1 Introduction ¶

GNU ddrescue is a data recovery tool. It copies data from one file or block device (hard disc, cdrom, etc) to another, trying to rescue the good parts first in case of read errors.

The basic operation of ddrescue is fully automatic. That is, you don’t have to wait for an error, stop the program, restart it from a new position, etc.

If you use the mapfile feature of ddrescue, the data are rescued very efficiently, (only the blocks needed are read). Also you may interrupt the rescue at any time and resume it later at the same point. The mapfile is an essential part of ddrescue’s effectiveness. Use it unless you know what you are doing.

Ddrescue does not write zeros to the output when it finds bad sectors in the input, and does not truncate the output file if not asked to. So, every time you run it on the same output file, it tries to fill in the gaps without wiping out the data already rescued.

Automatic merging of backups: If you have two or more damaged copies of a file, cdrom, etc, and run ddrescue on all of them, one at a time, with the same output file, you will probably obtain a complete and error-free file. This is so because the probability of having the same area damaged in all copies is low (if the errors are randomly located). Using the mapfile, only the blocks needed are read from the second and successive copies.

Ddrescue recommends lzip for compression of backups because the lzip format is designed for long-term archiving and provides data recovery capabilities which nicely complement those of ddrescue. (Ddrescue fills unreadable sectors with data from other copies, while lziprecover corrects corrupt sectors with data from other copies). If the cause of file corruption is a damaged medium, the combination ddrescue + lziprecover is the best option for recovering data from multiple damaged copies. See lziprecover-example.

Because ddrescue needs to read and write at random places, it only works on seekable (random access) input and output files. Therefore, the only way of creating a compressed image with ddrescue is to create a normal (uncompressed) image first, and then compress that image.

If your system supports it, ddrescue can use direct disc access to read the input file, bypassing the kernel cache.

One of the strengths of ddrescue is that it is interface-agnostic, and so can be used for any kind of device supported by your kernel (ATA, SATA, SCSI, old MFM drives, floppy discs, or even flash media cards like SD).

Ddrescue also features a ’fill mode’ able to selectively overwrite parts of the output file, which has a number of interesting uses like wiping data, marking bad areas, or even, in some cases, "repair" damaged sectors.

2 Basic concepts ¶

Block

Any amount of data. A block is described by its starting position and its size. The starting position (or beginning position) is the lowest position in the block. The end of the block is its starting position plus its size.

Cluster

Group of consecutive sectors read or written in one go.

Device

Piece of hardware containing data. Hard disc drives, cdrom drives, USB pendrives, are devices. /dev/hda, /dev/sdb, are device names (file names associated to devices).

File

Files are named units of data which are stored by the operating system for you to retrieve later by name. Devices and partitions are accessed by means of their associated file names.

Partition

Every part in which a device is divided. A partition normally contains a file system. /dev/hda1, /dev/sdb3, are partition names (file names associated to partitions).

Recoverable formats

As ddrescue uses standard library functions to read data from the device being rescued, only mountable device formats can be rescued with ddrescue. CD-ROMs and DVDs can be rescued, "compact disc digital audio" CDs can’t, "video CDs"[1] maybe.
[1] http://en.wikipedia.org/wiki/Video_CD

Rescue domain

Block or set of blocks to be acted upon (rescued, listed, etc). You may define it with the options --input-position, --size, and --domain-mapfile. The rescue domain defaults to the whole input file or mapfile. If ddrescue can’t determine the size of the input file, the rescue domain defaults to the maximum size of a block (at least 2^63 - 1 bytes, or 8 EiB minus 1 byte).

Ddrescue never tries to read any data outside the rescue domain except when unaligned direct disc access is requested (see Direct disc access). If it does, please, report it as a bug.

The data shown by ddrescue (amount of data rescued, number of bad areas, etc) may vary or even become zero if you limit the rescue domain. Don’t worry, they have not disappeared; they are simply out of the rescue domain specified.

Sector

Hardware block. Smallest accessible amount of data on a block device.

3 Using ddrescue safely ¶

Ddrescue is like any other power tool. You need to understand what it does, and you need to understand some things about the machines it does those things to, in order to use it safely.

Never try to rescue a r/w mounted partition. The resulting copy may be useless. It is best that the device or partition to be rescued is not mounted at all, not even read-only.

Never try to repair a file system on a drive with I/O errors; you will probably lose even more data.

If you use a device or a partition as destination, any data stored there will be overwritten.

Some systems may change device names on reboot (for example, udev enabled systems). If you reboot such a system, pass the option --ask to ddrescue and check that the device names match the model and serial number of the input and output devices before allowing ddrescue to proceed. You may also check the devices with commands like ‘hdparm -I /dev/sda’ or ‘smartctl -i /dev/sda’.

If you interrupt the rescue and then reboot, any partially copied partitions should be hidden before allowing them to be touched by any operating system that tries to mount and "fix" the partitions it sees.

When using a (graphical) frontend, be careful to not insert or remove any devices to avoid changes in the mapping of devices to device names between the moment the devices are selected in the frontend and the call to ddrescue. Be careful even if the frontend passes --ask to ddrescue, because outfile may be changed between user confirmation and actual opening.

4 Algorithm ¶

GNU ddrescue is not a derivative of dd, nor is related to dd in any way except in that both can be used for copying data from one device to another. The key difference is that ddrescue uses a sophisticated algorithm to copy data from failing drives causing them as little additional damage as possible.

Versions of ddrescue prior to 1.19 used a divide-and-conquer strategy to rescue the difficult parts of the drive. But that caused a lot of head movement, which is bad for the drive. Therefore, newer versions try to minimize head movement to minimize drive damage.

Ddrescue manages efficiently the status of the rescue in progress and tries to rescue the good parts first, scheduling reads inside bad (or slow) areas for later. This maximizes the amount of data that can be finally recovered from a failing drive.

The standard dd utility can be used to save data from a failing drive, but it reads the data sequentially, which may wear out the drive without rescuing anything if the errors are at the beginning of the drive.

Other programs read the data sequentially but switch to small size reads when they find errors. This is a bad idea because it means spending more time at error areas, damaging the surface, the heads, and the drive mechanics, instead of getting out of them as fast as possible. This behavior reduces the chances of rescuing the remaining good data.

The algorithm of ddrescue is divided in four phases: copying, trimming, scraping, and retrying. Each phase is described below. Disc sectors are marked successively as non-tried, non-trimmed, non-scraped, and bad-sector until they are successfully read and marked as finished. The user may interrupt the process at any point, but a bad drive can block ddrescue for a long time until the kernel gives up.

The amount of work remaining for a given phase can be calculated by comparing the current size of the corresponding areas with their size at the end of the previous pass. Namely the size of non-tried while copying, the size of non-trimmed while trimming, the size of non-scraped while scraping, and the size of bad-sector while retrying. See Meaning of ddrescue’s screen output.

The steps of the algorithm are:

1) Optionally read a mapfile describing the status of a multi-part or previously interrupted rescue. If no mapfile is specified, or is empty, or does not exist, mark all the rescue domain as non-tried.

2) (First phase; Copying) Copying is done in up to five passes. The first pass reads the non-tried parts of the input file, marking the failed blocks as non-trimmed and skipping beyond them. The second pass runs in the opposite direction as the first pass and delimits the blocks skipped by the first pass. The first two passes also skip beyond slow areas. The areas skipped are tried later in one or three additional passes (before trimming). The copying direction is reversed after each pass until all the rescue domain is tried.

The third and fourth passes read the blocks skipped due to slow areas (if any) by the first two passes, in the same direction that each block was skipped. For each block, passes 2 to 4 skip the rest of the block after finding the first error in the block. The last pass is a sweeping pass, with skipping disabled. The purpose of the multiple passes is to delimit large bad areas fast, recover the most promising areas first, keep the mapfile small, and produce good starting points for trimming.

Only non-tried areas are read in large blocks. Trimming, scraping, and retrying are done sector by sector. Each sector is tried at most two times: the first in the copying phase as part of a large block read, the second in one of the phases below as a single sector read.

