On 7/11/2014 12:42 PM, William wrote:
ddrescue -vvv -d -c 1 -r1 --skip-on-first-err /dev/sdb
Samsung-1500GB.dd Samsung-1500GB.dd.log
Interesting. You are reading one sector at a time and skipping. I have
not yet tested that combination. The advantage of skipping out of the
bad areas early on to get more good data first, but never read any
sector twice. Makes for slower reads of the good data, but overall would
still be faster.
PS: Why not creating a patch that only runs forwards with just one
pass, at least when using --no-retry; something like "-n
--only-forwards"? Would be hard to create it?
Let me understand what you are proposing. You would like an option to
stop after pass1 (would also include option to stop after pass2). This
option(s) would effectively stop the rescue after pass 1 or 2, before
getting to the slow no-skip pass. When ran again without the option,
ddrescue would continue from that point on (kind of like the no-trimming
and no-splitting options). If this is what you would like, I am not sure
how hard it would be without looking into it more. If it is doable, it
could make it into the next patch version.
PS: If you are running Linux, have you tried the sgio_passthrough patch
(I don't see it in your commands)? It can be used along with the
features patch. There has been great excitement about the features
patch, but I have heard nothing from anyone about the passthrough patch
which came out beforehand. It would appear that no one is excited about
something that can make ddrescue process errors 5 times faster.
Scott
_______________________________________________
Bug-ddrescue mailing list
[email protected]
https://lists.gnu.org/mailman/listinfo/bug-ddrescue
