Hacker News Reader

Comment by sillysaurus2 | Hacker News Reader

dispose of storage devices while being certain the data is unrecoverable

I use Derek's Boot and Nuke bootdisc for this purpose. http://www.dban.org/

It's straightforward to use, but it's also configurable if you want to be extra certain the data is gone. I think the default is 3 passes of filling the harddrive with random data generated via Mersenne Twister.

View on HN

Unless you're very paranoid then a single pass of zeroes should be sufficient.

The bigger problem is what to to with a dead hard disk. It's usually easier to buy a new one rather than replace it, but an attacker could perhaps repair the disk to steal the data. At work we send any non-wipable disk to be physically shredded. At home I think I'd just hold on to all disks indefinitely.

There's lots of ways to damage the platters, at which point the data probably won't justify exotic recovery methods.

For a drive with bitcoin on it, just move the coins to new addresses before disposing of it.

Open it, take the disc out and microwave it.

Dangerous?

"Is It A Good Idea To Microwave An XBOX 360?"

http://www.youtube.com/watch?v=vzodemYzswQ

"Is It A Good Idea To Microwave A PlayStation 3?"

http://www.youtube.com/watch?v=4rWyJXpezPs

"Is It A Good Idea To Microwave A Nintendo Wii?"

http://www.youtube.com/watch?v=OydTZpbp0EE

Drive a few nails through the case, it's fun and destructive! When drives were much larger, we used to take them apart, use the platter for mirrors/silly decorations and the magnets to magnetize tools around the office.

>> Unless you're very paranoid then a single pass of zeroes should be sufficient.

If you're paranoid it won't matter how many passes of zeroes you do. After 1 it's done.

If you're still paranoid, hard drives make great rifle targets.

These days there is a proper secure wipe built into most drives themselves. Its part of the ATA spec. https://ata.wiki.kernel.org/index.php/ATA_Secure_Erase

Yet if you ask how to wipe a disk on a forum, you'll get user-space ideas like `dd` and `shred` :(

A lot of people want a more physically certain level of data destruction than such a command provides. If I saw that a drive wrote 3-5 passed of junk data before zeroing the disk then I would trust it but such firmware level info is rare to impossible up find.

Many SSDs do on the fly AES encryption. Secure wipe is then just a destruction of the private key. All the encrypted data stays put.

Since when there is private key in AES? And that requires a paranoid person to trust that a company does not holds the keys in escrows somewhere or are by any way retrievable.

DBAN's good, there are a couple of other ways I've used including dd.

Here's a blog post about it http://alicious.com/secure-drive-data-wiping/ including references to some papers where they looked at the probability/possibility of recovery.

To clarify - the hard part is not the destruction of data you know about, it's the identification of data on devices you don't know about, or forget.

Computers with hard disks are obvious, but also consider mobile phones, tablets, games devices (that increasingly ask for sensitive information), PVRs, and so on.

There may be a point, if we haven't already reached it, where I can't dispose of a toaster without worrying about how to wipe its memory.

Don't forget office printers! Some have hard drives or other nonvolatile storage. You know, in case you want to run off a second copy of that fax from 2009.

How about cat /dev/random > file ?

Doesn't that effectively do:

    unlink(filename)
    create(filename)
    write(filename, random)

? Though something that opens the file in append mode then seeks to 0 would probably overwrite the same HDD locations.

Yes. It will truncate the file first, then allocate new blocks for the new random data, likely leaving the old blocks lying around.

You can avoid this with the 'conv=notrunc' option of 'dd'[1]. It will overwrite existing blocks instead of truncating (and possibly reallocating):

    notrunc  Do not truncate the output file.  This will preserve
             any blocks in the output file not explicitly written by
             dd.  The notrunc value is not supported for tapes.

[1]:https://developer.apple.com/library/mac/documentation/Darwin...

Do you mean that the system will try to compress whatever gigabytes of random data?

No, I mean it will store the new random data in different blocks, thus not overwriting the old blocks.

Even if the whole drive is 'empty' ?

I don't know whether a nearly-empty drive is better or worse than average. But the point is, you had a 1000-block file full of secret data and when you do

    > secrets.txt

the file is truncated, freeing blocks 1-999 (usually block 0 is zero-filled.) If you proceed to write random data it will go to newly-allocated blocks. Then a raw read of the original blocks will expose your secret data.

With dd and notrunc, the random data goes to the original blocks, overwriting your secrets.

Can you explain a bit more?

Are you saying this will not overwrite some of the blocks?

Reply on news.ycombinator.com