How hard would it be to build a PC system with redundant harddisk drives where they functioned as follows:
Each disk write would write to 2 or 3 different active physical drives.
Each disk read would read from the next active physical drive, as in a round-robin fashion.
Each active drive could have a failing or restoring attribute.
A failing drive can be removed and replaced with a new drive and change its attribute to restoring. The new drive should be at least as large as the smallest active drive.
When the new drive has been restored by a low priority background process, then the restoring attribute can be removed, and this drive can then be included in the available drives that can be read.
If you used 2 active disk drives, and a failed drive could be restored before the next drive failed, then your storage system would never be unavailable due to a hard disk drive failure. More active drives insure more protection.