1SAFECOPY(1) SAFECOPY SAFECOPY(1)
2
6 safecopy - rescue data from a source that causes IO errors
7
10 safecopy [OPTION]... SOURCE DEST
11
14 safecopy tries to get as much data from SOURCE as possible, even
15 resorting to device specific low level operations if applicable.
16 This is achieved by identifying problematic or damaged areas, skipping
18 over them and continuing reading afterwards. The corresponding area in
19 the destination file is either skipped (on initial creation that means
20 padded with zeros) or deliberately filled with a recognizable pattern
21 to later find affected files on a corrupted device.
22 Safecopy uses an incremental algorithm to identify the exact beginning
24 and end of bad areas, allowing the user to trade minimum accesses to
25 bad areas for thorough data resurrection.
26 Multiple passes over the same file are possible, to first retrieve as
28 much data from a device as possible with minimum harm, and then trying
29 to retrieve some of the remaining data with increasingly aggressive
30 read attempts.
31 For this to work, the source device or file has to be seekable. For
33 unseekable devices (like tapes) you can try to use an external script
34 to execute a controlled skip over the damaged part for you.
35 (For example by using "mt seek" and "mt tell" on an SCSI tape device)
37 See the "-S <seekscript>" parameter for details.
38 Performance and success of this tool depend extremely on the device
40 driver, firmware and underlying hardware.
41 Currently safecopy supports RAW access to CDROM drives to read data
43 directly of a CD, bypassing some driver dependant error correction.
44 This can speed up data retrieval from CDs and reduce system load during
45 recovery, as well as increase the success rate. Safecopy uses the disc
46 status syscall to determine sector size and addressing of CDs. This
47 fails on mixed-mode or multi-session CDs, since the sector layout can
48 change within the disk, but would still work on the the big majority of
49 disks. Other disks can still be recovered using normal high level data
50 access. Safecopy auto-detects the disk type involved during scan for
51 disk and block size.
52 Some CD/DVD drives are known to cause the ATAPI bus to crash on errors,
54 causing the device driver to freeze for times up to and beyond a minute
55 per error. Try to avoid using such drives for media recovery. Using
56 safecopys low level access features might help under some circum‐
57 stances.
58 Some drives can read bad media better than others. Be sure to attempt
60 data recovery of CDs and DVDs on several different drives and comput‐
61 ers. You can use safecopys incremental recovery feature to read previ‐
62 ously unreadable sectors only.
63
66 Data recovery from damaged media is a delicate task, in the worst case
67 its success or failure can safe or ruin whole companies and seal
68 affected peoples personal fate. It is paramount that any tools written
69 for that purpose are reliable and trustworthy.
70 A user needs to know what exactly the software is doing to his hardware
72 and data. The outcome of any operation needs to be both understandable
73 and predictable.
74 An "intelligent data resurrection wizard" with unknown complex internal
76 behaviour may be a nifty tool, but does not meet the above requirements
77 of predictable outcome, nor will the user know in advance what is done
78 to his data.
79 The operation sequence of safecopy has been kept relatively simple to
81 assure this predictability. Unfortunately feature additions have risen
82 the complexity and lead to undefined outcome in the past when include
83 and exclude lists had been mixed, especially when mixing different
84 block sizes. In the worst case this could have lead to overwritten data
85 in the destination file on a later incremental run with the mark (-M)
86 option.
87 From version 1.3 on, safecopy ships with a test suite that can be used
89 to verify safecopys behaviour in a set of test cases, simulating the
90 combination of bad blocks in input with different include and exclude
91 lists, both with and without marking. Releases are only made if safe‐
92 copy passes those test cases according to the specification.
93 This textual specification of behaviour of safecopy can be found in the
95 file specification.txt shipped with safecopy.
96
99 --stage1
100 Preset to rescue most of the data fast, using no retries and
101 avoiding bad areas.
102 Presets: -f 10% -r 10% -R 1 -Z 0 -L 2 -M BaDbLoCk \
104 -o stage1.badblocks
105 --stage2
107 Preset to rescue more data, using no retries but searching for
108 exact ends of bad areas.
