[Joe Towner]

Short version: RAID10 is a bunch of RAID1 pairs with a RAID0 laid down across the top of it. RAID01 is 2 RAID0 stripes that are then setup in a mirror.

In the event of a disk failure, in RAID10 the matching disk copies its data to the new drive. If a different drive fails during this rebuild, you’re fine, just as long as you have 1 disk from each RAID1 pair.

In a RAID01 situation, you’ve got 2 RAID0 volumes, so if a disk fails, once you replace it, it has to copy the entire data set from the other RAID0 volume. If a second disk fails, you may be in a world of hurt, as you are now hopeful the underlying software can assemble things from 2 broken RAID0 volumes.

In a single enclosure, it doesn’t matter if you do slot 1/2, 3/4 or 1/3, 2/4. It actually made a difference back in the day when you had external disk shelfs via external SCSI. You would mirror shelf 1 disk 1 with shelf 2 disk 1, shelf 1 disk 2 with shelf 2 disk 2, and so on – that way if a shelf had an issue (power, cooling, termination, etc) it wouldn’t take down the whole system. Lots and lots of 9.1gb SCSI & FC disks in those days, that’s how we rolled!