• Your Name (unregistered) in reply to marty
    marty:
    What makes me think it is made up is that pressure on the case of the server would not apply pressure to the heads onto the surface of the disk. Not quite the same thing as dust.
    TRWTF is that comment.
  • OldCoder (unregistered) in reply to OddOne
    OddOne:
    That's what thermite is for - very few people actually need secure erasure, and when secure erasure is a life-or-death matter the drive WILL be melted down to slag. Most high-security environments have thermite plates or grenades sitting atop equipment racks for precisely this purpose.
    Thermite? Why yes, I ordered it...

    What for? Why, uh, uuuh, for securing the data in my datacenter, yeah! That's right!

    Why did I need that much? Well, I wanted to make sure my data was destroyed completely, and I have a lot of data! Why do you ask?

    Terrorism? Really? I never thought of such a thing. Honest. You do believe me, don't you?

  • allun (unregistered) in reply to OldCoder
    OldCoder:
    OddOne:
    That's what thermite is for - very few people actually need secure erasure, and when secure erasure is a life-or-death matter the drive WILL be melted down to slag. Most high-security environments have thermite plates or grenades sitting atop equipment racks for precisely this purpose.
    Thermite? Why yes, I ordered it...
    Because rust and aluminium are so difficult to obtain separately?
  • Zapakh (unregistered) in reply to marty
    marty:
    Plus the overall made-upity nature of most wtf stories make them outlandish poorly imagined tall tales.
    ...Can I send you a résumé?
  • Tom (unregistered)

    I had a similar experience many years ago when I was a programmer in the Air Force. We had a IBM mainframe computer system with a very large magnetic tape library. The tapes were hung at multiple levels on numerous racks. One day we started experiencing corrupt tapes. The odd thing was it seemed to only be happening to tapes on the lowest rack levels, just off the floor. After several weeks of this happening, one of the admins was in the tape library one weekend and saw one of the cleaning crew come in the room with a large/industrial floor buffer (can we say electromagnet). He was effectively degaussing the tapes near the floor as he ran the buffer by them. Needless to say, that was the last time any floors in the data center were ever buffed.

  • NH (unregistered)

    That seems to be the time when you actually bill the construction company for all costs that did appear.

    Crashed disk, technician etc.

  • TC (unregistered)

    If it were me, I would blame the moron who decided to move LIVE servers into a room workman were still using. That's just plain stupid. The poor builder who decided to use the server as a saw-horse is just a consequence of a prior incorrect decision.

  • Dirk (unregistered) in reply to hikari
    hikari:
    No encryption is unbreakable, it's just a matter of the resources you have available; including time, money, tech, etc.
    A one time pad is unbreakable. It's been mathematically proven.
  • Santa Cklaus (unregistered)

    I should've saw that coming

  • Ian (unregistered) in reply to Mike

    Nah.. not on the first one.

    one disk failure is well within normal bounds, especially if the machines have been physically moved recently.

    when the second disk failed... yeah, you have a point.

  • Anonymous (unregistered)

    Yep, I remember this WTF. Not that I had to, since the punchline was given away in the title. Server failure during building works... "it doubles as a saw horse"... gee, I wonder what happened to the server?

  • Spudd86 (unregistered) in reply to Mike
    Mike:
    alegr:
    Mike:
    dkf:

    OTOH, I don't believe that any person with enough patience to do that exists anywhere in the world.

    The last company I worked for focused on exactly that issue. Many of our competitors simply drilled holes through a drive and its platters to "destroy" the data. Others would hit the drive with a hammer, cut it in half with a saw, or run it through a degaussing ring. Some just used DBAN or similar software. We went one step further with our 3-stage shredding system, which basically turned the drive into sand which was shipped to a smelter.

    With modern self-encrypting drives it's not necessary. You just rewrite the encryption key and all data becomes garbage.

    Until that encryption method is broken, possibly by a competitor with a great desire for your company's information, or a rogue government. Either of those entities may have billions of dollars to spend and plenty of spare time to burn on just that.

    Physical destruction > all other methods.

    We have RSA which we're pretty sure is secure against conventional computers (Factoring is (probably) not NP-Hard, but it does seem to hard enough, ie most people don't think it's in P), quantum computers can break it in polytime, but... nobody has them yet, and we don't know if they can be made remotely useful yet... (Also quantum computers cannot solve NP hard problems fast, unless P=NP, so if QP is quantum-polytime, we know that P is a subset of QP is a subset of NP, and P is probably not the same as QP)

    Then we have block ciphers... which as far as I know we don't have one with nice proof that it is not efficiently breakable unless P=NP or some problem everyone thinks is hard is actually easy or something like that...

    Basically if your data is encrypted with RSA and a reasonable key length (say 2048 or 4096 bits) you don't have much to worry about... AES on the other hand... may eventually be broken (but you'll hear about it, and probably get YEARS of advanced warning when someone finds a weakness... then everyone who isn't a cryptographer ignores it...)

  • Spudd86 (unregistered) in reply to Dirk
    Dirk:
    hikari:
    No encryption is unbreakable, it's just a matter of the resources you have available; including time, money, tech, etc.
    A one time pad is unbreakable. It's been mathematically proven.

    Yes... but it's also hard to use it right, and if you DON'T it's usually easy to break... also you need to securely store the pad... and if you've got an 80GB hd... the only place to do that is another hd... which kinda defeats (part of) the purpose of hardware encryption... (ie the key is in the hardware and there is no way to get it out ever)

  • Spudd86 (unregistered) in reply to hikari
    hikari:
    No encryption is unbreakable, it's just a matter of the resources you have available; including time, money, tech, etc.

    Of course, encryption isn't supposed to be unbreakable, it's just supposed to be unbreakable for long enough that the data it's protecting no-longer requires protecting.

    If you really want the data gone for good, physical destruction is the only way to go.

    Breaking RSA is intractable (ie run times that are longer than the lifetime of the sun) unless factoring numbers can be done MUCH more quickly than anyone thinks it can be done.

    Factoring is not known to be NP-Hard, but it is generally not thought to be in P either... so RSA is pretty safe until someone builds a practical quantum computer (which can solve the factoring problem in polytime...)

  • taltamir (unregistered)

    the real WTF is that an entire server rack (of the CORE server) contained only ONE drive.

Leave a comment on “Classic WTF: It Doubles as a Saw Horse”

Log In or post as a guest

Replying to comment #:

« Return to Article