Subject: Re: rm -P and strange file systems
To: Geert Hendrickx <ghen@NetBSD.org>
From: Daniel Carosone <dan@geek.com.au>
List: tech-userlevel
Date: 08/25/2006 07:51:07
--ev7mvGV+3JQuI2Eo
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
Content-Transfer-Encoding: quoted-printable

On Thu, Aug 24, 2006 at 09:16:21PM +0200, Geert Hendrickx wrote:
> On Thu, Aug 24, 2006 at 08:52:58PM +0200, Ignatios Souvatzis wrote:
> > I wonder whether rm should be able to find out about this. Consider the
> > case where a user is operating on, say, NFS with the remote server
> > exporting LFS transformed with maybe an intermediate layer... how would
> > the user know?=20
>=20
> The user can and should not make any assumptions about the storage of his
> files when they are on NFS.  If the data is sensitive, he should use the
> cgd-on-vnd-on-NFS trick as described in
> http://www.netbsd.org/guide/en/chap-cgd.html#chap-cgd-suggestions

Nor even on local disks, even with block-overwriting filesystems like
ffs or msdos.  As noted elsewhere, the underlying disk could have
remapped marginal blocks (and may even be triggered to do so by the -P
overwrite).

Or, heaven forbid, someone might actually have taken *backups*!

None of these things makes rm -P useless, but they are important
qualifications and caveats on its usage - and which the manpage
describes, iirc.  If you need more than that, you need more than that,
and cgd might well help.

--
Dan.
--ev7mvGV+3JQuI2Eo
Content-Type: application/pgp-signature
Content-Disposition: inline

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.4 (NetBSD)

iD8DBQFE7h9KEAVxvV4N66cRAijhAJ47rFm7tTZ1ZQzWY3sYcpbHJNUrSACgtJkh
h+Kre5fYIIWddw1276Rn6zQ=
=Gl8x
-----END PGP SIGNATURE-----

--ev7mvGV+3JQuI2Eo--