Path: utzoo!utgpu!jarvis.csri.toronto.edu!mailrus!tut.cis.ohio-state.edu!ucbvax!decwrl!labrea!rutgers!bellcore!texbell!uhnix1!sugar!ssd
From: ssd@sugar.uu.net (Scott Denham)
Newsgroups: comp.misc
Subject: Re: Looking for Computer Folklore
Summary: I've done that one, too!
Message-ID: <3438@sugar.uu.net>
Date: 15 Feb 89 19:51:25 GMT
References: <799@n8emr.UUCP> <7090@killer.DALLAS.TX.US>
Organization: Sugar Land Unix - Houston, TX
Lines: 33

In article <7090@killer.DALLAS.TX.US>, linimon@killer.DALLAS.TX.US (Mark Linimon) writes:
> In article <799@n8emr.UUCP>, lwv@n8emr.UUCP (Larry W. Virden) writes:
> > I was at a DECUS
> > conference about 6 yrs ago when a system programmer was laughing about
> > programming a Dec machine to seek around on a disk drive enough to cause the
> > cabinet to rock.
  
  
> I saw either this incident or a similar one -- firsthand.  PDP-11/20, Ampex
> add-on disk, "custom" (phew) controller, 1973.  Some late-night programming
> bums had tortured the diagnostic program to "full seek at switch register
> speed."  After that a quick binary search produced the resonant frequency
> of the machine.  At one point the electronics in the disk had crashed

 I did something similar (inadvertently) on and IBM 360/44 with 2314 disks
These were nasty, unreliable beasts, and we had a *terrible* problem with
seek errors.  One of the engineers decided we might be able to cure the 
problem by using a different type of oil in the head-actuator assebly;
it had some sort of oil-filled damper in it.  I wrote a program that 
alternately did random seeks across the whole pack as fast as possible
along with some "shaking" to attempt to heat things up.  We finally came
up with something that worked - and the program could be made to run for
long periods of time. For grins, somebody left it running all night one
Sunday night, and while the operator was either napping or out with his
girlfriend, 2 of the spare disk packs that had been sitting on TOP of the
disk cabinet walked off and took a quick 5 foot leap to the floor. One
survived.....the operations manager was not to happy about the other one,
which happened to contian a month's machine accounting data!!!!  
 
This same program, resurrected some years later for exercising some 
Memorex "washing machine" drives didn't do anything quite that drastic, 
but did manage to make 'em walk enough so a gap appeared between two
adjacent cabinets.