3) (Second phase; Trimming) Trimming retries sector by sector the edges of the large block reads failed during the copying phase. Trimming is done in one pass as follows. For each non-trimmed block, read forwards one sector at a time from the leading edge of the block until a bad sector is found. Then read backwards one sector at a time from the trailing edge of the block until a bad sector is found. Then mark the bad sectors found (if any) as bad-sector, and mark the rest of the block as non-scraped without trying to read it. If any edge is already adjacent to a bad sector, it is considered as already trimmed and is not trimmed again.

4) (Third phase; Scraping) Scrape together, sector by sector, the data not recovered by the copying or trimming phases. Scraping is done in one pass. Each non-scraped block is read forwards, one sector at a time. Any bad sectors found are marked as bad-sector.

5) (Fourth phase; Retrying) Optionally try to read again the bad sectors until the number of retry passes specified is reached. The direction is reversed after each pass. Every bad sector is tried only once in each pass. Ddrescue can’t know if a bad sector is unrecoverable or if it will be eventually read after some retries.

6) Optionally write a mapfile for later use.

When ddrescue finishes the steps above, any areas marked as bad-sector will remain untouched in the output file. If the output file is a regular file created by ddrescue, the areas marked as bad-sector will contain zeros. If it is a device or a previously existing file, the areas marked as bad-sector will still contain the data previously present there.

The mapfile is periodically saved to disc, as well as when ddrescue finishes or is interrupted. A backup copy of the mapfile with the extension ‘.bak’ is also periodically created (if possible). So in case of a crash you can resume the rescue with little recopying. The default interval between saves varies from 30 seconds to 5 minutes depending on mapfile size (larger mapfiles are saved at longer intervals), but may be overriden. See --mapfile-interval.

The same mapfile can be used for multiple commands that copy different areas of the input file, and for multiple recovery attempts over different subsets. See this example:

Rescue the most important part of the disc first.

ddrescue -i0 -s50MiB /dev/sdc hdimage mapfile
ddrescue -i0 -s1MiB -d -r3 /dev/sdc hdimage mapfile

Then rescue some key disc areas.

ddrescue -i30GiB -s10GiB /dev/sdc hdimage mapfile
ddrescue -i230GiB -s5GiB /dev/sdc hdimage mapfile

Now rescue the rest (does not recopy what is already done).

ddrescue /dev/sdc hdimage mapfile
ddrescue -d -r3 /dev/sdc hdimage mapfile

5 Meaning of ddrescue’s screen output ¶

The output of ddrescue looks like this:

GNU ddrescue 1.28
Press Ctrl-C to interrupt
Initial status (read from mapfile)
rescued: 1665 MB, tried: 0 B, bad-sector: 0 B, bad areas: 0

Current status
     ipos:    2874 MB, non-trimmed:        0 B,  current rate:  21479 kB/s
     opos:    2874 MB, non-scraped:        0 B,  average rate:  21023 kB/s
non-tried:   13603 MB,  bad-sector:        0 B,    error rate:       0 B/s
  rescued:    2401 MB,   bad areas:        0,        run time:         35s
pct rescued:   15.00%, read errors:        0,  remaining time:         10m
 slow reads:        5,        time since last successful read:          0s
Copying non-tried blocks... Pass 1 (forwards)

Ddrescue scrolls forward after each pass. This keeps on the screen the final status of the previous pass, making it easier to estimate the amount of work done by the current pass.

The meaning of each field is as follows:

ipos

Input position. The position in the input file where data are being currently read from.

opos

Output position. The position in the output file where data are being currently written to.

non-tried

Size of the part of the rescue domain pending to be tried. This is the sum of the sizes of all the non-tried blocks.

rescued

Size of the part of the rescue domain already successfully recovered. This is the sum of the sizes of all the finished blocks.

pct rescued

Percentage of the rescue domain that has been successfully recovered.

slow reads

Number of times that the read rate fell below --min-read-rate during the first two passes of the copying phase. See --min-read-rate.

tried

Size of the part of the rescue domain already tried but not yet rescued. This is the sum of the sizes of all the non-trimmed, non-scraped, and bad-sector blocks.

non-trimmed

Size of the part of the rescue domain pending to be trimmed. This is the sum of the sizes of all the non-trimmed blocks.

non-scraped

Size of the part of the rescue domain pending to be scraped. This is the sum of the sizes of all the non-scraped blocks.

bad-sector

Total error size. This is the size of the part of the rescue domain formed by known bad sectors. The total error size is the sum of the sizes of all the bad-sector blocks. It increases during the trimming and scraping phases, and may decrease during the retrying phase. A sector is not marked as bad-sector and considered part of a bad area until it has been tried individually instead of as part of a large block read. Note that as ddrescue retries the bad-sector blocks, the good data found may divide them into smaller blocks, decreasing the total error size but increasing the number of bad areas.

bad areas

Number of separate bad-sector blocks inside the rescue domain. Non-trimmed and non-scraped blocks are not considered bad areas. See --max-bad-areas.

read errors

Number of failed read attempts. See --max-error-rate.

current rate

The read rate measured during the last second.

average rate

The average read rate measured during the current run.

error rate

The read error rate measured during the last second.

run time

Time elapsed since the beginning of the current run.

remaining time

Estimated remaining time to rescue all the data in the rescue domain. The remaining time is calculated using the average rate of the last 30 seconds and does not take into account that some parts of the rescue domain may be excluded from the rescue (for example with --no-trim), or that some areas may be unrecoverable. Therefore it may be very imprecise, may vary widely during the rescue, and may show a non-zero value at the end of the rescue. In particular it may go down to a few seconds at the end of the first pass, just to grow to hours or days in the following passes. Such is the nature of ddrescue; the good parts are usually recovered fast, while the rest may take a long time.

time since last successful read

Time elapsed since the last successful read attempt.

6 Invoking ddrescue ¶

The format for running ddrescue is:

ddrescue [options] infile outfile [mapfile]

infile and outfile may be files, devices, or partitions. mapfile is a regular file and must be placed in an existing directory. If mapfile does not exist, ddrescue creates it. Be careful to not specify by mistake an old mapfile from an unrelated rescue.

Ddrescue tries to create a backup copy of the mapfile, with the name mapfile.bak, every time it is going to overwrite a fsynced mapfile. See --mapfile-interval.

Always use a mapfile unless you know you won’t need it. Without a mapfile, ddrescue can’t resume a rescue, only reinitiate it.

ddrescue supports the following options:

-h

--help

Print an informative help message describing the options and exit.

-V

--version

Print the version number of ddrescue on the standard output and exit. This version number should be included in all bug reports.

-a bytes

--min-read-rate=bytes

Minimum read rate of good non-tried areas, in bytes per second. If the read rate falls below this value during the first two passes of the copying phase, ddrescue skips ahead a variable amount depending on rate and error histories. The blocks skipped are tried in additional passes (before trimming). --min-read-rate is ignored in all passes but the first two.

If bytes is 0 (auto), the minimum read rate is recalculated every second as (average_rate / 10).

-A

--try-again

Mark all non-trimmed and non-scraped blocks inside the rescue domain as non-tried before beginning the rescue. Try this if the drive stops responding and ddrescue immediately starts scraping failed blocks when restarted. If --retrim is also specified, mark all failed blocks inside the rescue domain as non-tried.

-b bytes

--sector-size=bytes

Sector (hardware block) size of input device in bytes (usually 512 for hard discs and 3.5" floppies, 1024 for 5.25" floppies, and 2048 for cdroms). Defaults to 512.

In rescue mode, any non-finished subsector that is found during the initial read of the mapfile is joined to its corresponding sector (if it is also not finished), marking the whole sector with the less processed state, so as to make sure that sub-sector data is not discarded from a successful read during the rescue. (A subsector is a block smaller than sector size). Subsector joining is performed in all the mapfile, not only in the rescue domain.

-B

--binary-prefixes

Show units with binary prefixes (powers of 1024).
SI prefixes (powers of 1000) are used by default. (See table below).

-c sectors

--cluster-size=sectors

Number of sectors to copy at a time. Defaults to 64 KiB / sector_size. Try smaller values for slow drives. The number of sectors per track (18 or 9) is a good value for floppies.

