[dm-crypt] Filling a disk with random data - question

Arno Wagner arno at wagner.name
Tue Mar 18 03:33:51 CET 2014

On Mon, Mar 17, 2014 at 19:55:05 CET, Cpp wrote:
> Hello,
> I noticed a lot of online articles recommend to overwrite your hard
> disk with random data before creating an encryption volume on it.
> Normally this is done by:
> # dd if=/dev/urandom of=/dev/sda bs=4096

This is slooooooooooow.... ;-)
> This will of course take a while, and there is also the badblocks
> alternative. However recently I've seen another approach that uses an
> encrypted (non-luks) container that is later filled with blank data.
> It's said that this approach is much faster than the urandom method
> above. 

I came up with that when I had to securely wipe 50 disks 
a few years ago. Other people may have had a similar idea.

> I haven't used this yet, so I hope I got the command line
> right:
> # cryptsetup -c aes-xts-plain64 -h sha512 -s 512 -d /dev/urandom open
> /dev/sda --type plain cryptroot

Make ist easier on you, the defaults are really quite enough: 

# cryptsetup create -d /dev/urandom /dev/sda cryptroot

> # dd if=/dev/zero of=/dev/mapper/cryptroot bs=4096

And you can get a progress meter like this:

# dd_rescue -w /dev/zero /dev/mapper/cryptroot

or like this

# cat /dev/zero | wcs > /dev/mapper/cryptroot

(uses my sream-meter "wcs" from:  http://www.tansi.org/tools/index.html)
> My question is are there any serious drawbacks of using this method in
> place of the urandom one?


Arno Wagner,     Dr. sc. techn., Dipl. Inform.,    Email: arno at wagner.name
GnuPG: ID: CB5D9718  FP: 12D6 C03B 1B30 33BB 13CF  B774 E35C 5FA1 CB5D 9718
A good decision is based on knowledge and not on numbers. -  Plato

More information about the dm-crypt mailing list