RAID 0 + Backup

Question

I have two 1GB hard (A) drives, and one 2TB hard drive (B). What I want to do is set up a RAID 0 on (A) and then back it up to (B).

To my understanding, this would be different than a RAID 5 set up, which I believe only backs up essentially "half" the data (if both drives fail, then restoring from backup won't work). I've also read that a RAID 5 is not desirable in situations where you only have 3 drives. Essentially, this would be like setting up a RAID 0 between the two (A) drives, and then a RAID 1 between the (A) and (B) drives.

My question is, is this a viable setup? What problems could I run into?

Answer 1

I assume that you meant that you have two 1TB drives, and one 2TB drive.

First off, you need to be fully aware that RAID is not a backup in the first place. Redundant RAID (which RAID 0 is not) gives you a chance (by no means a guarantee that you will be able to!) to handle failures within the storage system without the storage system failing completely. This gives you uptime, but not backups.

By setting the two smaller drives up in RAID 0, both those drives must be working for the data to be accessible. If one fails, then that array fails, and the data stored on it becomes inaccessible. Some small amounts of data may be recoverable, but this depends on the particulars of the RAID implementation and is nothing that can be counted on in the general case.

Depending on your I/O pattern, RAID 0 can be either faster or slower than a single drive. This depends on whether the drives can work in parallell responding to requests, or if they need to work sequentially. In the best case, you can get the combined performance of the drives; in the worst case, you will get approximately the performance of the slower drive. Normally, RAID 0 should give you a performance benefit.

By backing up to the larger drive, you are creating a separate storage set. This is different from RAID 1, because RAID 1 creates a single storage set. Borrowing this excellent answer's illustration of how this is different:

If you accidentally overwrite your PhD thesis with garbage, redundancy ensures that you have multiple copies of garbage, in case one gets bad. A backup ensures that you can restore your PhD thesis.

In this case, RAID 1 would provide redundancy, but your proposed scheme would provide a (single) backup.

If you want to do something like what you are proposing, this (having a separate storage set for a backup) is how you should be doing it.

There should be no particular gotchas with such a setup, beyond the fact that you are relying on both 1 TB drives to work for your data to be accessible. Assuming that you make and test backups regularly, this should not be a major problem in most situations where RAID 0 would be considered (which is non-critical data where downtime can be tolerated).

Also, as pointed out in the comments, the backup drive should be physically separate; ideally, disconnected and taken off site between backups. This protects you from many common-mode failures, such as for example power supply malfunctions, fire, theft, software errors, malware, and so on.