In all seriousness, I think the best way to quickly wipe a drive is to get the drive firmware to do it.

This has a few advantages..

1. It's really fast, as its close to the metal.
2. If the wiping stops half way, due to power failure, when the drive next starts the wiping will continue, also if the drive has physical faults, the firmware will try to prevent read of the data even if it can't successfully wipe. 
3. Presumably reallocated bad sectors are also wiped (not true with dd).

Part of the ATA spec says that you can set a drive password. Another part says if you lose the password you can recover the drive, but only by wiping the data first. 

So how would you do this?

1. Use hdparm to set a password
2. Use hdparm to reset the password
(Wiping continues in the background).

Reference:
https://ata.wiki.kernel.org/index.php/ATA_Secure_Erase

-Noah

On Tuesday, September 3, 2013, Christopher M. Bailey wrote:
Thermite, yep that'll do it :-P



-------- Original message --------
From: Daniel Cross <daniel@ritualmedia.co.nz>
Date:
To: David Zuccaro <david.zuccaro@optusnet.com.au>
Cc: luv-main@luv.asn.au
Subject: Re: Disk Washing


Well, in that case probably Dban will be sufficient for your needs.  If still going with dd I agree with the bs=1M switch.

Incidentally, a discussion with a friend this evening lead us to the decision that a good way (although involved) to erase a drive would be:

1) Open disk enclusure.
2) Use bench grinder to create aluminium powder from the top of the disk enclosure
3) find something rusty, (or another source of iron oxide) and create powder.
4) mix 50:50 by volume.
5) if available add some magnesium (left over guy fawks sparklers?)
6) pour powder over HDD spindles.
7) Light and step well back while the platters melt.  

I might try this in coming months and report my results ;)

Daniel


On Mon, Sep 2, 2013 at 6:30 PM, David Zuccaro <david.zuccaro@optusnet.com.au> wrote:
On Mon, 2013-09-02 at 17:45 +1000, Daniel Cross wrote:

> I guess it depends on your threat model, so rewinding a bit - who are
> you wanting to prevent accessing your data? This greatly informs the
> conversation.  The methods offered above, and in many of the replies
> thus far, assume a very sophisticated adversary.
>
> Daniel

Let us assume that the adversary is not a
three-letter-government-agency; not that I consent to
three-letter-government-agencies accessing any of my information.