Paranoia and deletion: the wipe man page

http://www.boingboing.net/2011/01/14/paranoia-and-deletio.html

by julian37 • 16 years ago

93 70 16 years ago

Mithrandir• 16 years ago

The man page is correct about quite a few things, but it's also a bit sarcastic:

       The  best way to sanitize a storage medium is to subject it to tempera-
       tures exceeding 1500K.  As a cheap alternative, you might use  wipe  at
       your  own  risk.  Be  aware that it is very difficult to assess whether
       running wipe on a given file will actually wipe it -- it depends on  an
       awful  lot  of  factors,  such  as  :  the type of file system the file
       resides on (in particular, whether the file system is a journaling  one
       or not), the type of storage medium used, and the least significant bit
       of the phase of the moon.

But no matter what, wipe is a really great program.

Jach• 16 years ago

Funny, I was just having a conversation today with a friend who works for the company selling http://www.whitecanyon.com/wipedrive-erase-hard-drive.php and remarking that it may keep your data safe from identity thieves, it wouldn't protect you from the FBI.

MikeCapone2• 16 years ago

Did he explain more specifically why he thinks it would "protect you from the FBI"?

DanI-S• 16 years ago

The idea that they need actual evidence to lock you up is so 20th century...

MikeCapone2• 16 years ago

*I can't edit my previous message anymore, but the "would" should read as "wouldn't".

srean2• 16 years ago

An alternative could be to overwrite the disk with null bytes.

  dd if=/dev/zero of=/dev/hda bs=1M

Wonder if it is going to be slower than wipe.

ComputerGuru2• 16 years ago

....multiple times.

Overwriting it once prevents software reconstruction of the data, but magnetic analysis of the underlying disk itself can reveal (depending on the voltage returned by the resulting 0 or 1) whether the previous value was (within a degree of certainty) a 0 or 1.

srean2• 16 years ago

Ah good point +1. Did not think about that. Maybe a single pass from

  /dev/urandom

will suffice. But its going to be slow.

nodata• 16 years ago

Wait - how automated is magnetic analysis? How much would it cost to recover a gigabyte disk, for example? What about non-spinning disks, are they cheaper or more expensive?

amock2• 16 years ago

Do you have any sources for this? I've heard about this before but I've never seen any details about how it works.

__david__• 16 years ago

You may find this interesting: http://www.nber.org/sys-admin/overwritten-data-guttman.html

It's a counter argument to the urban legend that says data should be wiped multiple times to be truly deleted.

It's worth noting that with today's disks' PRML channels, the signal is barely there already. It's merely "guessed" at (ML in PRML stands for maximum likelyhood). It's seems crazy that anyone could recover something after it's been overwritten. Maybe in the past, but not any more.

JoshCole• 16 years ago

What happens if you only overwrite a bit if the bit that was there before writing is different? This way your not overwriting every bit, so guessing at the previous bit can lead you astray. Or is that impossible?

getsat3• 16 years ago

I have a script that mounts a tiny ramdisk, fetches a statically compiled shred binary I've compiled, and proceeds to shred each disk attached to the system up to 25 times. I generally use this when decommissioning servers not under my direct control.

algorias• 16 years ago

What about overwriting it with random data?

Someone• 16 years ago

That is nsufficient to be 100% sure that all our data gets overwritten. The disk's firmware could discover that a block is unreliable and never write to it again. Such a block could contain recoverable data.

Also, I am not sure that this is guaranteed to overwrite the last part of a disk whose size is not a multiple of 1M. I guess that will depend on how eagerly the device detects ENOSPC conditions.

ck22• 16 years ago

I always dissemble my old hard drives, it's therapeutic somehow.

I am not sure how they broke that disc into so many pieces though, in my experience they are incredibly strong and rigid.

The discs make great wind chimes too.

Harbor Freight has a $3 security bit set if you have drives with special screws.