109 Presets: -f 128* -r 1* -R 1 -Z 0 -L 2 \
111 -I stage1.badblocks -o stage2.badblocks
112 --stage3
114 Preset to rescue everything that can be rescued using maximum
115 retries, head realignment tricks and low level access.
116 Presets: -f 1* -r 1* -R 4 -Z 1 -L 2 \
118 -I stage2.badblocks -o stage3.badblocks
119 All stage presets can be overridden by individual options.
121 -b <size>
123 Blocksize for default read operations. Set this to the physical
124 sectorsize of your media.
125 Default: 1*
127 Hardware block size if reported by OS, otherwise 4096
128 -f <size>
130 Blocksize in bytes when skipping over badblocks. Higher set‐
131 tings put less strain on your hardware, but you might miss good
132 areas in between two bad ones.
133 Default: 16*
135 -r <size>
137 Resolution in bytes when searching for the exact beginning or
138 end of a bad area. If you read data directly from a device
139 there is no need to set this lower than the hardware blocksize.
140 On mounted filesystems however, read blocks and physical blocks
141 could be misaligned. Smaller values lead to very thorough
142 attempts to read data at the edge of damaged areas, but increase
143 the strain on the damaged media.
144 Default: 1*
146 -R <number>
148 At least that many read attempts are made on the first bad block
149 of a damaged area with minimum resolution. More retries can
150 sometimes recover a weak sector, but at the cost of additional
151 strain.
152 Default: 3
154 -Z <number>
156 On each error, force seek the read head from start to end of the
157 source device as often as specified. That takes time, creates
158 additional strain and might not be supported by all devices or
159 drivers.
160 Default: 1
162 -L <mode>
164 Use low level device calls as specified:
165 0 Do not use low level device calls
167 1 Attempt low level device calls for error recovery only
168 2 Always use low level device calls if available
169 Supported low level features in this version are:
171 SYSTEM DEVICE TYPE FEATURE
173 Linux cdrom/dvd bus/device reset
174 Linux cdrom read sector in raw mode
175 Linux floppy controller reset, twaddle
176 Default: 1
178 --sync Use synchronized read calls (disable driver buffering).
180 Default: Asynchronous read buffering by the OS is allowed
182 -s <blocks>
184 Start position where to start reading. Will correspond to posi‐
185 tion 0 in the destination file.
186 Default: block 0
188 -l <blocks>
190 Maximum length of data to be read.
191 Default: Entire size of input file
193 -I <badblockfile>
195 Incremental mode. Assume the target file already exists and has
196 holes specified in the badblockfile. It will be attempted to
197 retrieve more data from the listed blocks or from beyond the
198 file size of the target file only.
199 Warning: Without this option, the destination file will be emp‐
201 tied prior to writing. Use -I /dev/null if you want to continue
202 a previous run of safecopy without a badblock list.
203 Default: none
205 -i <bytes>
207 Blocksize to interpret the badblockfile given with -I.
208 Default: Blocksize as specified by -b
210 -X <badblockfile>
212 Exclusion mode. If used together with -I, excluded blocks over‐
213 ride included blocks. Safecopy will not read or write any data
214 from areas covered by exclude blocks.
215 Default: none
217 -x <bytes>
219 Blocksize to interpret the badblockfile given with -X.
220 Default: Blocksize as specified by -b
222 -o <badblockfile>
224 Write a badblocks/e2fsck compatible bad block file.
225 Default: none
227 -S <seekscript>
229 Use external script for seeking in input file. (Might be useful
230 for tape devices and similar). Seekscript must be an executable
231 that takes the number of blocks to be skipped as argv1 (1-64)
232 the blocksize in bytes as argv2 and the current position (in
233 bytes) as argv3. Return value needs to be the number of blocks
234 successfully skipped, or 0 to indicate seek failure. The exter‐
235 nal seekscript will only be used if lseek() fails and we need to
236 skip over data.
237 Default: none
239 -M <string>
241 Mark unrecovered data with this string instead of skipping it.
242 This helps in later finding corrupted files on rescued file sys‐
243 tem images. The default is to zero unreadable data on creation
244 of output files, and leaving the data as it is on any later run.
245 Warning: When used in combination with incremental mode (-I)
247 this may overwrite data in any block that occurs in the -I file.
