Is it absolutely necessary to smash the mass storage device's hardware to pieces to make the data on it unrecoverable? I knew about /dev/zero and /dev/urandom, which, from what I understand, would be the equivalent of a C program that, via syscall write, would write a binary file, either all 0s or random, that would occupy the "free" space (i.e., the space not marked in the index table) until the disk is full. Once it's full, the file is deleted from the index table. In theory, it should work like this, at least for mechanical mass storage devices... For NVMe and SSDs, I know they talk about trim, but I don't know if it's actually a real thing... In any case, why are both methods fallacious in preventing data recovery? On paper, it seems to make sense.
>>107769094Isn't it enough to just disk zeroing?
>>107769094https://www.kingston.com/en/blog/personal-storage/how-to-securely-erase-ssd
>>107769113Memes aside, I knew the NSA and Mossad (I'm not actually meming) were capable of restoring even zeroed or randomized disks to previous states. A guy at university explained to me years ago that they do this through small caches residing on the disk's hardware. On the hardware side, I think it's possible to maintain a history of some indexes, but I don't understand how to trace them back to the original block, given that, in theory, the index table contains indexes that point to a section of memory whose contents shouldn't actually be the same as the one it pointed to previously... So I don't know, but the idea still fascinates me, at least to understand how these devices actually work.>Isn't it enough to just disk zeroing?Do it on an SSD :)
>>107769094you could never know if the data is truly erased. use FDE and when you're done with it you won't have a problem of erasing data
>>107769162I must also include this doesn't defat information leaks from side channels which include write/read patterns. SSD sectors wear after each amount of write.
>>107769094>occupy the "free" spaceSmall files aren't written to disk as regular files, they're written to the MFT as resident files. Once the MFT grows it never shrinks again, and free entries are reused at random. So, merely filling the free space with zeroes or whatever will not securely remove your deleted data, not even from an old HDD.
>>107769220>>107769162So, what's the safest option for SSDs and HDDs? And what's the best way to do this with disks that are still in use?
>>107769094All modern storage devices are independent computers running proprietary firmware. You have no clue what they're really doing and they're free to ignore any drive erasing commands or make additional copies of data in the spare disc space. Even full disc encryption isn't a perfect defense because the drive can log access patterns that could leak information. Mechanical destruction is simple and reliable.
>>107769413and FUN!
>>107769148zeroed, sure. randomized highly unlikely.the best they could get from that cache file is a filetree of names, not the data within it.unless hdd's secretly all have futuristic invisible chips in it which would be manufactured by china.
>>107769094Lol> writing the drive with random ones and zeroes 10x never truly erases the drive! Mossad or the Chinese could recover it? More like> CAN ANYONE RECOVER THIS FAILED DRIVE! I'LL PAY ANYTHING > sorry mate that's impossible
>>107769220>small filesYou mean two sectors, a.k.a. 1024 bytes.
>>107770008>1024 bytesu mean 1kb?
>>107770405That would be 1 KiB.
>>107769413>All modern storage devices are independent computers running proprietary firmware. You have no clue what they're really doing and they're free to ignore any drive erasing commands or make additional copies of data in the spare disc space. Even full disc encryption isn't a perfect defense because the drive can log access patterns that could leak information.What if destruction isn't an option? Damn, with an argument that's valid to a third party, you force me to think that the hammer on mass storage is a mega meme.
>>107770846Smashing device with hammer always works. What are you on about. Destruction is always an option unless you don't own it. I personally think it's excessive, but I wouldn't set an HD/SSD in trash without physically trashing it first.
>>107770875It was just right to better understand how these devices work and whether disk encryption (mechanical or not) is actually worthwhile.
>>107771569Encryption does work; the issue with it is it can work too well... like losing the code to the encryption and thus access to the contents of the drive. t. just formatted a drive with that issue on Win10 machine after OS reinstall. Fortunately was already backed up.