InclinedPlane2• 16 years ago

Liquid nitrogen?

ck22• 16 years ago

I looked it up on flickr http://www.flickr.com/photos/jon_a_ross/1482849745/

and from the caption it sounds like it actually did that to itself which seems impossible unless maybe the platter was made defectively in the first place ... ? (note it was an "old" 40gb drive - also looks like 2.5 inch)

wazoox2• 16 years ago

Some drives have glass platters AFAIK. This would explain how this one was broken.

hnhg2• 16 years ago

Sorry to be a pedant, but I think you mean "disassemble" - I always get the two mixed up too.

ronnoch• 16 years ago

Currently your comment, which adds nothing to the discussion except a simple spelling correction, has more upvotes than the parent. Sign of the times for HN?

Not to say spelling corrections are never useful, I think they often are, but I don't understand the mindset that upvotes them over comments that actually participate in the discussion.

hexley2• 16 years ago

Glass platters?

rasur• 16 years ago

My method for dealing with old drives:

- remove top plate of drive - a number of heavy blows with a Sledgehammer. - incineration.

This usually does the trick.

mmaunder• 16 years ago

I sleep at night knowing that the teams of highly effective government data recovery personnel that Hollywood portrays will never exist.

chanux• 16 years ago

Any idea what's the difference between wipe and shred?

naner• 16 years ago

The shred man page warns you that the tool is not fully effective on journaled filesystems. I doubt wipe is, either.

Entlin• 16 years ago

To everybody spending thought on this: Relax, nobody is interested in your porn collection.

haribilalic• 16 years ago

I've used such tools to wipe my hard drives before selling or donating used computers (although I would use them before disposing of a hard drive in any way, really). I have heard of people buying up used computers and hard drives and extracting saved passwords from them.

A lot of people are smart enough to format it quickly, but not everyone will let (or know to) a computer sit for dozens of hours so that things might be wiped more securely.

halfasleep• 16 years ago

I generally recommend DBAN[1] for people who are disposing of old boxes.

[1]http://www.dban.org/

jws• 16 years ago

I've had the experience where people agree to buy my broken computers for parts and the deal is done, until I mention that the "drives are wiped and ready to go". Then I never hear from them again.

Granted, somehow visiting the craigslist site turns ordinary people into flakes, but the "wiped" statement correlates a bit too well with the vanishing interest.

troels• 16 years ago

    Of course this shifts the trust to the computing system,
    the CPU, and so on. I guess there are also "traps" in the
    CPU and, in fact, in every sufficiently advanced mass-
    marketed chip. Wealthy nations can find those. Therefore
    these are mainly used for criminal investigation and
    "control of public dissent".

I'm unsure in which way a government agency could benefit from having backdoors in the CPU? Even if they did, and even if it could detect that some sort of encryption was going on, where would it store the interesting data?

sesqu• 16 years ago

Well, there was/is that trusted computing thing, i.e. putting DRM on the PC in hardware. The benefit would be denying encryption from terrorists/pirates/dissenters, like he just said. Interesting data can be stored just about anywhere, from phoning home to hidden partitions to having a small flash on the BIOS, depending on what's deemed interesting (logs, sector addresses, cryptographic keys, painted files).

hexley2• 16 years ago

For those on Mac OS X (and perhaps *BSD?), we have srm.

The Secure Empty Trash function is a frontend for this IIRC.

stcredzero• 16 years ago

I wish there was a version that only overwrite once with 1's. That would be best for my SSD. (Which has wear leveling and is formatted with a journalling filesystem anyhow, so perhaps it's moot.)

suraj• 16 years ago

How would a hard disk controller detect encrypted data? It is essentially a (psudo-) random stream of bytes. Even if the controller is programmed to recognize such streams, it would be easy to first wipe disk by writing small random files all over and then wiping with 0/1 pattern. So any cached data is essentially worthless.

endian• 16 years ago

Does anyone know if any such FOSS works with SSDs and their wear-levelling of their writes?

tedunangst• 16 years ago

wipe is rather unlikely to overwrite the bits on an SSD. Even with TRIM (which wipe won't use, btw), the disk will prefer using a blank sector to one it has to erase. Reading the currently unmapped spare sectors is probably something anyone with a solder gun can do.

Confusion• 16 years ago

Silly conspiracy theory.

1) The scope is too large. Too many engineers at hard drive manufacturing companies would have to know about it.

