if you're managing it.  I've encountered too many people that would keep running their NAS RAID with a failed disk with no plans to immediately replace it.

Another issue is that most people buy their RAID disk in a single batch.  That increases the probability that a 2nd disk will fail within 2 weeks of a first failed disk.  You really have to have a spare disk on hand for immediate relpacement and be ready to buy a replacement for that spare that you just used up.

P.S.

I would also say that half the RAID 1 boot disk setups I've ever encountered were faulty and/or not true RAID 1.  They didn't properly duplicate the boot sector to the 2nd disk so it wouldn't boot when the first disk failed.  Everything but the boot sector was duplicated, or the card wouldn't boot the 2nd disk, even if it got moved.  They had "built-in" junk RAID and shouldn't have been made into a RAID.  Nobody knows until that first disk fails because nobody tested them when they were first created.

I test new systems now by pulling the disk to make sure it boots when either disk are removed.  I still periodically encounter other people's setups where the boot RAID just didn't work like it should.  If it doesn't work, there's no point in making a RAID 1 boot.  Just make it a single disk so you don't have false security.  You need to test that. Having backups is just much simpler for many small businesses and home users.