All disk drives suck, and one must assume that all disk drives will fail sooner or later. One also needs to assume that software errors will eventually corrupt some important data you need, and you will only find out much later. The only safe solution for data protection is on-site (for rapid recovery) and off-site (for protection against fire, etc) storage. Simple clones are inadequate, as they will efficiently copy the errors on your primary to the secondary. RAID can protect you against drive failures, but will happily write software-corrupted data across all the drives.
I happen to use Apple Time Machine for my on-site backup, and Crashplan for off-site. They run independently and make frequent copies of the primary. They keep older data indefinitely (or at least for a long time), allowing you to recover from corrupted data to any earlier date. But there are many acceptable solutions. They key is diverse locations, diverse software stacks, independent processes. You dramatically reduce your risk of loss in this way.
If you're really paranoid (and you should be), take a periodic full copy of your data (like once a year) and keep it at a different location.