This is a tale from the old days where hard drives were the size of washing machines.
The legend goes that there is a resonant frequency at which the drive, if periodically pushed, will rock back and forth and move around the room. Resonant frequencies being what they are, it's possible to write a program that causes the disk heads to seek at that frequency, and to start the box rocking, and eventually walking. Until it walks far enough to pull out its own plug. Operators come in in the morning and find the drive six feet from its original position, its cord forlornly dragging behind it. And the room was locked.
Or imagine the horror of being the graveyard shift operator and having the disk drive suddenly move towards you, methodically and inexorably, and apparently under its own power....
-------------------- What it all comes down to is, dyslexics have more nuf. Posts: 1090 | From: Area 51 | Registered: Feb 2000
| IP: Logged |
Having worked on both, I rather doubt the validity of the "walking drive" story. Those 3330-II's certaninly shook at times, but they couldn't really shake themselves out of position for a couple of reasons.
First of all, these drives were bolted together in groups, and I've never seen a group of less than 8 - and IBM isn't likely to have sold less than 2 to any customer. No one of them could move out of place without unbolting it from its neighbor - which usually only happened if the drive was down for maintenance. I believe this was true for other manufacturer's drives as well (ie, Sperry-Rand and CDC). For another, these things were connected to the mainframes with heavy bus and tag cables than run under the raised subfloor. These cables were thick and normally only as long as they needed to be to reach the DASD controller (which was usually located at the end of strings of 8 or 16) and they weren't easy to bend. For the drive to 'walk' out of position, it would have to pull it's bus and tag cable at a nearly 90 degree angle after only a few inches of movement. Thirdly, if that model drive was bumped too hard, it would destabilize the DASD platter stack probably to the point of a head crash. Any drive wandering across a floor under its own shaking is a drive that isn't going to be functioning for very long.
We certainly had interesting moments with the mod-2 at times. I'm convinced IBM didn't design these things with girls in mind, despite what is depicted in the first picture. The mod-1's and 2's had removable media, and the platter stack weighed about 60 pounds. We had shelves near the drives to store them, and I nearly dropped more than one stack that was placed on an upper shelf.
The most common failure was a 'head crash', where the soft metal heads or the arm holding them actually came in contact with the platter surface. There was usually never anything but fine shavings left of the heads after that, and the arm often left deep gouges in the platter surface. A drive that did that was usually down for a day or two while the tech vacuumed out the drive.
Much more rare was a platter or spindle failure. Imagine a platter stack, comprised of 8 16-inch metal plates weighing 60 pounds and spinning at high rpms (I don't remember the exact speed) coming unglued within the drive... It only happened once at the place I worked in all the time we had them (or any model DASD, for that matter) -- that drive never ran again.
As for the mod-1 3330 depicted in the first picture - that thing wasn't going anywhere it wasn't installed. There were smaller ones, but it was still a pretty heavy piece of machinery.
~Psihala (*The whoopie-cushion had nothing on a 3420 tape drive with dirty vacuum columns....)
Edited because the opening line sounded snarky which isn't what I was trying to be.... sorry, Alchemy...)