Wiping a disk clean

Wiping a disk clean

[Larry Hazel]

If I create a binary file, use sequential_io to write records of all 
ones until Use_Error is raised.  Then, delete the file and repeat with 
all zeros.
Will the unused portion of the disk be wiped clean so that no one could 
recover what was written there before?

Re: Wiping a disk clean

[Jeffrey Carter]

Larry Hazel wrote:
> If I create a binary file, use sequential_io to write records of all 
> ones until Use_Error is raised.  Then, delete the file and repeat with 
> all zeros.
> Will the unused portion of the disk be wiped clean so that no one could 
> recover what was written there before?

Probably not. The US Govt requires 3 passes with specific patterns, and 
the most exhaustive study of the issue indicates that different 
approaches are required for different kinds of drives. For most modern 
disks, several passes with random data will suffice. "Several" is 
probably 5 or 6.

And that doesn't even get into the issues of whether Sequential_IO will 
write on all the unused portion of the disk, much less the unused parts 
of partially used sectors, and so on.

-- 
Jeff Carter
"You couldn't catch clap in a brothel, silly English K...niggets."
Monty Python & the Holy Grail
19

Re: Wiping a disk clean

[Larry Kilgallen]

In article <PRWHe.7695$0C.7190@newsread3.news.pas.earthlink.net>, Jeffrey Carter <spam@spam.com> writes:
> Larry Hazel wrote:
>> If I create a binary file, use sequential_io to write records of all 
>> ones until Use_Error is raised.  Then, delete the file and repeat with 
>> all zeros.
>> Will the unused portion of the disk be wiped clean so that no one could 
>> recover what was written there before?
> 
> Probably not. The US Govt requires 3 passes with specific patterns, and 
> the most exhaustive study of the issue indicates that different 
> approaches are required for different kinds of drives. For most modern 
> disks, several passes with random data will suffice. "Several" is 
> probably 5 or 6.
> 
> And that doesn't even get into the issues of whether Sequential_IO will 
> write on all the unused portion of the disk, much less the unused parts 
> of partially used sectors, and so on.

Or the formerly used parts of the disk that have been "revectored"
in response to hardware problems.

In summary, there is a large body of knowledge in this area that is
not at all specific to Ada.

Re: Wiping a disk clean

[Keith Thompson]

Dennis Lee Bieber <wlfraed@ix.netcom.com> writes:
[...]
> 	I spent 20 years on a program where the only allowed means of
> removing a disk drive was to run it into a heavy demagnetizer... And I
> mean HEAVY -- it not only wiped the platter of magnetic traces, but also
> wiped the motor magnets and the head positioning magnets. IOW: no more
> drive (well, the drive controller card might be salvageble).

I've heard stories of requirements to erase disk drives using
thermite.

-- 
Keith Thompson (The_Other_Keith) kst-u@mib.org  <http://www.ghoti.net/~kst>
San Diego Supercomputer Center             <*>  <http://users.sdsc.edu/~kst>
We must do something.  This is something.  Therefore, we must do this.

Re: Wiping a disk clean

[Marc A. Criley]

Keith Thompson wrote:
> I've heard stories of requirements to erase disk drives using
> thermite.

Metal grinder.  Nothing left but powder and shards.

Re: Wiping a disk clean

[Bobby D. Bryant]

On Wed, 03 Aug 2005, "Marc A. Criley" <mcNOSPAM@mckae.com> wrote:

> Keith Thompson wrote:
>> I've heard stories of requirements to erase disk drives using
>> thermite.
> 
> Metal grinder.  Nothing left but powder and shards.

Which must then be thrown into a black hole.

Though with Hawkins conceding his famous wager, even that might not
be enough...

-- 
Bobby Bryant
Austin, Texas

Re: Wiping a disk clean

[Keith Thompson]

Dennis Lee Bieber <wlfraed@ix.netcom.com> writes:
> On Wed, 03 Aug 2005 07:09:56 GMT, Keith Thompson <kst-u@mib.org>
> declaimed the following in comp.lang.ada:
>
>> 
>> I've heard stories of requirements to erase disk drives using
>> thermite.
>
> 	That would have set off the Halon fire suppression system <G>

I'm reasonably sure they took them outside first.

