Redundancy is at it's simplest a means in which a hard drive (storage device) can fail but the data is still recoverable.
This is done using a mirror drive which is a second hard drive of the same size that the data gets written to as well as the primary drive. It is a direct, one-for-one copy of the first drive and so if the first drive fails the second has the same data and can be used to keep the system running and the data recoverable.
Another way involves a technique called striping in which you sacrifice the capacity of one of the drives to keep the system alive.
Usually this is done in a minimal of a 5-drive configuration (say 5 TB hard drives) and you have the capacity of three of them. One TB being sacrificed for data redundancy and the other being an empty drive that is built up with the missing data when one of the other drives fails.