-C

--complete-only

Limit rescue domain to the blocks listed in the mapfile. Don’t read new data beyond mapfile limits. This is useful when reading from devices of undefined size (like raw devices), when the drive returns an incorrect size, or when reading from a partial copy. -C can only be used after a first rescue attempt, possibly limited with the option --size, has produced a complete mapfile. If -C is not specified, ddrescue extends the mapfile from position 0 to the size of outfile, adding one non-tried block before the existing blocks and another one after them if needed.

-d

--idirect

Use direct disc access (see Direct disc access) to read from infile, bypassing the kernel cache. (Opens the file with the flag ‘O_DIRECT’). Sector size must be correctly set for this to work. Not all systems support this.

If your system does not support direct disc access, ddrescue warns you. If the sector size is not correctly set, an unaligned read error may happen, in which case ddrescue exits with error status 1.

-D

--odirect

Use direct disc access to write to outfile, bypassing the kernel cache. (Opens the file with the flag ‘O_DIRECT’). Sector size must be correctly set for this to work. Not all systems support this.

If your system does not support direct disc access, ddrescue warns you. If the sector size is not correctly set, a write error is reported and no data is rescued. Some OSs have a bug that prevents them from detecting write errors properly (or at all) on some devices if direct disc access is not used for outfile.

-e [+]n

--max-bad-areas=[+]n

Maximum number of bad areas allowed before giving up. Defaults to infinity. If n is preceded by ‘+’ the number refers to new bad areas found in this run, not counting those already present in the mapfile.

-E bytes

--max-error-rate=bytes

Maximum rate of read errors allowed before giving up, in bytes per second. Defaults to infinity. The rate being measured is that of actually failed reads, so ddrescue may exit because of this rate being exceeded even if the total error size (size of bad-sector areas) does not change because the areas being tried are being marked as non-trimmed or non-scraped, or are already marked as bad-sector.

-f

--force

Force overwrite of outfile. Needed when outfile is not a regular file, but a device or partition. This option is just a safeguard to prevent the inadvertent destruction of partitions, and is ignored for regular files.

-F types

--fill-mode=types

Fill the blocks in outfile specified as any of types in mapfile, with data read from infile. types contains one or more of the status characters defined in the chapter Mapfile structure (see Mapfile structure) and an optional ‘l’ for sector location data. See the chapter Fill mode (see Fill mode) for a complete description of the fill mode.

-G

--generate-mode

Generate an approximate mapfile from the infile and outfile of the original rescue run. Note that you must keep the original offset between --input-position and --output-position of the original rescue run. See the chapter Generate mode (see Generate mode) for a complete description of the generate mode.

-H file

--test-mode=file

Build a map of good/bad blocks using the mapfile file and use it to simulate read errors in infile. The blocks marked as finished in file are read normally. All other block types are considered read errors without even trying to read them from infile. The apparent size of infile is truncated to the extent of file. This mode is an aid in improving the algorithm of ddrescue and is also useful to check that ddrescue produces accurate results in presence of read errors. Use a hyphen ‘-’ as file to read the mapfile from standard input.

-i bytes

--input-position=bytes

Starting position of the rescue domain in infile, in bytes. Defaults to 0. This is not the point from which ddrescue starts copying. (For example, if you pass the option --reverse to ddrescue, it starts copying from the end of the rescue domain). In fill mode it refers to a position in the infile of the original rescue run. See the chapter Fill mode (see Fill mode) for details.

-I

--check-input-size

Compare the size of infile with the size calculated from the list of blocks contained in the mapfile, and exit with status 1 if they differ. This is not enabled by default because the size of some devices can’t be known in advance and because the size derived from the mapfile may be incomplete, for example after doing a partial rescue.

-J

--check-on-error

After every read error, read again the last good sector found and check that it returns the same data. Exit with status 2 if the read fails or returns inconsistent data. Exit with status 1 if a read error happens before a good sector is found.

This option performs one extra read after each error, wearing the drive faster. Use it only on drives that stop responding or return garbage data after finding errors. You may need to power cycle the drive before restarting ddrescue.

-K [initial][,max]

--skip-size=[initial][,max]

Set limits to skip size during the copying phase. At least one of initial or max must be specified. initial is the size to skip on the first read error or slow read, in bytes. max is the maximum size to skip. The values given are rounded to the next multiple of sector size. The skip size is doubled for each read error or slow read until it reaches max or, if max is omitted, 1% of the size of infile, and is reset to initial when good data are found. Valid values range from 64 KiB to 1 EiB. initial defaults to infile_size / 100_000 with a minimum value of 64 KiB. An initial value of 0 disables skipping entirely.

If ddrescue is having difficulties skipping away from a large area with scattered errors, or if the device has large bad areas at regular intervals, you may increase the initial skip size with this option. Inversely, if ddrescue is skipping too much, leaving large non-tried areas behind each error (which will be read later in the usually slower backwards direction), you may reduce the maximum skip size, or disable skipping.

--skip-size is independent from --cluster-size. The size to skip is calculated from the end of the block that just failed.

-L

--loose-domain

Accept an incomplete synthetic (user fabricated) domain mapfile or test-mode mapfile, and fill the gaps in the list of data blocks with non-tried blocks. The blocks in the mapfile may be unordered, may overlap other blocks of the same status, and don’t need to be contiguous. This option allows making quick edits to a mapfile without all the size calculations involved in making all data blocks contiguous again.

-m file

--domain-mapfile=file

Restrict the rescue domain to the blocks marked as finished in the mapfile file. This is useful for merging partially recovered images of backups, or if the destination drive fails during the rescue. Use a hyphen ‘-’ as file to read the domain mapfile from standard input. Specialized tools like ddrutility or partclone can produce a domain mapfile listing all the used blocks in a partition, making the rescue more efficient.

-M

--retrim

Mark all failed blocks inside the rescue domain as non-trimmed before beginning the rescue. The effect is similar to --retry-passes=1, but the bad sectors are tried in a different order, making perhaps possible to rescue some of them.

-n

--no-scrape

Skip the scraping phase. Avoids spending a lot of time trying to rescue the most difficult parts of the file.

-N

--no-trim

Skip the trimming phase. Especially useful in the first parts of a multi-part rescue.

-o bytes

--output-position=bytes

Starting position of the image of the rescue domain in outfile, in bytes. Defaults to --input-position. The bytes below bytes aren’t touched if they exist and truncation is not requested. Else they are set to 0.

-O

--reopen-on-error

Close infile and then reopen it after every read error encountered during the copying phase. If --min-read-rate is set, also close and reopen infile after every slow read encountered during the first two passes of the copying phase. Use this option if you notice a permanent drop in transfer rate after finding read errors or slow areas. But be warned that most probably the slowing-down is intentionally caused by the kernel in an attempt to increase the probability of reading data from the device.

-p

--preallocate

Preallocate space on disc for outfile. Only space for regular files can be preallocated. If preallocation succeeds, rescue will not fail due to lack of free space on disc. If ddrescue can’t determine the size to preallocate, you may need to specify it with some combination of the options --input-position, --output-position, --size, and --domain-mapfile.

-P[lines]

--data-preview[=lines]

Show lines lines of the latest data read in ‘16-byte hex + ASCII’ format. Valid values for lines range from 1 to 32. If lines is omitted, a default value of 3 is used.

-q

--quiet

Quiet operation. Suppress all messages.

-r n

--retry-passes=n

Exit after the given number of retry passes. Defaults to 0. -1 means infinity. Every bad sector is tried only once in each pass. The direction is reversed after each pass. To retry bad sectors detected on a previous run, you must specify a non-zero number of retry passes.

A command like ‘ddrescue -f -r-1 /dev/sdcard /dev/null mapfile’ can be used to read repeatedly until the device controller succeeds and remaps the bad sectors internally.

-R

--reverse

Reverse the direction of all passes (copying, trimming, scraping, and retrying). Every pass that is normally run forwards is now run backwards, and vice versa. -R does not modify the size of the blocks copied during each phase, just the order in which they are tried.

-s bytes

--size=bytes

Maximum size of the rescue domain in bytes. It limits the amount of input data to be copied. -1 removes any previous size limit. If ddrescue can’t determine the size of the input file, you may need to specify it with this option. Note that this option does not specify the size of the resulting outfile. For example, the following command creates an outfile 300 bytes long, but only writes data on the last 200 bytes:

ddrescue -i 100 -s 200 infile outfile mapfile

-S

--sparse

Use sparse writes for outfile. (The blocks of zeros are not actually allocated on disc). May save a lot of disc space in some cases. Not all systems support this. Only regular files can be sparse. Use this option only with an empty or zeroed outfile because if a block in outfile contains non-zero data, it won’t be overwritten by a corresponding block of zeros in infile, resulting in a corrupt copy.