2) Too America-centric. Engineers in foreign countries, that actually build most of the stuff, would know about it. It would be an enormous security gap that they could use as well.

The sad thing is that such paranoid fantasies, that are easy to debunk, blind you to the actual conspiracies, that are less comprehensive, more subtle and therefore much more threatening. I bet the TLA government agencies love these stories.

rapind• 16 years ago

Maybe it's the gov who posits these theories?... nah.

SoftwareMaven2• 16 years ago

The invisible yellow dots identifying printers went a long time before becoming common knowledge. I don't believe TFA for an instance, but it doesn't seem an impossibility.

SoftwareMaven2• 16 years ago

Ugh, sometimes I wish I could edit later. Make that "I don't believe for an instant". I hate responding on my phone.

coffeeaddicted• 16 years ago

But what to do against those brain-wave readers? I better start thinking encrypted!

sp332• 16 years ago

I'm sorry, but this info is nearly ten years out-of-date. First, the Gutmann method was designed for encodings back in the days of RLL/MFM encoded drives. For newer drives which really push the physics quite a lot further, two passes of random data are enough to throw the magnetic domains into a statistical dead heat. There just isn't any physical room on the drive to hold old data.

Secondly, new drives reserve a percentage of the room (invisible to the user), in case some of the sectors go bad the controller will re-map them transparently to new sectors. This might leave old data in the old sectors, where you can't normally see it but an investigator armed with the proper ATA commands can. (This isn't a conspiracy of the government and drive manufacturers, it's all there in the ATA spec.) The correct way to securely erase a drive is to send the drive the SECURITY-ERASE command. The drive controller will securely erase every part of the drive. https://ata.wiki.kernel.org/index.php/ATA_Secure_Erase The NSA actually recommends this to other government agencies, so it's probably OK.

jwr2• 16 years ago

But how do you know if the data actually gets erased by the drive?

mseebach• 16 years ago

You don't know, you hardly ever know anything. You evaluate the evidence (as laid out by the GP) and conclude that it's most likely the case.

This is really the approach that most conspiracy theories take. The fact that we don't know something to be true (but it's very likely to be) is used as an implication that it's probably false.

jerf• 16 years ago

If that possibility seriously concerns you, follow it up by physical destruction of the drive. Magnets strong enough to actually damage the data aren't a good cost/benefit from what I gather (too strong, too dangerous, too much hassle), so either shred or melt the drive platters.

I'm generally satisfied with df if=/dev/zero of=/dev/sda because I'm not actually dealing with anything that sensitive, but if you are, just take appropriate steps. Defense in depth.

icarus_drowning• 16 years ago

If you're that worried about it, there's always thermite.

TheAmazingIdiot• 16 years ago

Well, you don't know. It all depends on what kind of data was on the target drive.

If it's schoolwork or unimportant business data, a format is good enough. If we're talking CC and SSN numbers, the ATA command with writing 0's is good. If it's sensitive, a sledgehammer and an anvil is suitable.

I inherited a junk 1G hard drive, and just 'filed it away' in my junk drawer. After years, I actually checked the contents. Lo and behold: a flat database full of medical data, including SSN's and other revealing data. It was only 1G so it got the sledge (I use a small section of railroad tie instead of a blacksmiths anvil).

gus_massa• 16 years ago

I think that incinerating the disk should erase the data. Magnetic material become normal materials after some temperature http://en.wikipedia.org/wiki/Curie_temperature . Usually it is no more than 1000 C/ 1800 F, so a good fire should be enough.

Warning: I don't know if the fumes are toxic.

deutronium• 16 years ago

I love this section:

"I strongly recommend to call wipe directly on the corresponding block device with the appropriate options. However THIS IS AN EXTREMELY DANGEROUS THING TO DO. Be sure to be sober."

ams6110• 16 years ago

I use DBAN to wipe hard drives before disposal. If you are replacing a failed drive, the only reasonable thing you can do is physically destroy the platters.

cloudwalking• 16 years ago

When I was younger, I erased a harddrive by covering it with burning thermite. It was pretty impressive, melting the drive, turning the sand underneath to glass, and burning through the asbestos pad into the cement.

Paranoia and deletion: the wipe man page

Related Stories