Re: Wiping a disk partition

On Thu, Jul 3, 2025 at 4:09 AM David Christensen <dpchrist@holgerdanske.com>
wrote:

> On 7/2/25 01:18, Odhiambo Washington wrote:
> > My question was "What is the fastest way to wipe all data on
> /dev/ada1p2?"
> > It looks like delete, wipe, and erase all have different meanings here.
> > I should have explained that "rm -rf" would take a long time, so I
> needed a
> > faster command.
> > I didn't even want a secure erase, just the deletion.
> > I concluded that there is no straight answer to my question.
>
>
> I tend to think of "remove" in terms of file system metadata only --
> e.g. rm(1).
>
>
> I tend to think of "erase" in terms of file system data and metadata --
> e.g. shred(1).  But, journals, copy-on-write, snapshots, wear leveling,
> etc., can defeat true destruction of data if the software tool and/or
> storage API/DDI stack are not specified designed for such.  Research and
> verification would be required.
>
>
> I tend to use "wipe" for entire drives -- security erase if available,
> shred(1) or dd(1) /dev/zero if not.
>
>
> If rm(1) is too slow because you have a huge number of small files and
> if you do not care about leftover data on disk, the only tool I can
> think of that might be faster would be newfs(8).
>
>
> > What I ended up doing was
> > dd if=/dev/ada1 of=/dev/ada2  bs=1g status=progress
> > The 2nd disk is a replica of the 1st disk, and I use it only for
> creating a
> > backup of the first disk.
>
>
> I assume you mean (?):
>
> dd if=/dev/ada0 of=/dev/ada1 bs=1g status=progress
>
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
Yes!


That should work if you boot from ada0p2 into single-user mode, ada0p2
> is mounted read-only, and ada1p2 not mounted.  But, if both drives are
> installed (such as to recover files), you might have problems if
> fstab(5), etc., uses UUID's.
>

What I did was:

umount /disk2
dd if=/dev/ada0 of=/dev/ada1 bs=1g status=progress

-- 
Best regards,
Odhiambo WASHINGTON,
Nairobi,KE
+254 7 3200 0004/+254 7 2274 3223
 In an Internet failure case, the #1 suspect is a constant: DNS.
"Oh, the cruft.", egrep -v '^$|^.*#' ¯\_(ツ)_/¯ :-)
[How to ask smart questions:
http://www.catb.org/~esr/faqs/smart-questions.html]