On 05/08/2015 at 01:20 PM, German wrote:
> On Fri, 08 May 2015 12:10:38 -0400 Gary Dale <garydale@torfree.net>
> wrote:
>
>> On 08/05/15 10:32 AM, German wrote:
>>> Hi list. Ok, now I have a spare 2TB USB drive where I can save
>>> .img file. Is that the right procedure? Do I have make a snapshot
>>> of failed drive and transfer it as a .img file to a spare drive,
>>> correct? R-studio for linux can display files of failed drive (
>>> TestDisk coudn't do it ). So now I think I'll proceed. What is
>>> exact command to do it with ddrescue and what file system the
>>> spare drive has to be formated? Thank you very much!
>>>
>>>
>> You can try ddrescue if=/dev/sdb1 of=failed.img where /dev/sdb1
>> would be the partition that you want to recover.
>
> Thanks, but some clarification is needed. Now I have two drives,
> failed and a spare. Both are 2TB in size. Failed drive probably has
> 1.6 TB data I'd like to recover. It has only one partition I
> suppose.
That's bad.
If the drive has only one partition, it probably has a single filesystem
taking up all of its space.
When you create a ddrescue image from that partition, the new image will
take up _at least as much_ space as the original filesystem. That's not
the 1.6TB of "used" space; it's the full 2TB of "total" space. (Plus
however much space is taken up by the "index" file used by ddrescue
while doing its work.)
That means that if your two 2TB drives are actually the same size, the
"good" one will not have enough space to store the image you need to
rescue from the "bad" one.
> So, if failed drive is for instance /dev/sdb and spare drive is for
> isntance /dev/sdc, the right command will be ddrescue if=/dev/sdb
> of=/dev/sdc/failed.img ?
No.
With ddrescue, you do not (want to) create a new filesystem directly on
the new device.
What you want to do is create a filesystem _image_, in a file which is
stored on the new device.
Step-by-step, what you do is:
* Create a filesystem on your "good" drive (/dev/sdc). You can use any
filesystem you want.
* Mount the newly created filesystem somewhere. For example, if you want
to mount it to the empty directory '/mnt/new_disk', you could run the
command:
mount /dev/sdc1 /mnt/new_disk
* Run the command
ddrescue /dev/sdb /mnt/new_disk/sdb_failed.img
/tmp/sdb_failed.ddrescuelog
Note that this will create _two_ files: the rescued filesystem image,
and a "log file" which ddrescue uses to keep track of what it has
already successfully rescued and where it has encountered errors. This
second file will take up additional space.
* If you have enough free space, copy the resulting sdb_failed.img to a
separate location, so you have a backup copy if something goes wrong.
* Run recovery tools (starting with fsck, as Gary Dale suggested) on
sdb_failed.img, and see how much you can recover.
* Mount sdb_failed.img, using a loopback mount, and copy out whatever
files you can to a new location.
Recovering data from a failed drive using ddrescue, dd_rescue, or
myrescue is entirely possible; I've done it myself. However, it is not
trivial, and in my experience the process requires considerably more
space than the size of whatever drive is being rescued.
Personally, I would probably recommend the use of myrescue rather than
of ddrescue or dd_rescue, but any of them can work if you use them right.
--
The Wanderer
The reasonable man adapts himself to the world; the unreasonable one
persists in trying to adapt the world to himself. Therefore all
progress depends on the unreasonable man. -- George Bernard Shaw
Attachment:
signature.asc
Description: OpenPGP digital signature