Want to read Slashdot from your mobile device? Point it at m.slashdot.org and keep reading!


Forgot your password?
It's funny.  Laugh. Hardware Science

Physical Hard-Disk Data Arrangements and Drive Failures? 46

Tadau asks: "Knowing not much of the low-level and molecular aspects of a hard drive platter, I'm wondering if it is possible to cause a weight change/imbalance on a hard drive platter by say writing solid 1's to approximately 1/2 of a side of the platter? If there is a weight change, then could that attribute to drive vibrations by an ever-so-slightly unbalanced platter, which may result in an eventual drive failure?"
This discussion has been archived. No new comments can be posted.

Physical Hard-Disk Data Arrangements and Drive Failures?

Comments Filter:
  • I mean, seriously, does a one weigh more than a zero?
    • I forget, but doesn't an electron have no mass? :)
    • No. The ones and zeros are produced by differences in magnetic fields, but magnetic fields are not caused by electron imbalances-- they're caused by electron motion. For hard disks, this is the orbit of the electron around the nucleus of its atom. The alignment of electron orbits is what produces a 1 or zero, but no matter what, the electrons stay where they belong.
  • c'mon guys, can't you create an April Fools thread with a bit more credibility... If I see this same story in this month's GamePro I'm gonna want my money back. 'from the theoretical-weight-of-information dept.' sheesh! let's just give it away from the start. oh well, better luck next year.
  • by nytes ( 231372 ) on Tuesday April 01, 2003 @12:58PM (#5638473) Homepage
    The answer is yes. If you write all 1's to one side of a drive, and all 0's to the other side, the drive will eventually fail.

  • Simple answer (Score:5, Interesting)

    by missing000 ( 602285 ) on Tuesday April 01, 2003 @01:00PM (#5638499)

    If you could however write all 1's to all the harddrives in the Southern Hemisphere it would likely cause a polarity shift on the earth however. (AF)
  • I think it's the hole in the "zero" that makes it weigh less than a "one." -wjc

  • For reminding me what day it is.
  • by Y Ddraig Goch ( 596795 ) on Tuesday April 01, 2003 @01:11PM (#5638555)
    to an online utility and an explanation of how it works Hard Drive Balancer [gbrockman.com]
    • Very funny link.

      I hope you enjoyed the grins, and if you were fooled into wanting to buy these, don't feel bad. You are in very good company . Save this site for the next April Fools day and see if you can find a victim or just a friend/co-worker with a good sense of humor.
  • The day that Slashdot becomes utterly fucking useless.

    Come on, guys, most of these "joke" stories aren't even mildly amusing or convincing. They're just plain dumb. Put a little effort into it, will ya?

  • 1's more then 0's? (Score:2, Insightful)

    by voot ( 609611 )
    If all the hardrives were ones it would have the north and south poles more defined and the hard drive would cave in to itself.
  • The real risk is that if you write all ones to the entire surface of the disk, you will effectively create a spinning monopole. This will cause your data to 'go back into time' and erase itself.
  • RLL-encoding makes long strings of "ones" and "zeros" unlikely.

    Back in the old FM days (where there was a direct relationship between flux and data) you might have had a problem such as you describe, but remember that the disks rotated at much lower RPMs then.

    These days you don't have anything to worry about.
  • by pmz ( 462998 )
    I'd be more worried about all those '1's standing up on that one half of the drive. This causes an aerodynamic drag imbalance that will induce harmonic vibration. Not only is the buzz annoying, I have heard from an aquaintance that the drive bearings will wear into an elliptical pattern causing all the cylinders on half the drive to be off-by-one. If the drive begins malfunctioning, tapping one side of the drive with a 20-oz. hammer will reset the bearing geometry.
  • You know, a 0 is bigger than a 1, so who's to say the 0s aren't heavier? Think of how many more atoms it must take to write that 0, not to mention the increased time...
  • As we all know, the more data on a drive, the heavier the unit. This was first demonstrated on the old "Maytag" style drives. The experiments demonstrated it wasn't so much the state of the bits, it was merely the data that increased the weight. Yes, a 1 bit can increase the weight of a single byte, but in the long run the effect is delta. You will *probably* have to worry when they start coming out with exobyte RAIDs, but who needs one of those?

    If you are really concerned, you will have to make sure

  • by MarkGriz ( 520778 ) on Tuesday April 01, 2003 @02:00PM (#5638944)
    PHB: This laptop computer weighs too much. Do we have anything lighter?
    Wally: Why don't you just delete files to lower the weight on that one?
    PHB (curiously studying laptop): That's a thought.
    Wally: Technically, I only asked why not.
  • by shoppa ( 464619 ) on Tuesday April 01, 2003 @02:06PM (#5638982)
    You are, of course, 100% right historically-speaking. Hard drives and tape drives used to use "NRZ" or "Non-Return-to-Zero" recording, where ones were recorded with a magnetic flux change and a zero without a flux change. The problem was actually much more severe with 7-track 556 BPI tape recordings, where the weight imbalance would cause the tape drive to actually jump up and down on the floor.

    Once a year, (traditionally, the first day in April) all disk and tape drives were rebalanced by redistibruting ones and zeroes. The "bit buckets" were also emptied on this hallowed day.

    This isn't a problem anymore because all modern recording media use "MFM", "RLL", or "GCR" encoding methods, where ones and zeroes are automatically balanced.

    One minor technical nit: "ones" actually weigh less than "zeroes". This led to the conclusion that the more data you put on your punched cards, the less they cost to mail :-)

  • On what file system your using. Modern file systems (like ReiserFS and ext3) take this into account and distribute the data accordingly using complex algorythms I don't pretend to understand. It's worth noting that fat16/32 don't have this problem either due to drive fragmentation. So if you use either of those two you're safe as long as you don't run a disk defrag tool.
  • Actually, because data is stored as 2's complement, the so-called "1" is actually a *negative* 1, causing it to actually levitate. This helps reduce friction on the disk, ultimately leading to fewer disk failures. Ahh, the wonders of modern science.
  • What an excellent question. Not many people are as perceptive as you. As a matter of fact it does make a difference, starting at the 80 gigabyte 7200 rpm range and it was part of the Microsoft suite of nanocode system support that had to be worked on starting with Windows 2000. It is an automatic system service to move data sectors with low zero counts automatically to maintain 'Odd-Set Symmetry' known as OSS in the Windows world. Macintosh has something similar, I believe it's known as 'Bipolar Symmetry' o
  • And the story is apocryphal.

    But it goes something like this:

    In an early NASA launch project the engineers needed to know the exact weight of every component in the craft. The programmers swore that the software didn't weigh anything. The engineer in question finally slams down a bunch of punch cards in frustration and demands to know how much the software weighs.

    The programmer replies, "We only use the holes."
  • There are more applications for harddrives. By letting the heads flow freely on the harddisk and look at the data it sees, you can measure your acceleration.
  • no -- the answer

"Never face facts; if you do, you'll never get up in the morning." -- Marlo Thomas