[onlit4regs]

hi,

My 500Gb Seagate 7200.11 had the famous bug of BSY state, it wasn't seen in BIOS.
I have successfully made the unbrick trick with serial cable, hyper terminal and so on. Now the drive is seen in the BIOS (and with the right size).
But Windows see a RAW partition on it.
I've tried easus recovery, it has seen all my files, but any restore of any file results in an unreadable file. (nothing was writen on the faulty disk at the moment)
I've tried testdisk, it has seen the partition, I have made "WRITE PARTITION TABLE" to the faulty disk, and reboot. same thing under windows.

so now, what can I do ?

I attach the HDHACKER first sector of logical drive and first sector of physical drive, if this can help.

thanks a lot for your help

Attached File(s)

•  HDHACKER.zip (1.07K)
  Number of downloads: 5

This post has been edited by onlit4regs: 17 August 2012 - 07:19 AM

[jaclaz]

onlit4regs, on 17 August 2012 - 07:16 AM, said:

I attach the HDHACKER first sector of logical drive and first sector of physical drive, if this can help.

At first sight there is nothing "wrong" in them.
BUT how exactly (under which OS, with which tools) was the disk partitioned originally?
The sectors you posted seems like a "normal" XP partitioned (French), single NTFS:

#0 07 80 0 1 1 1023 254 63   63 976768002 

The data in the bootsector is as well "standard":

 3      0003   OEM String:    NTFS  
11     000B    Bytes per sector:     0200   512
21     0015    Media type:             F8   248
24     0018    Sectors per Head:     003F   63
26     001A    Number of Heads:      00FF   255
28     001C    Sectors Before:   0000003F   63
40     0028    Total Sectors:                000000003A384C01   976768001
48     0030    LCN for $MFT::                00000000000C0000   786432
56     0038    LCN for $MFTMirr::                0000000003A384C0   61048000
64     0040    Clusters per $MFT record:            000000F6   246
68     0044    Clusters per Index record:            00000001   1
72     0048    Volume Serial:    6424B80A24B7DD6C

Which OS are you running/can run?

Ideally you should try imaging the disk on a (larger) device as a file, first thing, do you have (or can buy) a 640 or 750 Gb disk?

You could check with a disk viewer/editor the presence at the designed addresses of the $MFT and of it's mirror.
$MFT is at 63+786432*8=sector 6,291,519
$MFT Mirror is at 63+61048000*8=sector 488,384,063
The first sector of both should begin with "File0" or in hex "46494C4530".

jaclaz

[onlit4regs]

jaclaz, on 17 August 2012 - 07:51 AM, said:

onlit4regs, on 17 August 2012 - 07:16 AM, said:

I attach the HDHACKER first sector of logical drive and first sector of physical drive, if this can help.

At first sight there is nothing "wrong" in them.
BUT how exactly (under which OS, with which tools) was the disk partitioned originally?
The sectors you posted seems like a "normal" XP partitioned (French), single NTFS:

#0 07 80 0 1 1 1023 254 63   63 976768002 

The data in the bootsector is as well "standard":

 3      0003   OEM String:    NTFS  
11     000B    Bytes per sector:     0200   512
21     0015    Media type:             F8   248
24     0018    Sectors per Head:     003F   63
26     001A    Number of Heads:      00FF   255
28     001C    Sectors Before:   0000003F   63
40     0028    Total Sectors:                000000003A384C01   976768001
48     0030    LCN for $MFT::                00000000000C0000   786432
56     0038    LCN for $MFTMirr::                0000000003A384C0   61048000
64     0040    Clusters per $MFT record:            000000F6   246
68     0044    Clusters per Index record:            00000001   1
72     0048    Volume Serial:    6424B80A24B7DD6C

Which OS are you running/can run?

Ideally you should try imaging the disk on a (larger) device as a file, first thing, do you have (or can buy) a 640 or 750 Gb disk?

You could check with a disk viewer/editor the presence at the designed addresses of the $MFT and of it's mirror.
$MFT is at 63+786432*8=sector 6,291,519
$MFT Mirror is at 63+61048000*8=sector 488,384,063
The first sector of both should begin with "File0" or in hex "46494C4530".

jaclaz

thanks a lot for your message jaclaz.

it was originally parted in windows XP (French), with standard XP storage manager.
I have a 1Tb hard drive available for imaging, and I am running actually Win 7 Pro x64, but I can also plug it again on my win XP.
I'll try to give a look at the MFT with disk editor and I'll tell you.

[onlit4regs]

hmmm, $MFT begins with the good code "File0". the problem is $MFT mirror, impossible to access :
"System Error: Code 1117 " and in french something like: I/O error, unable to satisfy query.

