• Fluke@feddit.uk
    link
    fedilink
    arrow-up
    11
    ·
    edit-2
    6 hours ago

    Then 1s, then a pattern of 1s and 0s, then the inverse of that pattern, then another pattern, for a number of cycles.

    Data can actually be recovered beyond multiple overwrites, if enough time and money is thrown at it.

    • JackbyDev@programming.dev
      link
      fedilink
      English
      arrow-up
      2
      ·
      2 hours ago

      They keep saying that but those Bitcoins are still in the dump. (I’m aware it’s not comparable since having the drive in hand versus missing is a huge difference. Just a little joke.)

    • wabasso@lemmy.ca
      link
      fedilink
      English
      arrow-up
      16
      ·
      6 hours ago

      If there is something on your disk that a state actor is going to use magnetic microscopy to try to recover, it seems absurd to worry about still being able to use that hard drive and not just crush/melt it to be sure.

    • Buddahriffic@lemmy.world
      link
      fedilink
      arrow-up
      8
      ·
      6 hours ago

      Is that still the case with SSDs? I understood it to be a property of magnetic disks, and only possible because the drives can be disassembled and then read with a more sensitive reading head. I can’t think of a way to do that with flash circuitry unless it’s already designed to do that.

        • Buddahriffic@lemmy.world
          link
          fedilink
          arrow-up
          5
          ·
          4 hours ago

          Oh yeah, the bit where SSDs move sectors around for wear evening is important. Because of that, it’s possible to completely fill up an SSD after deleting files and still have those files recoverable from the flash chips themselves. Without that secure erase, as I understand it, if a sector gets marked “bad”, whatever data is there might stay there forever (or at least as long as the cells hold a charge).

          So there’s no benefit to writing multiple passes over deleted data on SSDs as far as the flash is concerned, but multiple passes might make it more likely for the controler to actually direct those extra writes to a sector actually storing the data (though the odds might be low unless you’re overwriting all free space, though even that depends on how much space is free vs how many “spare” sectors there are, and even then it might be impossible to get it to write to a sector marked “bad”).