248 Blocks not in the -I file, or covered by the file specified with
249 -X are save from being overwritten.
250 Default: none
252 -h, --help
254 Show the program help text.
255
258 valid parameters for -f -r -b <size> options are:
259 <integer>
261 Amount in bytes - i.e. 1024
262 <percentage>%
264 Percentage of whole file/device size - e.g. 10%
265 <number>*
267 -b only, number times blocksize reported by OS
268 <number>*
270 -f and -r only, number times the value of -b
271
274 description of output symbols:
275 . Between 1 and 1024 blocks successfully read.
277 _ Read of block was incomplete. (possibly end of file) The block‐
279 size is now reduced to read the rest.
280 |/| Seek failed, source can only be read sequentially.
282 > Read failed, reducing blocksize to read partial data.
284 ! A low level error on read attempt of smallest allowed size leads
286 to a retry attempt.
287 [xx](+yy){
289 Current block and number of bytes continuously read successfully
290 up to this point.
291 X Read failed on a block with minimum blocksize and is skipped.
293 Unrecoverable error, destination file is padded with zeros.
294 Data is now skipped until end of the unreadable area is reached.
295 < Successful read after the end of a bad area causes backtracking
297 with smaller blocksizes to search for the first readable data.
298 }[xx](+yy)
300 current block and number of bytes of recent continuous unread‐
301 able data.
302
305 How do I...
306 - resurrect a file from a mounted but damaged media, that copy will
308 fail on:
309 safecopy /path/to/problemfile ~/saved-file
310 - create an filesystem image of a damaged disk/cdrom:
312 safecopy /dev/device ~/diskimage
313 - resurrect data as thoroughly as possible?
315 safecopy source dest -f 1* -R 8 -Z 2
317 (assuming logical misalignment of blocks to sectors)
318 safecopy source dest -f 1* -r 1 -R 8 -Z 2
320 - resurrect data as fast as possible, or
322 - resurrect data with low risk of damaging the media further:
324 (you can use even higher values for -f and -r)
325 safecopy source dest -f 10% -R 0 -Z 0
327 - resurrect some data fast, then read more data thoroughly later:
329 safecopy source dest -f 10% -R 0 -Z 0 -o badblockfile
331 safecopy source dest -f 1* -R 8 -Z 2 -I badblockfile
332 Alternate approach using the new preset features:
334 safecopy source dest --stage1
336 safecopy source dest --stage2
338 safecopy source dest --stage3
340 - utilize some friends CD-ROM drives to complete the data from my dam‐
342 aged CD:
343 safecopy /dev/mydrive imagefile <someoptions> -b <myblocksize>
344 -o myblockfile
345 safecopy /dev/otherdrive imagefile <someoptions> -b <otherblock‐
346 size> -I myblockfile -i <myblocksize> -o otherblockfile
347 safecopy /dev/anotherdrive imagefile <someoptions> -b <anoth‐
348 erblocksize> -I otherblockfile -i <otherblocksize>
349 - interrupt and later resume a data rescue operation:
351 safecopy source dest
352 <CTRL+C> (safecopy aborts)
353 safecopy source dest -I /dev/null
354 - interrupt and later resume a data rescue operation with correct bad‐
356 blocks output:
357 safecopy source dest <options> -o badblockfile
358 <CTRL+C> (safecopy aborts)
359 mv badblockfile savedbadblockfile
360 safecopy source dest -I /dev/null -o badblockfile
361 cat badblockfile >>savedbadblockfile
362 - interrupt and resume in incremental mode:
364 (this needs a bit of bash scripting to get the correct badblock
365 lists)
366 safecopy source dest <options> -o badblockfile1
367 safecopy source dest <options> -I badblockfile1 -o badblockfile2
368 <CTRL+C> (safecopy aborts)
369 latest=$( tail -n 1 badblockfile2 )
370 if [ -z $latest ]; then latest=-1; fi;
371 cat badblockfile1 | while read block; do
372 [ $block -gt $latest ] && echo $block >>badblockfile2;
373 done;
374 safecopy source dest <options> -I badblockfile2 -o badblockfile3
375 - find the corrupted files on a partially successful rescued file sys‐
377 tem:
378 safecopy /dev/filesystem image -M CoRrUpTeD
379 fsck image
380 mount -o loop image /mnt/mountpoint
381 grep -R /mnt/mountpoint "CoRrUpTeD"
382 (hint: this might not find all affected files if the unreadable
383 parts are smaller in size than your marker string)
384 - exclude the previously known badblocks list of a filesystem from
386 filesystem image creation:
387 dumpe2fs -b /dev/filesystem >badblocklist
388 safecopy /dev/filesystem image -X badblocklist -x <blocksize of
389 your fs>
390 - create an image of a device that starts at X and is Y in size:
392 safecopy /dev/filesystem -b <bsize> -s <X/bsize> -l <Y/bsize>
393 - combine two partial images of rescued data without access to the
395 actual (damaged) source data:
396 (This is a bit tricky. You need to get badblocks lists for both
397 files somehow to make safecopy know where the missing data is.
