Andrea Mastellone wrote, On 09/24/2008 01:10 PM:
Todd Denniston wrote:
Andrea Mastellone wrote, On 09/24/2008 11:12 AM:
Hi,
if I had less space (and no time to buy an extra USB hard drive) and
could translate the CHS to bytes/blocks of disk space, I would do
something like:
assume 512 blocks, and HeadSector size of 7697074bytes
bc of 7697074*6402/512 yields 96243491 which is probably completely
wrong.
(is this roughly a 55GB device?)
Yes, it is about 55 GiB device. Can i proceed so with this ?
No warranties, you get to keep the broken chunks.
I was just playing about with the math based on the info that fdisk puts out
in the 'p' information, which for my disk was:
Units = cylinders of 16065 * 512 = 8225280 bytes
and I was guessing was:
Units = cylinders of 15033 * 512 = 7697074 bytes
based on you indicating the partition in question was 10GB and THAT partition
having 7798-6403=1395 cylinders
10*1024*1024*1024/1395=7697074
So these are VERY round about numbers... If you are not going to capture the
whole disk, then I would capture significantly more than the calculations show
and then trim that down to work with.
If I _/HAD/_ to capture less than 55GB of the device, with only the
insufficient data I have now, I would do a something more like setting seek to
the order of 7697074*6000/512=90200085 and then start looking for filesystem
headers in the data.
My preference though would be capture the whole device, because it gives you a
lot more chances to come up with the correct data. One example of more
chances, would be to 'fix' the partition table in a copy of the image, i.e,
you know the partition ran from 6403 to 7798 and something like Partitioned
loopback devices[1] MIGHT be able to help.
Granted we are already WAY outside my comfort zone in giving advise about
recovering data from a device that I don't own. If it is really important,
there are folks who will do the whole job for a fee.
Again no warranties, you get to keep the broken chunks.
dd if=/dev/whole_solid_state_disk \
of=/path/to/spare11GB/SSD.image \
seek=96243491 conv=noerror bs=512
Andrea
And remember, the device is already failing... you MIGHT only get one
more read, so take that into consideration when choosing the method
for getting the image.
[1] http://lwn.net/Articles/110468/
http://lwn.net/Articles/251908/
http://lwn.net/Articles/274113/
--
Todd Denniston
Crane Division, Naval Surface Warfare Center (NSWC Crane)
Harnessing the Power of Technology for the Warfighter
--
fedora-list mailing list
fedora-list@xxxxxxxxxx
To unsubscribe: https://www.redhat.com/mailman/listinfo/fedora-list
Guidelines: http://fedoraproject.org/wiki/Communicate/MailingListGuidelines
