LinuxQuestions.org

LinuxQuestions.org (/questions/)
-   Linux - Hardware (https://www.linuxquestions.org/questions/linux-hardware-18/)
-   -   need help for recover NTFS partition! (https://www.linuxquestions.org/questions/linux-hardware-18/need-help-for-recover-ntfs-partition-499942/)

taiwf 11-08-2006 11:58 PM

need help for recover NTFS partition!
 
hi,

i have am IBM 60G hard disk (7200RMP) which quite often come out with 'chee chee chee chee' sound while use.
Last week i was playing game all of sudden my game become a very slow sliding-show like and i reset the computer and only realize that my hd is dead.

I have tried spinrite for a week but the number of hours to complete jump from 1000 to 4xxx. So i turn it off and try to use ddrescue but it still only to copy little bit of the full disk

===PS: the knoppix still see 2 partition of the first NTFS hard disk(IBM) and single partition of 2nd hard disk. I presume this mean partition table still ok. But i get input/outpt error while trying to use knoppix to read first hard disk. I can read about half of 2nd hard disk from knoppix though.

[first attempt]
root@1[hdb5]# ddrescue /dev/hda /mnt/samba/backup.img


Press Ctrl-C to interrupt
rescued: 25165 kB, errsize: 0 B, current rate: 8456 B/s
rescued: 25165 kB, errsize: 0 B, current rate: 0 B/s
ipos: 25165 kB, errors: 0, average rate: 17026 B/s
opos: 25165 kBr: Input/output error

[2nd attempt]
root@1[samba]# ddrescue /dev/hda /mnt/samba/backup.img


Press Ctrl-C to interrupt
rescued: 12255 kB, errsize: 0 B, current rate: 131 kB/s
rescued: 12255 kB, errsize: 0 B, current rate: 0 B/s
ipos: 12255 kB, errors: 0, average rate: 17995 B/s
opos: 12255 kBr: Input/output error

syg00 11-09-2006 05:40 PM

If you have a real disc failure, rather than just a read error, I suspect you are out of luck. People have reported success with things such as putting the disk (wrapped) in a freezer overnight. I've tried it on a couple of dead disks, but it didn't work for me.
I recently tried all I knew to recover an NTFS disk that had been eaten by a power spike. Got nowhere - the business data was backed up, but lots of (family) stuff was lost when we had to re-format the whole lot away.

Brian1 11-09-2006 05:43 PM

If default Window tools cannot recover then one last resort is contact one of those disk recovery sites to see the cost to recover what data can be recovered. Quite expensive but can recover a lot but some of the data is not really complete files. more like bit and pieces. It can also depend on the drive. The places that do this type of recovery usually are able to remove the platters in the drive and place in a good working drive as far as harddrive controller board interface. It works best if you swap platters out of the exact same brand, model, revicsion, and size.

Brian

tamoneya 11-09-2006 10:10 PM

I have used the freezer method and have had marginal success with it. Just barely enough to extract some important documents. I try leaving it in the freezer for a couple hours in a plastic bag so no moisture gets in. This causes some of the parts to contract and allows it to move more freely. Then use a livecd distro like backtrack(just google for it). Boot your system and the ntfs should be mounted and you can recover the data. If this doesnt work there really isnt much you can do except some data recovery center. It depends how valuable the data is i guess.

taiwf 11-09-2006 11:58 PM

I am not sure if the hard disk is really dead coz it still spin at least... one thing i dont' understand is whether winxp crash will cause the other hard disk crash? I ask so coz i have winxp swap file on the 2nd hard disk. When crash happen, both my hard disk gone!

The 2 hard disk are different brand ,segate and IBM. I think the chance they fail at the same time is pretty remote.


Netherless, i will probably trying the frezzing metho as last resort. I found that if i use /dev/hda1 instead of /dev/hda , i can get more data out into the img. file

But i am not so sure is there anyway i can mount the img? i dont' have spare hard drive the moment. Do i must get a new hard disk to read whatever extracted into img file(from gddrescue)?

syg00 11-10-2006 01:51 AM

Quote:

Originally Posted by taiwf
I am not sure if the hard disk is really dead coz it still spin at least... one thing i dont' understand is whether winxp crash will cause the other hard disk crash? I ask so coz i have winxp swap file on the 2nd hard disk. When crash happen, both my hard disk gone!