398 If you used the -M (mark) feature you might be able to automati‐
399 cally compute these, however this feature is not provided by
400 safecopy. Lets assume you have two badblocks files.
401 you have:
403 image1.dat
404 image1.badblocks (blocksize1)
405 image2.dat
406 image2.badblocks (blocksize2)
407 The file size of image1 needs to be greater or equal to that of
409 image2. (If not, swap them) )
410 cp image2.dat combined.dat
412 safecopy image1.dat combined.dat -I image2.badblocks -i block‐
413 size2 -X image1.badblocks -x blocksize1
414 (This gets you the combined data, but no output badblocklist.
415 The resulting badblocks list would be the badblocks that are
416 a: in both badblocks lists, or
417 b: in image1.badblocks and beyond the file size of image2 It
418 should be reasonably easy to solve this logic in a short shell
419 script. One day this might be shipped with safecopy, until then
420 consider this your chance to contribute to a random open source
421 project.)
422 - rescue data of a tape device:
424 If the tape device driver supports lseek(), treat it as any
425 file, otherwise utilize the "-S" option of safecopy with a to be
426 self-written script to skip over the bad blocks. (for example
427 using "mt seek") Make sure your tape device doesn't auto-rewind
428 on close. Send me feedback if you had any luck doing so, so I
429 can update this documentation.
430
433 Q: Why create this tool if there already is something like dd-res‐
434 cue and other tools for that purpose?
435 A: Because I didn't know of dd(-)rescue when I started, and I felt
437 like it. Also I think safecopy suits the needs of a user in data
438 loss peril better due to more readable output and more under‐
439 standable options than some of the other tools. (Then again I
440 am biased. Compare them yourself) Meanwhile safecopy supports
441 low level features other tools don't.
442 Q: What exactly does the -Z option do?
444 A: Remember back in MS-DOS times when a floppy would make a "neek
446 nark" sound 3 times every time when running into a read error?
447 This happened when the BIOS or DOS disk driver moved the IO head
448 to its boundaries to possibly correct small cylinder misalign‐
449 ment, before it tried again. Linux doesn't do that by default,
450 neither do common CDROM drives or drivers. Nevertheless forcing
451 this behaviour can increase your chance of reading bad sectors
452 from a CD __BIG__ time. (Unlike floppies where it usually has
453 little effect)
454 Q: Whats my best chance to resurrect a CD that has become unread‐
456 able?
457 A: Try making a backup image on many different computers and
459 drives. The abilities to read from bad media vary extremely. I
460 have a 6 year old Lite On CDRW drive, that even reads deeply and
461 purposely scratched CDs (as in with my key, to make it unread‐
462 able) flawlessly. A CDRW drive of the same age at work doesn't
463 read any data from that part of the CD at all, while most DVD
464 and combo drives have bad blocks every couple hundred bytes.
465 Make full use of safecopys RAW access features if applicable.
466 (-L 2 option)
467 As a general guideline: -CDRW drives usually do better than
469 read-only CD drives.
470 -CD only drives sometimes do better on CDs than DVD drives.
471 -PC drives are sometimes better than laptop ones.
472 -A drive with a clean lens does better than a dirtball.
473 -Cleaning up CDs helps.
474 -Unless you use chemicals.
475 -Using sticky tape on a CD will rip of the reflective layer per‐
476 manently rendering the disk unreadable.
