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

[Antonio Diaz Diaz]

Hello Narcis,

Narcis Garcia wrote:
> --max-errors didn't work as expected with ddrescue 1.17

Apparently it did. Every time the number of errors grown by more than 1, 
ddrescue exited with "Too many errors in input file".


> This is my command:
> $ ddrescue --force -n -e +1 /dev/sdb /dev/sdc /var/tmp/ddrescue-map.log
>
> This is the output when running repeatedly:
> (why errsize 35563TB on a 1TB disk ???)

Most probably because the OS misrecognized the drive size.


> GNU ddrescue 1.17
> Press Ctrl-C to interrupt
> Initial status (read from logfile)
> rescued:    97277 MB,  errsize:    299 kB,  errors:       8
> Current status
> rescued:    97277 MB,  errsize:  35563 TB,  current rate:        0 B/s
>    ipos:    35563 TB,   errors:       8,    average rate:        0 B/s
>    opos:    35563 TB,    time since last successful read:     1.1 m
> Copying non-tried blocks...terminate called after throwing an instance
> of 'std::bad_alloc'
>   what():  std::bad_alloc
> Aborted

This is very strange. Ddrescue just needs a few MB of RAM, and the 
amount of RAM used by ddrescue depends on the number of blocks (lines in 
the mapfile). But this mapfile seems to be quite short. 
('std::bad_alloc' usually means 'out of memory').


> rescued:    97277 MB,  errsize:   2355 TB,  current rate:        0 B/s
>    ipos:     2355 TB,   errors:       8,    average rate:        0 B/s
>    opos:     2355 TB,    time since last successful read:       5 s
> Copying non-tried blocks...^C
> Interrupted by user

Could you, please, interrupt ddrescue before it aborts and tell us the 
result of the following command?

  ddrescuelog -t /var/tmp/ddrescue-map.log


Best regards,
Antonio.