-t

--truncate

Truncate outfile to zero size before writing to it. Only works for regular files, not for drives or partitions.

-T interval

--timeout=interval

Maximum time since last successful read allowed before giving up. Defaults to infinity. interval is an integer or rational number (like 1.5 or 1/2) optionally followed by one of ‘s’, ‘m’, ‘h’, or ‘d’, meaning seconds, minutes, hours, and days respectively. If no unit is specified, it defaults to seconds. interval has a resolution of one second; fractions of a second are not allowed.

-u

--unidirectional

Run all passes in the same direction. Forwards by default, or backwards if the option --reverse is also given.

-v

--verbose

Verbose mode. Further -v’s (up to 4) increase the verbosity level. Some large numbers in messages (like device sizes) are printed in groups of 3 digits separated by underscore characters to make them more readable.

-w

--ignore-write-errors

Make fill mode ignore write errors. This is useful to avoid ddrescue exiting because of new bad sectors developing while wiping the good sectors of a failing drive. Fill mode normally writes to outfile one cluster at a time. With this option, after the first write error is found in an area, the rest of that area is filled sector by sector.

Note that in rescue mode a write error is fatal, which means that the rescue needs to be repeated or else outfile needs to be copied to a third drive using mapfile as domain (see --domain-mapfile).

-W

--compare-before-write

Omit superfluous writes in rescue mode. Before writing each block of data to outfile, compare the data already present there with the data about to be written and, if they match, omit the write. Possible advantages are that on some devices reads are faster than writes and it may reduce wear when writing to a solid state device.

-x bytes

--extend-outfile=bytes

Extend the size of outfile to make it at least bytes long. If the size of outfile is already equal or longer than bytes, then this option does nothing. Use this option to guarantee a minimum size for outfile. Only regular files can be extended.

-X n

--max-read-errors=n

Maximum number of read errors allowed before giving up. Defaults to infinity. Exit with status 1 if more than n read errors are encountered. --max-read-errors=0 is similar but different to --timeout=0, which waits until the screen status is refreshed (at least 1 second). If there is at least one successful read per second, --timeout=0 does not make ddrescue to exit.

--max-read-errors=0 is also similar but different to --max-bad-areas=+0, which exits when a new bad area is found. If the read errors are adjacent to existing bad areas, no new bad areas are produced (just enlarged), and --max-bad-areas=+0 does not make ddrescue to exit.

-y

--synchronous

Use synchronous writes for outfile. (Issue a fsync call after every write). May be useful when forcing the drive to remap its bad sectors. Use it to make sure that all writes have been committed to disc when ddrescue finishes. Else the kernel may cache all the writes and pretend that it has finished.

-Z bytes

--max-read-rate=bytes

Maximum read rate, in bytes per second. If bytes is too small, the actual read rate is rounded up to the equivalent of a whole number of cluster reads per second. Use this option to limit the bandwidth used by ddrescue, for example when recovering over a network.

--ask

Ask for user confirmation before starting the copy. If the first letter of the answer is ‘y’, ddrescue starts copying. Else it exits with status 1.

If they can be obtained, ddrescue shows the model and serial number of the input and output devices. Ddrescue also shows the size in bytes of the corresponding file or device if it exists. The format used is [model::serial_number] (size)

--command-mode

Read commands from the standard input and execute them, copying parts of the input file on demand. Command-line arguments controling the display (like --data-preview) or the automatic algorithm (like --max-errors or --reverse) have no effect in command mode. See Copying parts of the input file on demand, for a complete description of the command mode.

--cpass=range

Select what pass(es) to run during the copying phase. Valid pass values range from 1 to 5. To run only the given pass(es), specify also --no-trim and --no-scrape. --cpass=0 skips the copying phase entirely.

Examples of range	Passes run
1	1
1,2,3	1, 2, 3
2-4	2, 3, 4
1,3-5	1, 3, 4, 5
1-3,5	1, 2, 3, 5

--delay-slow=interval

Initial delay before ddrescue starts checking for slow reads. Defaults to 30 seconds. interval is formatted as in the option --timeout above.

--log-events=file

Log all significant events (start of each pass and end of run) in file. If file already exists, the new events are appended at the end of file. For each event a line is printed containing a time stamp, the percentage rescued, and a message describing the event. The ‘end of run’ line also contains the current position and status. If ddrescue exits because of an error or interruption, the cause is also logged in file.

--log-rates=file

Log rates and error sizes every second in file. If file already exists, it is overwritten. Every time the screen is updated with new details, some of those details (time, input position, current and average rates, number of bad areas, and total error size) are written to file in a format usable by plotting utilities like gnuplot. This allows a posterior analysis of the drive to see if it has any weak zones (areas where the transfer rate drops well below the sustained average).

--log-reads=file

Log all read operations in file. If file already exists, it is overwritten. Every read attempt and its result (position, size, copied size, and error size) is written to file. (The position written is always the beginning of the block tried, even if reading backwards). A line is also written at the beginning of each phase (copying, trimming, scraping, and retrying). Finally, a line with a time mark is written every second (unless the read takes more time). Use this option with caution because file may become very large very quickly. Use lzip to compress file if you need to store or transmit it.

--mapfile-interval=[save_interval][,sync_interval]

Change the interval at which ddrescue saves and fsyncs the mapfile. At least one of save_interval or sync_interval must be specified. A save_interval of -1 chooses the default automatic interval (from 30 seconds to 5 minutes depending on mapfile size). A save_interval of 0 saves the mapfile after every read (use with caution). sync_interval is the interval between fsync calls. Default sync_interval is 5 minutes. Minimum sync_interval is 5 seconds. sync_interval must be greater or equal than save_interval. Intervals are formatted as in the option --timeout above.

In practice, fsyncs are a subset of saves. I.e., some of the times when the mapfile is saved, it is also fsync’ed. Therefore, --mapfile-interval=30,45 is really --mapfile-interval=30,60. The time needed to write the mapfile is excluded from the mapfile save and sync intervals. (Some mapfiles may take several seconds to write).

--max-slow-reads=n

Maximum number of slow reads allowed before giving up. Defaults to infinity. Exit with status 1 if more than n slow reads are encountered during the first two passes of the copying phase. Only works if a minimum read rate has been set with --min-read-rate.

--pause-on-error=interval

Time to wait after each read error. Also after each slow read if a minimum read rate has been set with --min-read-rate. Defaults to 0. interval is formatted as in the option --timeout above. If interval begins with ‘s’, the pause is simulated and interval can be smaller than one second; the time displayed is increased by interval but without performing any pause. Pause simulation can be useful in combination with --test-mode for testing purposes.

--pause-on-pass=interval

Time to wait between passes. Defaults to 0. interval is formatted as in the option --timeout above.

--reset-slow

Reset the slow reads counter every time the read rate reaches or surpasses --min-read-rate. With this option, ddrescue only exits after the read rate has remained below --min-read-rate for at least as many seconds as the argument given to --max-slow-reads.

--same-file

Allow infile and outfile to be the same file or device. This may be used to test the writing ability of a drive. It may also be used to copy part of a file to another location inside or beyond the end of the same file by setting different values for --input-position and --output-position. If the data to be copied overlap with the destination, the right copying direction must be chosen to avoid overwriting the overlapping part before it is copied.

Numbers given as arguments to options (positions, sizes, rates, etc) may be expressed as decimal, hexadecimal, or octal values (using the same syntax as integer constants in C++), and may be followed by a multiplier and an optional ‘B’ for "byte". The ‘s’ multiplier may be appended to any of the other multipliers. For example, ‘ks’ means kilosectors (1000 * sector_size), and ‘Kis’ means kibisectors (1024 * sector_size).

Table of SI and binary prefixes (unit multipliers):

Exit status: 0 for a normal exit, 1 for environmental problems (file not found, invalid command-line options, I/O errors, etc), 2 to indicate a corrupt or invalid input file, 3 for an internal consistency error (e.g., bug) which caused ddrescue to panic.