[jaclaz]

onlit4regs, on 17 August 2012 - 08:02 AM, said:

it was originally parted in windows XP (French), with standard XP storage manager.

Then the partition and the bootsector are seemingly OK.

onlit4regs, on 17 August 2012 - 08:02 AM, said:

I have a 1Tb hard drive available for imaging, and I am running actually Win 7 Pro x64, but I can also plug it again on my win XP.

Then a good idea would be to image it.
The reference app is Datarescuedd, see here:
http://reboot.pro/7783/

You might want to do a few tests with "smallish" parts of the disk , see this for a possible approach:
http://reboot.pro/15040/#entry133567

I have no idea if it works ok under 7 64 bit, it should, but cannot say.
If you have a XP available it should be "safer" (in the sense of "known to be working")

onlit4regs, on 17 August 2012 - 08:26 AM, said:

hmmm, $MFT begins with the good code "File0". the problem is $MFT mirror, impossible to access :
"System Error: Code 1117 " and in french something like: I/O error, unable to satisfy query.

Hmmm, strange.
It is possible that there is a bunch of bad sectors (or a translation table in the disk that was cleared during the unbricking) but a failed $MFT Mirror should not prevent the filesystem to be recognized .

Once you have the image done, we will see what TESTDISK finds about those....

jaclaz

[onlit4regs]

hello jaclaz
ok the image is done, it took a long time, a lot of I/O errors. the image is only 136Go
what should I do next, you were talking about something with testdisk, can you tell me more please ?

thanks a lot

[jaclaz]

onlit4regs, on 22 August 2012 - 09:06 AM, said:

hello jaclaz
ok the image is done, it took a long time, a lot of I/O errors. the image is only 136Go
what should I do next, you were talking about something with testdisk, can you tell me more please ?

thanks a lot

Sorry, I missed your reply.

Anyway if the "image" is 136 Gb it is very UNLIKE an image.

You seemingly did not follow the proposed approach:

jaclaz, on 17 August 2012 - 11:44 AM, said:

You might want to do a few tests with "smallish" parts of the disk , see this for a possible approach:
http://reboot.pro/15040/#entry133567

I seem to remember to have read *somewhere* that there is a rather common issue (though I don't seem to remember affecting 7200.11 specifically) where after an unbricking the actually accessible data is about 1/3 of the total, maybe this is the case.

Now the whole point is:

Quote

does the "whatever" you have now, 136 Gb in size represent the first 136 Gb of the disk?

If yes, most probably you can recover partially the data present in that part of the disk.

What I would do:

• create a sparse file (on a NTFS partition) sized 63+976768002=976,768,128 sectors x 512 = 500,105,281,536 bytes in size
  
• "dd to it" the 136 Gb *whatever* you have
  
• analyze it with TESTDISK

How to do that in practice (in a command prompt window, after having collected the tools and put them in a directory like C:\hdtools\, and navigating to that directory):

• mksparse <path>\my500GB.img 500105281536
  
• dsfi <path>\my500GB.img 0 0 <path>\thewhatever136GB.img
  
• testdisk <path>\my500GB.img

mksparse: see here:
http://reboot.pro/31...__25#entry70583
or:
http://wayback.archi...se/mksparse.zip

dsfi (part of the dsfok toolkit), here (you can use instead any other "dd-like" tool yu may be more familiar with):
http://members.ozema...eezip/freeware/

testdisk, here:
http://www.cgsecurit...stDisk_Download

Follow this EXACTLY (you want to create a log, Intel, analyse, N to "search for Vista created partitions", ):
http://www.cgsecurit...sk_Step_By_Step
see if pressing "P" you see (at least some of) the files .

Report what happens.

jaclaz

[onlit4regs]

hi,
thanks for your reply and your time !

testdisk have seen the NTFS partition of 500Go, said structure OK.
when pressing "P", there is only one directory displayed, and when entering it, it's empty ....

I think I'll try again the image processing of the hard drive this week-end, I've done it with USB connexion, but I'll try with direct SATA connexion, to see if it can go up to 500gb image. I'll tell you what.
thanks again

[jaclaz]

onlit4regs, on 31 August 2012 - 06:32 AM, said:

testdisk have seen the NTFS partition of 500Go, said structure OK.
when pressing "P", there is only one directory displayed, and when entering it, it's empty ....