-- 
Keith Thompson (The_Other_Keith) kst-u@mib.org  <http://www.ghoti.net/~kst>
San Diego Supercomputer Center             <*>  <http://users.sdsc.edu/~kst>
We must do something.  This is something.  Therefore, we must do this.

Re: Wiping a disk clean

[Marc A. Criley]

Dennis Lee Bieber wrote:
> On Wed, 03 Aug 2005 22:45:03 GMT, Keith Thompson <kst-u@mib.org>
> declaimed the following in comp.lang.ada:
>>I'm reasonably sure they took them outside first.
> 
> 	Ah, but you see -- ours couldn't "be taken outside" (except in
> double locked security containers)

Right, then.  Vacate the building and call in a missile strike.  :-)
We'll get rid of those pesky bits!

Re: Wiping a disk clean

[Jean-Pierre Rosen]

Dennis Lee Bieber a �crit :

> 	I spent 20 years on a program where the only allowed means of
> removing a disk drive was to run it into a heavy demagnetizer... And I
> mean HEAVY -- it not only wiped the platter of magnetic traces, but also
> wiped the motor magnets and the head positioning magnets. IOW: no more
> drive (well, the drive controller card might be salvageble).
> 

If you need *that* level of security, I would say nothing beats putting 
the drive under a road roller...
-- 
---------------------------------------------------------
            J-P. Rosen (rosen@adalog.fr)
Visit Adalog's web site at http://www.adalog.fr

Re: Wiping a disk clean

[Larry Kilgallen]

[-- Warning: decoded text below may be mangled, UTF-8 assumed --]
[-- Attachment #1: Type: text/plain, Size: 696 bytes --]

In article <a9tpcd.pl.ln@hunter.axlog.fr>, Jean-Pierre Rosen <rosen@adalog.fr> writes:
> Dennis Lee Bieber a �crit :
> 
>> 	I spent 20 years on a program where the only allowed means of
>> removing a disk drive was to run it into a heavy demagnetizer... And I
>> mean HEAVY -- it not only wiped the platter of magnetic traces, but also
>> wiped the motor magnets and the head positioning magnets. IOW: no more
>> drive (well, the drive controller card might be salvageble).
>> 
> 
> If you need *that* level of security, I would say nothing beats putting 
> the drive under a road roller...

A participant in another newsgroup says her agency requires slagging.

Check with your accreditors.

Re: Wiping a disk clean

[Keith Thompson]

Jean-Pierre Rosen <rosen@adalog.fr> writes:
> Dennis Lee Bieber a �crit :
>
>> 	I spent 20 years on a program where the only allowed means of
>> removing a disk drive was to run it into a heavy demagnetizer... And I
>> mean HEAVY -- it not only wiped the platter of magnetic traces, but also
>> wiped the motor magnets and the head positioning magnets. IOW: no more
>> drive (well, the drive controller card might be salvageble).
>>
>
> If you need *that* level of security, I would say nothing beats
> putting the drive under a road roller...

Since disk platters are quite flat to begin with, I can imagine that a
lot of information could still be recovered (unless the road roller is
heavily magnetized, of course).

-- 
Keith Thompson (The_Other_Keith) kst-u@mib.org  <http://www.ghoti.net/~kst>
San Diego Supercomputer Center             <*>  <http://users.sdsc.edu/~kst>
We must do something.  This is something.  Therefore, we must do this.

Re: Wiping a disk clean

[Florian Weimer]

* Larry Hazel:

> Will the unused portion of the disk be wiped clean so that no one could 
> recover what was written there before?

No, some file system allocation units are probably only partly used
and its unused parts are not overwritten by this procedure.

Re: Wiping a disk clean

[Bobby D. Bryant]

On Tue, 02 Aug 2005, Larry Hazel <lhhazel@otelco.net> wrote:

> If I create a binary file, use sequential_io to write records of all 
> ones until Use_Error is raised.  Then, delete the file and repeat with 
> all zeros.
> Will the unused portion of the disk be wiped clean so that no one could 
> recover what was written there before?

Modulo the caveats a couple of others have already mentioned re
partially used allocation units, it would be easier and probably more
effective to repeat

	$ cat /dev/random > temp.dat ; rm temp.dat

a few dozen times.

As I understand it, the need for repetition comes from the fact that
a drive wobbles somewhat during use, so whenever you overwrite a
sector it may leave a "fringe" of the earlier data to one side or the
other of a track.

-- 
Bobby Bryant
Austin, Texas