I am a long-time fan of Linux software RAID. But in reading an online discussion which happened to mention variations in actual sizes of storage devices (disks or flash) of supposedly identical capacity, I realized that there is yet another advantage of Linux software RAID: the fact that it can build arrays out of disk partitions, not necessarily entire disks. This means you can create the partitions to be a standard size, slightly smaller than the maximum capacity of the disk. (Each disk will have just one partition on it, the rest of the disk being unused.) Then if you have to replace a failed disk, you can create a new partition of the same size on the replacement. Otherwise, if you use the entire disk, you run the risk that the new disk is a bit smaller than the old one, and trying to add it to the RAID container may fail.