Try having a look at the "my500GB.img" with dmde:
http://softdm.com/
even if it is a tool that is not ( like TESTDISK) suitable to be used with a less then advanced knowledge of the NTFS filesystem, you should be able to understand if there is an issue with the $MFT or with the actual filesystem contents.
Another thing that you could do is to extrract some sectors starting from 6,291,519 and use on them this tool:
http://www.forensicf...ewtopic/t=8010/
http://code.google.com/p/mft2csv/
just to understand if the $MFT contains valid data or if it is "the issue".
If this latter is the case, PHOTOREC may still be able to find many files....


jaclaz

[onlit4regs]

jaclaz, on 31 August 2012 - 08:07 AM, said:

onlit4regs, on 31 August 2012 - 06:32 AM, said:

testdisk have seen the NTFS partition of 500Go, said structure OK.
when pressing "P", there is only one directory displayed, and when entering it, it's empty ....

Try having a look at the "my500GB.img" with dmde:
http://softdm.com/
even if it is a tool that is not ( like TESTDISK) suitable to be used with a less then advanced knowledge of the NTFS filesystem, you should be able to understand if there is an issue with the $MFT or with the actual filesystem contents.
Another thing that you could do is to extrract some sectors starting from 6,291,519 and use on them this tool:
http://www.forensicf...ewtopic/t=8010/
http://code.google.com/p/mft2csv/
just to understand if the $MFT contains valid data or if it is "the issue".
If this latter is the case, PHOTOREC may still be able to find many files....


jaclaz

softdm shows me all the files and directories of my hard drive, but trying to recover a dozen of files results with unreadable files.

I tried making a new image with ddrescue, with direct SATA connection in my PC. it started faster than the previous one on USB, but it is still running, since 72 hours !!!
only 24,5Go done .... so much errors: "unable to satisfy request because of an I/O error".
I have stopped the process.

about the last thing you asked, I'm not sure to understand what you really want me to do. Extract how many sectors from 6291519 ? and then, mft decode ? or mft2csv ?

thanks a lot for your help. I'm getting less and less hope to recover anything

[onlit4regs]

jaclaz, on 31 August 2012 - 08:07 AM, said:

onlit4regs, on 31 August 2012 - 06:32 AM, said:

testdisk have seen the NTFS partition of 500Go, said structure OK.
when pressing "P", there is only one directory displayed, and when entering it, it's empty ....

Try having a look at the "my500GB.img" with dmde:
http://softdm.com/
even if it is a tool that is not ( like TESTDISK) suitable to be used with a less then advanced knowledge of the NTFS filesystem, you should be able to understand if there is an issue with the $MFT or with the actual filesystem contents.
Another thing that you could do is to extrract some sectors starting from 6,291,519 and use on them this tool:
http://www.forensicf...ewtopic/t=8010/
http://code.google.com/p/mft2csv/
just to understand if the $MFT contains valid data or if it is "the issue".
If this latter is the case, PHOTOREC may still be able to find many files....


jaclaz

softdm shows me all the files and directories of my hard drive, but trying to recover a dozen of files results with unreadable files.

I tried making a new image with ddrescue, with direct SATA connection in my PC. it started faster than the previous one on USB, but it is still running, since 72 hours !!!
only 24,5Go done .... so much errors: "unable to satisfy request because of an I/O error".
I have stopped the process.

about the last thing you asked, I'm not sure to understand what you really want me to do. Extract how many sectors from 6291519 ? and then, mft decode ? or mft2csv ?

thanks a lot for your help. I'm getting less and less hope to recover anything

[jaclaz]

onlit4regs, on 03 September 2012 - 12:58 PM, said:

softdm shows me all the files and directories of my hard drive, but trying to recover a dozen of files results with unreadable files.

This should mean that the $MFT is OK (i.e. no need to analyze it manually with mft2csv).
But if you ran the DMDE on the (incomplete) image, this may still be "normal".

onlit4regs, on 03 September 2012 - 12:58 PM, said:

I tried making a new image with ddrescue, with direct SATA connection in my PC. it started faster than the previous one on USB, but it is still running, since 72 hours !!!
only 24,5Go done .... so much errors: "unable to satisfy request because of an I/O error".
I have stopped the process.

And, AGAIN, you are using a WRONG approach (attempting to image the whole disk at once).

For the THIRD time, please read again this:
http://www.msfn.org/...ost__p__1007535

jaclaz, on 17 August 2012 - 11:44 AM, said:

You might want to do a few tests with "smallish" parts of the disk , see this for a possible approach:
http://reboot.pro/15040/#entry133567

You might want to try with even smaller "chunks".

