[Grml] How to use GRML to check whether a hard disk is failing

[Michael Schierl]

[sorry for off-topic posting...]

Am 22.12.2009 20:21, schrieb Josh Lawrence:
> On Tue, Dec 22, 2009 at 1:06 PM, David Maus <maus.david at gmail.com> wrote:
>> On the question how to check and/or fix a broken ntfs filesystem, I am
>> lost.
> 
> I missed the original post, but if you want to check the ntfs file
> system, you need an XP disk.  look up CHKDSK.

When you have cloned a dying hard disk with ddrescue (and were able to
only copy a fraction of the information) your first choice should not be
CHKDSK. CHKDSK primarily tries to restore the filesystem into a
consistent state, which means that directories that are only partially
readable will be completely lost completely afterwards - not very good
if you try to rescue data.

When you have the disk space available, create a clone from the clone
and try several tools on it (free and commercial ones). When the data is
"important", buying an additional external hard disk may be worth the
money (especially since it can be used afterwards, regardless whether
the data recovery was successful or not). Give each tool a clean clone
after having copied all the data it could rescue.

Note that cloning from a failing drive can take a long time (hours to
days), cloning the healthy clone afterwards is much faster.

In the (two) cases where I tried to rescue lost data from a friend's
hard disk, the commercial Windows tool "Get Data Back for NTFS" (from
Runtime software) was able to rescue most of the files where free
alternatives failed. There is also a trial version available that can
list you the files it could rescue (and can save a small number of files
so that you see it is right).

Of course, it's better to make backups and never need these tools.


Regards,

Michael