If ddrescue is interrupted by a signal, it updates mapfile and then terminates by raising the signal received.

7 Mapfile structure ¶

NOTE: In versions of ddrescue prior to 1.20 the mapfile was called ‘logfile’. The format is the same; only the name has changed.

The mapfile is a text file easy to read and edit. It is formed by three parts, the heading comments, the status line, and the list of data blocks. The character ‘#’ at begin of line or after whitespace starts a comment that extends to the end of the line.

The heading comments contain the version of ddrescue or ddrescuelog that created the mapfile, the command line used, and the time when the program started. If the mapfile was created by ddrescue it also contains the current time when the mapfile was saved and a copy of the status message from the screen describing the operation being performed (copying, trimming, finished, etc). They are intended as information for the user.

The first non-comment line is the status line. It contains a non-negative integer, a status character, and a positive decimal integer. The first integer is the position being tried in the input file. (The beginning of the block being tried in a forward pass or the end of the block in a backward pass). The status character is one of these:

Finally, the last integer is the number of the current pass in the current phase. The status line allows ddrescue to resume the copying phase instead of restarting it from pass 1. It also allows the retrying phase to resume in the same direction it was interrupted.

The blocks in the list of data blocks must be contiguous and non-overlapping.

Every line in the list of data blocks describes a block of data. It contains 2 non-negative integers and a status character. The first integer is the starting position of the block in the input file, the second integer is the size (in bytes) of the block. The status character is one of these:

And here is an example mapfile:

# Mapfile. Created by GNU ddrescue version 1.28
# Command line: ddrescue -d -c18 /dev/fd0 fdimage mapfile
# Start time:   2015-07-21 09:37:44
# Current time: 2015-07-21 09:38:19
# Copying non-tried blocks... Pass 1 (forwards)
# current_pos  current_status  current_pass
0x00120000     ?               1
#      pos        size  status
0x00000000  0x00117000  +
0x00117000  0x00000200  -
0x00117200  0x00001000  /
0x00118200  0x00007E00  *
0x00120000  0x00048000  ?

If you edit the file, you may use decimal, hexadecimal, or octal values, using the same syntax as integer constants in C++, except for current_pass, which must be a decimal integer.

8 Saving the mapfile in case of trouble ¶

The mapfile is an essential part of ddrescue’s effectiveness. Without a mapfile, ddrescue can’t resume a rescue, only reinitiate it. Given that a difficult rescue may take days to complete, it would be a serious drawback if the mapfile were lost because of a solvable problem like a lack of space on the device the mapfile is written to.

In case of trouble writing the mapfile, ddrescue prints a message like this:

Error writing mapfile 'mapfile': No space left on device
Fix the problem and press ENTER to retry,
                     or E+ENTER for an emergency save and exit,
                     or Q+ENTER to abort.

You may try to fix the problem, for example deleting some files to make room for the mapfile, and press Return to retry.

If the problem can’t be fixed, you may press e followed by Return to try an emergency save and exit. Ddrescue tries to write the mapfile to the file ddrescue.map in the current directory or, if this fails, to $HOME/ddrescue.map. If the mapfile is written successfully, ddrescue exits with status 1. Else it prints the above message again.

Or you may press q followed by Return to quit and exit with status 1. In this case the content of the mapfile is lost.

9 Copying CD-ROMs and DVDs ¶

Ddrescue may be better than dd for copying recordable CD-ROMs because the two lead out sectors at the end of some of them may cause a read error that prevents the whole last record from being copied by dd, potentially losing data. Also dd may create an image larger than the original if the ‘sync’ conversion and a block size larger than the sector size are specified.

In the special case of reading CD-ROMs (but not DVDs), the specialized tool dvdisaster may be a better option than ddrescue for recovering data because dvdisaster can read and analyze raw CD sectors, which ddrescue can’t.

Recordable CD and DVD media keep their data only for a finite time (typically for some years). After that time, data loss develops slowly with read errors growing from the outer region towards the inside. It is a good idea to make two (or more) copies of every important CD-ROM/DVD you burn so that you can later recover them with ddrescue.

If you have only one copy of a CD-ROM or DVD that fails when being copied, and if you have access to multiple optical media drives, you have a better chance of recovering the bad sectors since one drive may fail to read a particular sector, but another drive might be able to squeeze the data out of it, depending on the laser frequency and the sensitivity of the laser-sensor that reads the reflected laser light.

Example 1: Rescue a CD-ROM in /dev/cdrom.

ddrescue -n -b2048 /dev/cdrom cdimage mapfile
ddrescue -d -r1 -b2048 /dev/cdrom cdimage mapfile
  (if bad-sector size is zero, cdimage now contains a complete image
   of the CD-ROM and you can write it to a blank CD-ROM)

Example 2: Rescue a CD-ROM in /dev/cdrom from two copies.

ddrescue -n -b2048 /dev/cdrom cdimage mapfile
ddrescue -d -b2048 /dev/cdrom cdimage mapfile
  (insert second copy in the CD drive)
ddrescue -d -r1 -b2048 /dev/cdrom cdimage mapfile
  (if bad-sector size is zero, cdimage now contains a complete image
   of the CD-ROM and you can write it to a blank CD-ROM)

Example 3: Rescue a CD-ROM in /dev/cdrom using two CD drives from two different computers, writing the image into an USB drive mounted on /mnt/mem.

ddrescue -n -b2048 /dev/cdrom /mnt/mem/cdimage /mnt/mem/mapfile
ddrescue -d -r1 -b2048 /dev/cdrom /mnt/mem/cdimage /mnt/mem/mapfile
  (umount the USB drive and move both USB drive and CD-ROM to second
   computer)
ddrescue -d -r1 -b2048 /dev/cdrom /mnt/mem/cdimage /mnt/mem/mapfile
  (if bad-sector size is zero, /mnt/mem/cdimage now contains a complete
   image of the CD-ROM and you can write it to a blank CD-ROM)

Example 4: Merge the partially recovered images of 3 identical DVDs using their mapfiles as domain mapfiles.

ddrescue -m mapfile1 dvdimage1 dvdimage mapfile
ddrescue -m mapfile2 dvdimage2 dvdimage mapfile
ddrescue -m mapfile3 dvdimage3 dvdimage mapfile
  (if bad-sector size is zero, dvdimage now contains a complete image
   of the DVD and you can write it to a blank DVD)

Example 5: Rescue a lzip compressed backup from two copies on CD-ROM with error-checked merging of copies. See the lziprecover manual for details about lziprecover.

ddrescue -d -r1 -b2048 /dev/cdrom cdimage1 mapfile1
mount -t iso9660 -o loop,ro cdimage1 /mnt/cdimage
cp /mnt/cdimage/backup.tar.lz rescued1.tar.lz
umount /mnt/cdimage
  (insert second copy in the CD drive)
ddrescue -d -r1 -b2048 /dev/cdrom cdimage2 mapfile2
mount -t iso9660 -o loop,ro cdimage2 /mnt/cdimage
cp /mnt/cdimage/backup.tar.lz rescued2.tar.lz
umount /mnt/cdimage
lziprecover -m -v -o backup.tar.lz rescued1.tar.lz rescued2.tar.lz
  Input files merged successfully.
lziprecover -tv backup.tar.lz
  backup.tar.lz: ok

10 A small tutorial with examples ¶

This tutorial is for those already able to use the dd command. If you don’t know what dd is, better search the net for some introductory material about dd and GNU ddrescue first.

A failing drive tends to develop more and more errors as time passes. Because of this, you should rescue the data from a drive as soon as you notice the first error. Be diligent because every time a physically damaged drive powers up and is able to output some data, it may be the very last time that it ever will.

You should make a copy of the failing drive with ddrescue, and then try to repair the copy. If your data are really important, use the first copy as a master for a second copy, and try to repair the second copy. If something goes wrong, you have the master intact to try again.

If you are trying to rescue a whole partition, first repair the copy with e2fsck or some other tool appropriate for the type of partition you are trying to rescue, then mount the repaired copy somewhere and try to recover the files in it.