Another test (but be careful):
What happens with DMDE on the original disk?
I find it strange that the $MFT is "perfect" (as it seemingly is) but *all* the disk is unreadable (I could understand some areas, but not the large majority of the disk)

jaclaz

[onlit4regs]

ok, will try that sorry for not understanding this approach

I've just read the post indicated, I just have a question about reassembling the different parts ? how to do that ? it says to create an empty file (with which tool ?) of the size of the disk and then merge all parts.

thanks

[jaclaz]

onlit4regs, on 04 September 2012 - 08:47 AM, said:

ok, will try that sorry for not understanding this approach

I've just read the post indicated, I just have a question about reassembling the different parts ? how to do that ? it says to create an empty file (with which tool ?) of the size of the disk and then merge all parts.

thanks

Normally I would use fsz.exe (part of the dsfok toolkit):
http://members.ozema...eezip/freeware/
that is if you have the space available to create a "whole" file, but It should be more convenient to use instead mksparse, here:
http://reboot.pro/31...__25#entry70583
this way the file will grow only with the actual "chunks" that you write to it.
And the reference app to write these chunks is dsfi (still part of the dsfok toolkit).
I can give you specific instructions if this other approachs actual succeeds in getting more data.

But if the test with DMDE on the original disk gives the same results, than there is a more serious issue somewhere "before" (partial unbricking or "failed" unbricking) though as said it sounds "strange".

jaclaz

[onlit4regs]

hi jaclaz

so, I have tried DMDE on the original hard drive, It couldn't display the directory/file structure , it was so long on "reading MFT", more than 4 days to complete only 3% !! so I aborted

on this disk, there is a dozen of "most wanted" files for me, which may represents 2 or 3go. I've made my recovery tests on these files. maybe other are readable, but they are not necessary for the moment.

so, do you think I should try to image the disk in smaller chunks ?

thanks

[jaclaz]

onlit4regs, on 13 September 2012 - 05:45 AM, said:

so, I have tried DMDE on the original hard drive, It couldn't display the directory/file structure , it was so long on "reading MFT", more than 4 days to complete only 3% !! so I aborted

I am confused by your reports:

onlit4regs, on 03 September 2012 - 01:32 PM, said:

softdm shows me all the files and directories of my hard drive, but trying to recover a dozen of files results with unreadable files.

Were you previously using dmde to look at the $MFT of the "image" you made?

If yes, that image - even if partial - actually holds a seemingly valid $MFT.

You can try the following:

• open the (partial) image with dmde
  
• check in it's directory structure the names of the fiiles you most value
  
• try using this tool getFileExtents
  http://www.wd-3.com/...e/luserland.htm
  to understand the location of the files (one by one) and/or of the various fragments of them
  
• try imaging just the relevant sectors found above with ddrescue from the original disk to new files
  
• re-assemble the new files into the "original" ones

If this strategy works, you will be able to get just the "strictly needed" files in a smaller time.

jaclaz

[onlit4regs]

I can't get "getfileextents" to work
should I use it on my hard drive or on my image ?
how to tell it to search on the drive or image ? parameter seems to be only the filename

thanks

[jaclaz]

onlit4regs, on 14 September 2012 - 12:37 AM, said:

I can't get "getfileextents" to work
should I use it on my hard drive or on my image ?
how to tell it to search on the drive or image ? parameter seems to be only the filename

thanks

On the image (I mean if the dmde can access the $MFT entries of the image, as I seem to understand, asked you but had no specific answer about).
The image must be mounted to a drive letter (using a virtual disk driver) in order to let getfilextents work. <- Sorry I omitted this piece of info.

Possibly the most easy to use one (and "good enough" for the task) could be IMDISK:
http://www.ltr-data.se/opencode.html/
http://reboot.pro/forum/59/

See if it can mount the image to a drive letter.

jaclaz

[onlit4regs]

yes dmde has no problem seeing the directory/file structure on the image file, I see all my favorite files.

I've mounted it with IMDisk, with default parameters of size of virtual disk, etc. It showed a new letter, but impossible to browse this letter ! (no filesystem type indicated in IMDisk, and windows can't see the size of partition, file or directory unreable or corrupted ...)
so, can't get fileextents to work on it too.

??
thanks

[jaclaz]

onlit4regs, on 14 September 2012 - 05:16 AM, said:

yes dmde has no problem seeing the directory/file structure on the image file, I see all my favorite files.

Good.

onlit4regs, on 14 September 2012 - 05:16 AM, said:

I've mounted it with IMDisk, with default parameters of size of virtual disk, etc. It showed a new letter, but impossible to browse this letter ! (no filesystem type indicated in IMDisk, and windows can't see the size of partition, file or directory unreable or corrupted ...)
so, can't get fileextents to work on it too.

Wait a minute, are you still using the "resulting" image as ddrescue created it, or did you "grow" it to it's full size (creating a new file with mksparse and dd-ing to it the image)?


jaclaz