477 Q: Whats my best chance to resurrect a floppy that became unread‐
479 able?
480 A: Again try different floppy drives. Keep in mind that it might be
482 easier to further damage data on a bad floppy than on a CD.
483 (Don't overdo read attempts)
484 Q: What about BlueRay/HDDVD disks?
486 A: Hell if I knew, but generally they should be similar to DVDs.
488 It probably depends how the drives firmware acts up.
489 Q: My hard drive suddenly has many bad sectors, what should I do?
491 A: Speed is an essential factor when rescuing data from a bad hard
493 drive. Accesses to bad areas or even just time running can fur‐
494 ther damage the drive and make formerly readable areas unread‐
495 able, be it due to temperature rise, damaged heads scratching
496 still good parts of the surface, bearings degrading due to
497 vibration, etc. Its advisable to shut the system down and
498 remove the hard drive from the computer as soon as errors occur
499 and as fast as possible without causing further damage.
500 (Don't pull the plug! Press reset to force reboot and then power
502 down via power button/ACPI)
503 Set up a rescue system with enough disk space to store all the
505 data from the damaged drive (and possibly multiple copies of
506 it). If you have an external hard drive case that connects via
507 USB SATA or SCSI, allowing a hot plug of the drive into a run‐
508 ning system, use it. This allows you to prepare everything
509 without the need for the drive to power up and possibly BIOS or
510 operating system involuntarily accessing it. You also get eas‐
511 ier access to the drive to check temperature or noise during
512 operation.
513 When you rescue data, rescue good data first before attempting
515 to access bad sectors. Safecopy allows you to skip known problem
516 sectors using a badblock exclude file (-X) which you might be
517 able to retrieve from system logs or from the drive internal
518 logs, via smartmontools or similar software. Be aware that you
519 might possibly need to convert physical sector numbers into log‐
520 ical block numbers depending on your source.
521 Also you should tell safecopy to jump out of any problematic
523 areas ASAP and continue somewhere else. Parameters "-f 10% -r
524 10% -R 0 -Z 0" would do the trick by making safecopy skip 10% of
525 the device content and continue there without backtracking. You
526 can always attempt to get the data in between later, first get
527 the supposedly good data on the rest of the drive. Depending on
528 the method of data recovery you plan on using, it may make sense
529 to mark the bad data with the "-M" option. This allows you to
530 later find files affected by data corruption more easily. Use
531 the "-o" option to make safecopy write a badblock list with all
532 blocks skipped or unreadable.
533 When safecopy is done with this first run, you can attempt a
535 second go trying to get more data. Using smaller values for "-f"
536 and allowing safecopy to backtrack for the end of the affected
537 area "-r 1*". Be sure to use incremental mode "-I" to only read
538 the blocks skipped in the first run.
539 It may make sense to let the drive cool down between runs. Once
541 you got all the data from the "good" areas of the drive you can
542 risk more "thorough" access to it. Increase the numbers of
543 retries "-R" in case of bad blocks, maybe together with a head
544 realignment "-Z 1". "-f 1*" would make safecopy try to read on
545 every single block, not skipping over bad areas at all.
546 If your drive stops responding, power it down, let it cool down
548 for a while, then try again.
549 (I heard from people who brought dead drives back to live for a
551 short time by cooling them to low temperatures with ice-packs.)
552 !!! If the data is really important, go to a professional data
554 recovery specialist right away, before doing further damage to
555 the drive.
556
559 safecopy always returns 0 (zero) regardless of success rate and/or
560 aborted execution.
561
564 safecopy and its manpage have been designed and written by CorvusCorax.
565
568 Please use the project page on sourceforge
569 <http://www.sf.net/projects/safecopy> to get in contact with project
570 development if you encounter bugs or want to contribute to safecopy.
571
574 Copyright © 2009 CorvusCorax
575 This is free software. You may redistribute copies of it under the
576 terms of the GNU General Public License
577 <http://www.gnu.org/licenses/gpl.html>. There is NO WARRANTY, to the
578 extent permitted by law.
579
582 Programs with a similar scope are among others
583 ddrescue(1), dd-rescue(1), ...
584 2009-05-27 SAFECOPY(1)