DOS ain't dead

Hi,
I just wanted to know if there is a working wipeout tool for FAT32 that wipes the whole free parts of a FAT32 partition (2 TB).
As an alternative: Is there a working wipeout tool that wipes a WHOLE 2 TB HD including the MBR?

The reason:
I ran some tests (for time reasons only on a 30 GB virtual HD), filled it up with trash files and then deleted them so that the HD was "empty" again. So I could see if the programs really wiped with a hex editor.
Then I ran 3 different tools:
Wipefree from Uwe Sieber (https://www.uwe-sieber.de/util.html), it says that it should do 2 TB in about 55 hours, but the result was that I got less than 2GB. Additionally it only shows the amount of wiped size when you exit the program. It ran and ran over hours but only less than 2 GB were done when I aborted. Source Code is included but I saw nothing about the license.

Zerofill from Nikkho (https://sourceforge.net/projects/nikkhokkho/files/ZEROFILL/) 1.09 only wipes 2 GB (4x500 MB), but has a UI. It is the wiper delivered with FD.

The third tool that I tested was not free, created in about 1040 (!!) 100 MB files , but then it crashed as it had a unlucky filename tempxxxx.0 till tempxxxx.999 - so it fell out of 8.3 an ended in reality at .999 and in about 100 GB.

So all programs are far away from wiping a 2 TB HD.

> As an alternative: Is there a working wipeout tool that wipes a WHOLE 2 TB
> HD including the MBR?

WDe may do that . However, it's no cmdline tool, but interactive:

- launch WDe, using the disk# as argument, i.e. "WDe 0"
- press F6 ( functions )
- press F2 ( fill )
- enter the pattern you want to fill the HD with

After you entered the no of sectors to be written, WDe starts, without further confirmation. The write process tells its progress and can be aborted at any time. Good luck .

---
MS-DOS forever!

Hi! If you want to wipe the entire drive, it does not matter whether you have FAT16, FAT32 or other contents on it.

If you only want to wipe free space (areas not currently used by files or directories) then you do of course need something which is FAT32 aware.

What you can do as a partial solution: Fill the disk with uninteresting files (in Linux you could for example dd if=/dev/zero of=onehugefile) until there are no unused clusters left. Then at least nothing can be undeleted. This will not wipe things stored in unused areas of half-used clusters, of course.

You could also use FORMAT x: /u to wipe all files, directories and anything which could be undeleted otherwise, but it will take ages on large FAT32 drives.

Easiest would be something similar to what the Linux command dd if=/dev/zero of=/dev/yourwholeharddisk would do. Sometimes things can be quite simple.

Alternatively, you could invoke the ATA security erase command to tell the disk to completely erase itself, which might be faster: https://wiki.osdev.org/ATA_Command_Matrix

Not sure which DOS tools can trigger that on IDE, ATA, SATA or other drives, though? In Linux, you can for example use hdparm to trigger a security erase.

---
FreeDOS / DOSEMU2 / ...

You want to delete the entire physical drive or only one partition?
Imagine that you have disk divided to partitions C: and D: and one non-DOS partition.
You want to delete the C: only, or completely everything? (even including MBR)?

---
DOS-u-akbar!

> The third tool that I tested was not free, created in about 1040 (!!) 100
> MB files , but then it crashed as it had a unlucky filename tempxxxx.0 till
> tempxxxx.999 - so it fell out of 8.3 an ended in reality at .999 and in
> about 100 GB.

There's also "my" Wipe tool, but it currently has the same problem - creates multiple files (each of 65535 * cluster size), but doesn't allow that it will need more than 1000. If you find it suitable otherwise, I could fix it (although you'd probably be better off with something else).

> Wipefree from Uwe Sieber (https://www.uwe-sieber.de/util.html), it says
> that it should do 2 TB in about 55 hours, but the result was that I got
> less than 2GB.

It only shows the result of the last file it writes, not the sum of all files (it too creates multiple files, a byte at a time).

If you're not given any other suggestions I'll see about updating Wipe.

> > Wipefree from Uwe Sieber (https://www.uwe-sieber.de/util.html)...
> ... creates multiple files, a byte at a time

For speed reasons, I suggest a cluster at a time or 32 or 64 kB at a time, or possibly a strategy which shrinks blocks only at the end when the disk is almost full

---
FreeDOS / DOSEMU2 / ...

> > The third tool that I tested was not free, created in about 1040 (!!)
> 100
> > MB files , but then it crashed as it had a unlucky filename tempxxxx.0
> till
> > tempxxxx.999 - so it fell out of 8.3 an ended in reality at .999 and in
> > about 100 GB.
>
> There's also "my" Wipe
> tool, but it currently has the same problem - creates multiple files (each
> of 65535 * cluster size), but doesn't allow that it will need more than
> 1000. If you find it suitable otherwise, I could fix it (although you'd
> probably be better off with something else).

I will give wipe a try, the description sounds good. Will last a while as I am out of home for a few days. But it would be great if it could really support 2 TB.

From my former job I know how important it can be to delete files secure.
I found a lot of important deleted files in the last 20 years


For this reason I asked for a partition wiper and a whole HD wiper.

Hi, I just started wipe. Would it be possible to add the source code and a license when you fix the 2 TB problem? Thx.

> > As an alternative: Is there a working wipeout tool that wipes a WHOLE 2
> TB
> > HD including the MBR?
>
> WDe may do that . However, it's no cmdline tool, but interactive:
>
> - launch WDe, using the disk# as argument, i.e. "WDe 0"
> - press F6 ( functions )
> - press F2 ( fill )
> - enter the pattern you want to fill the HD with
>
> After you entered the no of sectors to be written, WDe starts, without
> further confirmation. The write process tells its progress and can be
> aborted at any time. Good luck .

I ran a first quick test over in about 5 mio sectors with random values on a 300 GB virtual IDE HD in ca. 8 minutes. Works great too, but for a regular user this will be too complex and dangerous to handle.
Nevertheless: Thanks!

> Would it be possible to add the source code

Source is shared by all three programs, so I would rather leave it separate.

> and a license

I've contacted the original author (if he's still monitoring that address; it didn't bounce, so that's something).

Edit: he got back to me, we'll go with zlib.