On Wed, Oct 8, 2014 at 10:35 AM, Carlos E. R. carlos.e.r@opensuse.org wrote:
What about shred /dev/sdx. Does it terminate on the bad sector?
etc, etc, etc.
I see. Yes, that's easier.
Question: would not the write attempt to a failed sector automatically remap it, to one of those sectors reserved by the manufacturer for that purpose?
In that case, I'm not sure it can be undone.
If things work as described in the literature, yes the mapping is done permanently unless you have some exotic tools that allow bad sectors to be put back in service (I don't own those).
Anyway, I had a client pay me to confirm "shred"s reality matched theory on a range of drives.
That client runs shred against 10's of thousands of hard disks a year. They wanted an independent assessment that it would work even in the presence of bad media, bad cables, etc. or that shred would properly error out if it could not run to completion for whatever reason. That client would then note in their normal process that shred failed to complete and physically destroy the drive. They are extremely conscientious of data destruction.
Greg -- Greg Freemyer
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1
On 2014-10-08 17:59, Greg Freemyer wrote:
On Wed, Oct 8, 2014 at 10:35 AM, Carlos E. R. <> wrote:
Question: would not the write attempt to a failed sector automatically remap it, to one of those sectors reserved by the manufacturer for that purpose?
In that case, I'm not sure it can be undone.
If things work as described in the literature, yes the mapping is done permanently unless you have some exotic tools that allow bad sectors to be put back in service (I don't own those).
Ah. Ok, noted that they do exist, but nobody has seen them, as the Yeti ;-)
Then I don't think I will test this marking thing, in case the sector is remapped for ever. I will not sacrifice myself for science ;-))
Anyway, I had a client pay me to confirm "shred"s reality matched theory on a range of drives.
That client runs shred against 10's of thousands of hard disks a year. They wanted an independent assessment that it would work even in the presence of bad media, bad cables, etc. or that shred would properly error out if it could not run to completion for whatever reason. That client would then note in their normal process that shred failed to complete and physically destroy the drive. They are extremely conscientious of data destruction.
I see.
- -- Cheers / Saludos,
Carlos E. R. (from 13.1 x86_64 "Bottle" at Telcontar)
-
Carlos E. R. -
Greg Freemyer