If the drive is so damaged that the file system in the rescued partition can’t be repaired or mounted, you will have to browse the rescued data with an hex editor and extract the desired parts by hand, or use a file recovery tool like photorec.

If the partition table is damaged, you may try to rescue the whole disc, then try to repair the partition table and the partitions on the copy.

If the damaged drive is not listed in /dev, then you cannot rescue it. At least not with ddrescue.

See Copying CD-ROMs and DVDs, for rescue examples of CD-ROMs and DVDs.

Example 1: Fully automatic rescue of a whole disc with two ext2 partitions in /dev/sda to /dev/sdb.
Note: you don’t need to partition /dev/sdb beforehand, but if the partition table on /dev/sda is damaged, you’ll need to recreate it somehow on /dev/sdb.

ddrescue -f -r3 /dev/sda /dev/sdb mapfile
fdisk /dev/sdb
e2fsck -v -f /dev/sdb1
e2fsck -v -f /dev/sdb2

Example 2: Rescue an ext2 partition in /dev/sda2 to /dev/sdb2.
Note: you need to create the partition sdb2 with fdisk first. sdb2 should be of appropriate type and size.

ddrescue -f -n /dev/sda2 /dev/sdb2 mapfile
ddrescue -d -f -r3 /dev/sda2 /dev/sdb2 mapfile
e2fsck -v -f /dev/sdb2
mount -t ext2 -o ro /dev/sdb2 /mnt
  (read rescued files from /mnt)

Example 3: While rescuing the whole drive /dev/sda to /dev/sdb, /dev/sda freezes up at position 12345678.

ddrescue -f /dev/sda /dev/sdb mapfile         # /dev/sda freezes here
  (restart /dev/sda or reboot computer)
  (restart copy at a safe distance from the troubled sector)
ddrescue -f -i 12350000 /dev/sda /dev/sdb mapfile
  (then copy backwards down to the troubled sector)
ddrescue -f -R /dev/sda /dev/sdb mapfile

Example 4: While rescuing the whole drive /dev/sda to /dev/sdb, /dev/sdb fails and you have to rescue the data to a third drive, /dev/sdc.

ddrescue -f -n /dev/sda /dev/sdb mapfile1       # /dev/sdb fails here
ddrescue -f -m mapfile1 /dev/sdb /dev/sdc mapfile2
ddrescue -f -n /dev/sda /dev/sdc mapfile2
ddrescue -d -f -r3 /dev/sda /dev/sdc mapfile2

Example 5: While rescuing a partition in /dev/sda1 to the file hdimage, /dev/sda1 stops responding and begins returning read errors, causing ddrescue to mark the rest of the partition as non-scraped.

ddrescue -n /dev/sda1 hdimage mapfile          # /dev/sda1 fails here
  (restart /dev/sda or reboot computer)
ddrescue -n -A -i<pos> -O /dev/sda1 hdimage mapfile
  (if /dev/sda1 fails again, restart /dev/sda or reboot computer and
   then repeat the above command as many times as needed until it
   succeeds. <pos> is the position where the drive stopped responding)
ddrescue -d -r3 /dev/sda1 hdimage mapfile

Example 6: While rescuing a partition in /dev/sda1 to the file hdimage, sda1 disappears from /dev.

ddrescue -n /dev/sda1 hdimage mapfile          # /dev/sda1 fails here
  (restart /dev/sda or reboot computer and then repeat the above
   command as many times as needed until it succeeds)
ddrescue -d -r3 /dev/sda1 hdimage mapfile

Example 7: While rescuing a partition in /dev/sda1 to the file hdimage, the partition table of /dev/sda becomes unreadable and the OS no longer shows sda1 in /dev. The solution is to shift the mapfile and read the rest of the partition sda1 from /dev/sda.
Note: you need to know the offset of the partition sda1 in the drive sda and the size of sda1.

ddrescue /dev/sda1 hdimage mapfile       # partition table fails here
ddrescuelog --shift -o<offset> mapfile > shifted_mapfile
ddrescue -i<offset> -o0 -s<size> /dev/sda hdimage shifted_mapfile

Example 8: After rescuing a partition in /dev/sda1 to the file hdimage, expand hdimage to copy the whole drive in /dev/sda without recopying the already copied partition /dev/sda1. The solution is to shift the mapfile, move the data of sda1 to its final position in hdimage, and then read the rest of the data from /dev/sda.
Note: you need to know the offset of the partition sda1 in the drive sda and the size of sda1.

ddrescue /dev/sda1 hdimage mapfile             # rescue partition
ddrescuelog --shift -o<offset> mapfile > shifted_mapfile
ddrescue --same-file -o<offset> -s<size> --reverse hdimage hdimage
ddrescue /dev/sda hdimage shifted_mapfile      # rescue rest of drive

11 Direct disc access ¶

If you notice that the positions and sizes in mapfile are always multiples of the sector size, maybe your kernel is caching the disc accesses and grouping them. In this case you may want to use direct disc access for infile, or read from a raw device, to bypass the kernel cache and rescue more of your data.

NOTE! Sector size must be correctly set with the option --sector-size for direct disc access to work.

NOTE: Direct disc access can copy arbitrary domains by reading whole sectors and then writing only the requested part. This is the only case where ddrescue tries to read data outside the rescue domain.

Try the option --idirect first. If direct disc access is not available in your system, try raw devices. Read your system documentation to find how to bind a raw device to a regular block device. Some OSs provide raw access through especial device names, like /dev/rdisk.

Ddrescue aligns its I/O buffer to the sector size so that it can be used for direct disc access or to read from raw devices. For efficiency reasons, also aligns it to the memory page size if page size is a multiple of sector size. On some systems, ddrescue can’t determine the size of a raw device, so an explicit --size or --complete-only option may be needed.

Using direct disc access, or reading from a raw device, may be slower or faster than normal cached reading depending on your OS and hardware. In case it is slower you may want to make a first pass using normal cached reads and use direct disc access, or a raw device, only to recover the good sectors inside the failed blocks.

Example 1: using direct disc access.

ddrescue -f -n /dev/sdb1 /dev/sdc1 mapfile
ddrescue -d -f -r3 /dev/sdb1 /dev/sdc1 mapfile
e2fsck -v -f /dev/sdc1
mount -t ext2 -o ro /dev/sdc1 /mnt

Example 2: using a raw device.

raw /dev/raw/raw1 /dev/sdb1
ddrescue -f -n /dev/sdb1 /dev/sdc1 mapfile
ddrescue -C -f -r3 /dev/raw/raw1 /dev/sdc1 mapfile
raw /dev/raw/raw1 0 0
e2fsck -v -f /dev/sdc1
mount -t ext2 -o ro /dev/sdc1 /mnt

12 Copying parts of the input file on demand ¶

The command mode of ddrescue implements a scripting interface similar to the one of ed. In this mode commands are read from the standard input and executed to copy parts of the input file to the output file, retrieve information from the mapfile, or write the mapfile to disc. Ddrescue’s responses are written to standard output. "done\n" for a successfully executed command, "error\n" for a failed command (for example because of wrong arguments), and "error: error message\n" for serious or fatal errors like write errors.

If end-of-file is detected on standard input, ddrescue discards any partial command being read and executes the ‘f’ (finish) command.

All ddrescue commands are single characters, though some require additonal parameters separated by spaces. Only one command is allowed per line. Ddrescue recognizes the following commands:

c pos size

Copy command. Copies a block of data from infile to outfile and updates the internal copy of the mapfile. The areas already marked as finished in the mapfile are not copied again.

f

Finish command. Compacts the internal copy of the mapfile and writes it to mapfile (if it was specified in the command line). Then prints "done\n" to standard output and quits. On startup, the mapfile is first compacted and then split following the rescue domain borders. The finish command compacts the mapfile again before exiting. If writing mapfile to disc fails, a non-interactive emergency save is tried before exiting. See Saving the mapfile in case of trouble.

q

Quit command. Exits ddrescue. Does not update the mapfile.

s pos size

Status command. Writes to standard output one or more lines in mapfile format (see Mapfile structure), showing the status of the areas included in the block requested. A line consisting of the string "done" marks the end of the list.

u

Update mapfile command. Writes the internal copy of the mapfile to mapfile (if it was specified in the command line). If update mapfile fails, you may try to fix the problem, for example deleting some files to make room for mapfile, before trying to update it again.

