Re: [Bug-ddrescue] Option to stop copy at first error

[Narcis Garcia]

If ddrescue queries disk size all the time to the Kernel, once the
device is unreadable/unreachable the Kernel cannot report true
information from that disk.

I'm working now on a solution to use simple dd with a script that
records a pointer of last sector read from source disk while that disk
is readable.


On 05/02/16 12:43, Antonio Diaz Diaz wrote:
> Narcis Garcia wrote:
>> current pos:     2355 TB,  current status: copying
>> domain size:     9223 PB,  in    1 area(s)
> 
> As I suspected, the kernel stopped reporting the size of /dev/sdb, and
> ddrescue is using the maximum size (9223 PB). You will need to
> power-cycle /dev/sdb, reboot, or provide ddrescue the right size with
> the '-s' option. (And maybe use the '--complete-only' option in
> subsequent runs).
> 
> 
>>     errsize:     2355 TB,  errors:       8  (  0.02%)
>> non-trimmed:     2355 TB,  in 2193946 area(s)  (  0.02%)
>>   non-split:         0 B,  in    0 area(s)  (  0%)
>>  bad-sector:     1123 MB,  in 2193938 area(s)  (  0.00%)
> 
> Ddrescue 1.17 treats this case specially bad because of the way it marks
> failed blocks. Maybe you could update to a newer version.
> 
> 
> Best regards,
> Antonio.