Re: IBM Engineer
IBM midrange systems were inherently RAID 0 by default since time immemorial. Of course they never referred to it as "RAID 0", as they were doing it waaaay before the RAID concept was ever dreamed of. Besides, the bloody disks were anything but "Inexpensive". Still, good old IBM were trying to wring every drop of performance out of the hardware and forcing parallel disk channel use by deliberately fragmenting the data across all the available disks was a big win.
At least by the time the '400 came out, the backup systems had evolved to the state where you could feasibly backup the access paths (indices) as well as the data. Back in the days when tape bpi was measured in three digits, on the '38, even if you could restore everything, you were then looking at a week for the thing to rebuild its access paths before it would let anyone login.
Thus finding one day that some eejit had caught a sleeve on the handle of the isolator wallbox for the disk array, causing an immediate and ungraceful shutdown of all the disks and controllers, looked like being a bit of an issue. Good old System/38. I turned the disks back on and it picked up and carried on like nothing had happened. I've never seen another machine before or since that would do that.
Chief Operator: "You mean you can turn the disks off for a few minutes and then back on and it just carries on where it left off?"
Me: "Apparently. Let's not try it again though."