13 Fill mode ¶

When ddrescue is invoked with the option --fill-mode it operates in "fill mode", which is different from the default "rescue mode". That is, in "fill mode" ddrescue does not rescue anything. It only fills with data read from infile the blocks of outfile whose status character from mapfile coincides with one of the type characters specified in the argument to --fill-mode.

If the argument to --fill-mode contains an ‘l’, ddrescue writes location data (position, sector number, and status) into each sector filled. With bad sectors filled in this way, it should be possible to retry the recovery of important files, as the location of the error is known by looking into the unfinished copy of the file.

In fill mode infile does not need to be seekable and it may be of any size. If it is too small, the data is duplicated as many times as necessary to fill the input buffer. If it is too big, only the data needed to fill the input buffer is read. Then the same data is written to every cluster or sector to be filled.

Note that in fill mode infile is always read from position 0. If you specify a --input-position, it refers to the original infile from which mapfile was built, and is only used to calculate the offset between input and output positions.

Note also that when filling the infile of the original rescue run you should not set --output-position, whereas when filling the outfile of the original rescue run you should keep the original offset between --input-position and --output-position.

The option --fill-mode implies --complete-only.

In fill mode mapfile is updated to allow resumability when interrupted or in case of a crash, but as nothing is being rescued mapfile is not destroyed. The status line is the only part of mapfile that is modified.

The fill mode has a number of uses. See the following examples:

Example 1: Mark parts of the rescued copy to allow finding them when examined in an hex editor. For example, the following command line fills all blocks marked as ‘-’ (bad-sector) with copies of the string ‘BAD-SECTOR ’:

ddrescue --fill-mode=- <(printf "BAD-SECTOR ") outfile mapfile

And the following command line fills all the non-finished areas in the destination file with copies of the string ‘NON-RESCUED-SECTOR ’:

ddrescue --fill-mode='?*/-' <(printf "NON-RESCUED-SECTOR ") outfile mapfile

Example 2: Wipe only the good sectors, leaving the bad sectors alone. This way, the drive will still test bad (i.e., with unreadable sectors). This is the fastest way of wiping a failing drive, and is especially useful when sending the drive back to the manufacturer for warranty replacement.

ddrescue --fill-mode=+ --force /dev/zero bad_drive mapfile

Example 3: Force the drive to remap the bad sectors, making it usable again. If the drive has only a few bad sectors, and they are not caused by drive age, you can probably just rewrite those sectors, and the drive will reallocate them automatically to new "spare" sectors that it keeps for just this purpose. WARNING! This may not work on your drive.

ddrescue --fill-mode=- -f --synchronous /dev/zero bad_drive mapfile

Fill mode can also help you to figure out, independently of the file system used, what files are partially or entirely in the bad areas of the disc. Just follow these steps:

1) Copy the damaged drive with ddrescue until finished. Don’t use sparse writes. This yields a mapfile containing only finished (‘+’) and bad-sector (‘-’) blocks.

2) Fill the bad-sector blocks of the copied drive or image file with a string not present in any file, for example "DEADBEEF". Use --fill-mode=l- if you want location data.

3) Mount the copied drive (or the image file, via loopback device) read-only.

4) Grep for the fill string in all the files. Those files containing the string reside (at least partially) in damaged disc areas. Note that if all the damaged areas are in unused space, grep will not find the string in any file, which means that no files are damaged.

5) Take note of the location data of any important files that you want to retry.

6) Unmount the copied drive or image file.

7) Retry the sectors belonging to the important files until they are rescued or until it is clear that they can’t be rescued.

8) Optionally fill the bad-sector blocks of the copied drive or image file with zeros to restore the disc image.

Example 4: Figure out what files are in the bad areas of the disc.

ddrescue -b2048 /dev/cdrom cdimage mapfile
printf "DEADBEEF" > tmpfile
ddrescue --fill-mode=l- tmpfile cdimage mapfile
rm tmpfile
mount -t iso9660 -o loop,ro cdimage /mnt/cdimage
find /mnt/cdimage -type f -exec grep -l "DEADBEEF" '{}' ';'
  (note that my_thesis.txt has a bad sector at pos 0x12345000)
umount /mnt/cdimage
ddrescue -b2048 -i0x12345000 -s2048 -dr9 /dev/cdrom cdimage mapfile
ddrescue --fill-mode=- /dev/zero cdimage mapfile
mount -t iso9660 -o loop,ro cdimage /mnt/cdimage
cp -a /mnt/cdimage/my_thesis.txt /safe/place/my_thesis.txt

14 Generate mode ¶

When ddrescue is invoked with the option --generate-mode it operates in "generate mode", which is different from the default "rescue mode". That is, in "generate mode" ddrescue does not rescue anything. It only tries to generate a mapfile for later use.

So you didn’t read the manual and started ddrescue without a mapfile. Now, two days later, your computer crashed and you can’t know how much data ddrescue managed to save. And even worse, you can’t resume the rescue; you have to restart it from the very beginning.

Or maybe you started copying a drive with ‘dd conv=noerror,sync’ and are now in the same situation described above. In this case, note that you can’t use a copy made by dd unless it was invoked with the ‘sync’ conversion argument.

Don’t despair (yet). Ddrescue can in some cases generate an approximate mapfile, from infile and the (partial) copy in outfile, that is almost as good as an exact mapfile. It makes this by simply assuming that sectors containing all zeros were not rescued.

However, if the destination of the copy was a drive or a partition, (or an existing regular file and truncation was not requested), most probably you will need to restart ddrescue from the very beginning. (This time with a mapfile, of course). The reason is that old data may be present in the drive that have not been overwritten yet, and may be thus non-tried but non-zero.

For example, if you first tried one of these commands:

ddrescue infile outfile
or
dd if=infile of=outfile conv=noerror,sync

then you can generate an approximate mapfile with this command:

ddrescue --generate-mode infile outfile mapfile

Note that you must keep the original offset between --input-position and --output-position of the original rescue run.

15 Ddrescuelog ¶

Ddrescuelog is a tool that manipulates ddrescue mapfiles, shows mapfile contents, converts mapfiles to/from other formats, compares mapfiles, tests rescue status, and can delete a mapfile if the rescue is done. Ddrescuelog operations can be restricted to one or several parts of the mapfile if the domain setting options are used.

When performing logic operations (AND, OR, XOR) on mapfiles of different extension, only the blocks present in both files are processed. Other blocks are left untouched.

Here are some examples of how to use ddrescuelog, alone or in combination with other tools.

Example 1: Delete the mapfile if the rescue is finished (all data have been recovered without errors left).

ddrescue -f /dev/sda /dev/sdb mapfile
ddrescuelog -d mapfile

Example 2: Rescue two ext2 partitions in /dev/sda to /dev/sdb and repair the file systems using badblock lists generated with ddrescuelog. File system block size is 4096.
Note: you do need to partition /dev/sdb beforehand.

fdisk /dev/sdb                                   # partition /dev/sdb
ddrescue -f /dev/sda1 /dev/sdb1 mapfile1
ddrescue -f /dev/sda2 /dev/sdb2 mapfile2
ddrescuelog -l- -b4096 mapfile1 > badblocks1
ddrescuelog -l- -b4096 mapfile2 > badblocks2
e2fsck -v -f -L badblocks1 /dev/sdb1
e2fsck -v -f -L badblocks2 /dev/sdb2

Example 3: Rescue a whole disc with two ext2 partitions in /dev/sda to /dev/sdb and repair the file systems using badblock lists generated with ddrescuelog. Disc sector size is 512, file system block size is 4096. Arguments to options -i and -s are the starting positions and sizes of the partitions being rescued.
Note: you don’t need to partition /dev/sdb beforehand, but if the partition table on /dev/sda is damaged, you’ll need to recreate it somehow on /dev/sdb.

ddrescue -f /dev/sda /dev/sdb mapfile
fdisk /dev/sdb                                  # get partition sizes
ddrescuelog -l- -b512 -i63s -o0 -s767457s -b4096 mapfile > badblocks1
ddrescuelog -l- -b512 -i767520s -o0 -s96520s -b4096 mapfile > badblocks2
e2fsck -v -f -L badblocks1 /dev/sdb1
e2fsck -v -f -L badblocks2 /dev/sdb2

16 Invoking ddrescuelog ¶

The format for running ddrescuelog is:

ddrescuelog [options] mapfile

Use a hyphen ‘-’ as mapfile to read the mapfile from standard input (also in the options taking a mapfile argument) or to write the mapfile created by --create-mapfile to standard output.

