DISKRESCUE(1)              OpenBSD Reference Manual              DISKRESCUE(1)

NAME
     diskrescue - disk recovery tool

SYNOPSIS
     diskrescue [-aqv] [-R results file] [-b block size] [-o out file]
                [-r raw device] operation

DESCRIPTION
     diskrescue is a tool to verify and/or recover as much data as possible
     from failed disks.

     The idea is to read all sectors of a disk and let the disk reallocate bad
     or soon to be bad sectors.  While the disk is being read it can clone the
     data to a file or a disk as well.  The resulting disk/file can be checked
     with fsck(8) and analyzed with dumpfs(8) and fsdb(8) while not having to
     deal with a failing disk.

     When a sector can not be read it will be filled with zeroes and
     diskrescue continues on to the next sector.  For all intents and purposes
     this is a fancy dd(1) that takes some of the complexity out of dealing
     with a failed disk.

     The options are as follows:

     -R results file
             Write the results of the rescue process to the specified file.
             If the results file already exists the program will attempt to
             restart the recovery.  Other parameters such as block size and
             out file can not be specified when attempting to restart a recov-
             ery.

     -a      Abort on read error.

     -b block size
             This is the block size that the disk uses to read and write data.
             A large value speeds things up and once a bad spot on the disk is
             hit diskrescue reverts to the natural block size of 512 bytes
             during the recovery process.  Default is 512, which is very slow!
             The argument accepts the following quantifiers: b, k, m, g and
             the following multipliers: * and x.  For example 1 mega byte can
             be expressed as: -b 1m or -b 1k*1k

     -o out file or disk
             This is the target file or disk of the operation.

     -q      Don't report completion percentage.

     -r raw disk
             This is the raw disk device that will be attempted to be recov-
             ered.  For example: /dev/rsd2c

     -v      Print version number and exit.

     operation
                   verify  The verify operation reads the entire disk and lets
                           it do its thing to reallocate bad or questionable
                           blocks.
                   recover
                           Essentially the same as verify but write the read
                           data to a file or disk.

SEE ALSO
     dd(1) fsck(8) dumpfs(8) fsdb(8)

HISTORY
     diskrescue was written to recover a disk of a friend and it became useful
     enough for the general public.

AUTHORS
     diskrescue was written by Marco Peereboom <marco@peereboom.us>.

EXAMPLES
     Verify a disk and force it to reallocate bad blocks:

           diskrescue -b 1m -r /dev/rsd2c verify

     Recover a disk to a file and write a results file:

           diskrescue -r /dev/rsd2c -R res -o disk.img -b 1m recover

     Restart a recovery operation with a results file:

           diskrescue -r /dev/rsd2c -R res recover

DISCLAIMER
     Use this at your own risk!  Recovering disks can make them worse and the
     author takes no responsibility for any data loss or corruption.  Also
     when used wrong a good disk could be overwritten permanently losing all
     data on it.  Use this tool wisely and leave the author alone if you mess
     up.

BUGS
     Lots of features are still missing.

OpenBSD 4.6                   September 27, 2009                             2