Ddrescuelog supports the following options:

-h

--help

Print an informative help message describing the options and exit.

-V

--version

Print the version number of ddrescuelog on the standard output and exit. This version number should be included in all bug reports.

-a old_types,new_types

--change-types=old_types,new_types

Change the status of every block in the rescue domain from one type in old_types to the corresponding type in new_types, much like the command ‘tr’ does, and write the resulting mapfile to standard output. old_types and new_types are strings of block status characters as defined in the chapter Mapfile structure (see Mapfile structure). Blocks whose status is not in old_types are left unchanged. If new_types is shorter than old_types the last type of new_types is repeated as many times as necessary.

-A

--annotate-mapfile

Add comments containing the human-readable positions and sizes of the blocks in mapfile which are included in the rescue domain, and write the resulting mapfile to standard output.

-b bytes

--block-size=bytes

--sector-size=bytes

Block size used by ddrescuelog. Depending on the requested operation it may be the sector size of the input device, the block size of the rescued file system, etc. Defaults to 512.

-B

--binary-prefixes

Show units with binary prefixes (powers of 1024).
SI prefixes (powers of 1000) are used by default. (See table above, Invoking ddrescue).

-c[type1type2]

--create-mapfile[=type1type2]

Create a mapfile from a list of sectors read from standard input. The option --format determines the format of the input: either a list of sector numbers, or a bitmap. The sector numbers may be unordered. Only sectors included in the rescue domain are added to mapfile. If the mapfile is being created displaced from the input domain, the offset between -i and -o must be a multiple of the sector size. In this case, remember to specify -o because it defaults to the same value given to -i, producing a default offset of 0.

type1 and type2 are block status characters as defined in the chapter Mapfile structure (see Mapfile structure). type1 sets the type for blocks included in the list, while type2 sets the type for the rest of mapfile. If not specified, type1 defaults to ‘+’ and type2 defaults to ‘-’.

-C[type]

--complete-mapfile[=type]

Complete a synthetic (user fabricated) mapfile by filling the gaps with blocks of type type, and write the completed mapfile to standard output. type is one of the block status characters defined in the chapter Mapfile structure (see Mapfile structure). If type is not specified, the gaps are filled with non-tried blocks. All gaps in mapfile are filled. Domain options are ignored.

-d

--delete-if-done

Delete the given mapfile if all the blocks in the rescue domain have been successfully recovered. The exit status is 0 if mapfile could be deleted, 1 otherwise. If the mapfile is read from standard input, behave like --done-status. (There is nothing to delete).

-D

--done-status

Test if all the blocks in the rescue domain have been successfully recovered. The exit status is 0 if all tested blocks are finished, 1 otherwise.

-f

--force

Force overwrite of mapfile.

-F name

--format=name

Select the input format for --create-mapfile, or the output format for --list-blocks. The valid names are ‘list’, ‘bitmap’, ‘bitmap-be’ (big endian), and ‘bitmap-le’ (little endian). Plain ‘bitmap’ is equivalent to ‘bitmap-le’. The default format is ‘list’ (a list of block numbers). In a big endian bitmap the first block is represented by the most significant bit of the first byte. In a little endian bitmap the first block is represented by the least significant bit of the first byte.

-i bytes

--input-position=bytes

Starting position of the rescue domain, in bytes. Defaults to 0. It refers to a position in the original infile.

-l types

--list-blocks=types

By default print on standard output the sector numbers of the blocks specified as any of types in mapfile and included in the rescue domain. The list format is one sector number per line in decimal, like the output of the program ‘badblocks’, so that it can be used as input for e2fsck or other similar filesystem repairing tool.

If a bitmap --format is specified, write to standard output a bitmap of the endianness chosen, with ones for the sectors specified and zeros for sectors from other block types. The bits set to 1 in the bitmap should be equivalent to the list of blocks.

types contains one or more of the block status characters defined in the chapter Mapfile structure (see Mapfile structure). If the output numbers or bits are being created displaced from their position in the input domain, the offset between -i and -o must be a multiple of the block size. In this case, remember to specify -o because it defaults to the same value given to -i, producing a default offset of 0.

A sector is listed (or set to 1 in bitmap output) if selected, even partially. (The positions and sizes of the mapfile blocks are not required to be multiples of the sector size).

-L

--loose-domain

Accept an incomplete synthetic (user fabricated) domain mapfile or compare-as-domain mapfile, and fill the gaps in the list of data blocks with non-tried blocks. The blocks in the mapfile may be unordered, may overlap other blocks of the same status, and don’t need to be contiguous. This option allows making quick edits to a mapfile without all the size calculations involved in making all data blocks contiguous again.

-m file

--domain-mapfile=file

Restrict the rescue domain to the blocks marked as finished in the mapfile file.

-n

--invert-mapfile

Invert the types of the blocks in mapfile which are included in the rescue domain, and write the resulting mapfile to standard output. Finished blocks (‘+’) are changed to bad-sector (‘-’), all other types are changed to finished. --invert-mapfile is equivalent to --change-types=?*/-+,++++-

-o bytes

--output-position=bytes

Starting position of the image of the rescue domain in the original outfile, in bytes. It is used by the options --create-mapfile, --list-blocks, and --shift. Defaults to --input-position.

-p file

--compare-mapfile=file

Compare the types of the blocks included in the rescue domain. The exit status is 0 if all the blocks tested are the same in both file and mapfile, 1 otherwise.

-P file

--compare-as-domain=file

Compare only the blocks marked as finished in the rescue domain. The exit status is 0 if all the blocks tested are the same in both file and mapfile, 1 otherwise. Two files comparing equal with this option are equivalent when used as domain mapfiles.

-q

--quiet

Quiet operation. Suppress all messages.

-s bytes

--size=bytes

Maximum size of the rescue domain in bytes. It refers to a size in the original infile. -1 removes any previous size limit.

-t

--show-status

Print a summary of the contents of each mapfile to the standard output. This option allows more than one mapfile. If the domain setting options are used, the summary can be restricted to one or several parts of mapfile.

-v

--verbose

Verbose mode. Further -v’s (up to 4) increase the verbosity level.

-x file

--xor-mapfile=file

Perform a logical XOR (exclusive OR) operation between the finished blocks in file and those in mapfile, and write the resulting mapfile to standard output. In other words, in the resulting mapfile a block is only shown as finished if it was finished in either of the two input mapfiles but not in both.

-y file

--and-mapfile=file

Perform a logical AND operation between the finished blocks in file and those in mapfile, and write the resulting mapfile to standard output. In other words, in the resulting mapfile a block is only shown as finished if it was finished in both input mapfiles.

-z file

--or-mapfile=file

Perform a logical OR operation between the finished blocks in file and those in mapfile, and write the resulting mapfile to standard output. In other words, in the resulting mapfile a block is shown as finished if it was finished in either of the two input mapfiles.

--shift

Shift the positions of all the blocks in mapfile by the offset (--output-position - --input-position), and write the resulting mapfile to standard output. Either --input-position or --output-position must be 0. Any blocks beyond the end of the rescue domain are removed before performing the shift. The remaining blocks are shifted even if they are outside the rescue domain. If the offset is positive, a non-tried block is inserted before the first block to fill the gap.

Exit status: 0 for a normal exit, 1 for environmental problems (file not found, invalid command-line options, I/O errors, etc), 2 to indicate a corrupt or invalid input file, 3 for an internal consistency error (e.g., bug) which caused ddrescuelog to panic.

17 Reporting bugs ¶

There are probably bugs in ddrescue. There are certainly errors and omissions in this manual. If you report them, they will get fixed. If you don’t, no one will ever know about them and they will remain unfixed for all eternity, if not longer.

If you find a bug in GNU ddrescue, please send electronic mail to bug-ddrescue@gnu.org. Include the version number, which you can find by running ‘ddrescue --version’.