Yes, I wondered about that too.
Maybe you have a failure of the (E)IDE (presumably on-board) controller.
If you have the opportunity, swap the disks into another machine. Maybe buy a cheap PCI controller card. If you are lucky the disks are o.k.
As for recovering the image, the various recovery liveCDs should allow you to mount a samba share.

archtoad6 11-16-2006 12:20 PM

(I assume you were trying SpinRite 6.)
The reason SpinRite was taking so long is that is will try to read each problematic sector up to 2,000 times (yes, each sector, 2k times). Furthermore, time-to-complete algorithms being what they are, the estimated time will skyrocket in the middle of a bad section & then come back down in the good parts.

How far (%) did SpinRite get before you gave up?
Did you keep any of the SpinRite logs?

Have you tried smartctl (from smartmontools)?
Code:

smartctl -a <device>
will show everything S.M.A.R.T reports about a drive, while
Code:

smartctl -il error <device>
will show basic info plus the error log.

taiwf 11-19-2006 04:21 PM

spinrite after 1 week running is only 2% and i saw it jump from 1% to 2% on the 2nd day running. Since then, the hours keep increasing and progress is really slow on the DOS GUI. Thats why i decide to turn off because i heard running hard disk on that recovery mode over months will simply kill the harddisk.

I've been busy haven't tried swap hard disk to diff computer yet. Will probably do it tonight and try the smartctl thing too. I will post the result back here.

Electro 11-19-2006 07:11 PM

On my almost dead hard drive. I used manufacture's disk utility and cross my fingers while hitting enter to start fixing sectors. After one run, it booted Linux further, so repeat fixing the sectors. After the second time, Linux finished booted. I was lucky all files were not effected except for one that I did not care for.

The hard drive that is almost dead is an Hitachi 60GXP 40 GB. It still works, but it is not as efficient at finding files like it was before.

If two IDE hard drives are on one channel and one hard drive fails, the second hard drive will fail. Some hard drives are resistant to this problem. While others are not and this goes true to IDE controllers too. Though it depends on the signal state at the time of the hard drive failed.

I recommend that nobody tries the freezer trick. The mechanics components do not like this. Also freezer bags eventually make foods get frost bite. The cold temperatures (68 degrees F) is the victim that made my hard drive's lifespan shorter.

taiwf 11-20-2006 04:19 AM

just now change the 2 hard disk to my working celeron machine (2.5GHz) and boot up using knoppix but still no luck. when trying to access hard disk i got :

Could not mount device
the reported error was
Error reading $Mft record(s): input/output error
Failed to open inode: input/output error

couldn't mount device '/dev/hdb1': input/output error

Mount Failed.

================`smartctl -a <device>`=============
root@1[knoppix]# smartctl -a /dev/hdb
smartctl version 5.36 [i686-pc-linux-gnu] Copyright (C) 2002-6 Bruce Allen
Home page is http://smartmontools.sourceforge.net/

=== START OF INFORMATION SECTION ===
Model Family: IBM Deskstar 60GXP series
Device Model: IC35L060AVER07-0
Serial Number: SZPTZ1C6563
Firmware Version: ER6OA46A
User Capacity: 61,492,838,400 bytes
Device is: In smartctl database [for details use: -P show]
ATA Version is: 5
ATA Standard is: ATA/ATAPI-5 T13 1321D revision 1
Local Time is: Mon Nov 20 21:19:59 2006 EST

==> WARNING: IBM Deskstar 60GXP drives may need upgraded SMART firmware.
Please see http://www.geocities.com/dtla_update/index.html#rel and
http://www-3.ibm.com/pc/support/site...cid=MIGR-42215 or
http://www-1.ibm.com/support/docview...psg1MIGR-42215

SMART support is: Available - device has SMART capability.
SMART support is: Disabled

SMART Disabled. Use option -s with argument 'on' to enable it.
root@1[knoppix]# smartctl -a /dev/hda
smartctl version 5.36 [i686-pc-linux-gnu] Copyright (C) 2002-6 Bruce Allen
Home page is http://smartmontools.sourceforge.net/

=== START OF INFORMATION SECTION ===
Model Family: IBM/Hitachi Deskstar GXP-180 family
Device Model: IC35L060AVV207-0
Serial Number: VNVB30G8THN8VH
Firmware Version: V22OA66A
User Capacity: 61,492,838,400 bytes
Device is: In smartctl database [for details use: -P show]
ATA Version is: 6
ATA Standard is: ATA/ATAPI-6 T13 1410D revision 3a
Local Time is: Mon Nov 20 21:20:19 2006 EST
SMART support is: Available - device has SMART capability.
SMART support is: Disabled

SMART Disabled. Use option -s with argument 'on' to enable it.


============='smartctl -il error <device>'===========
root@1[knoppix]# smartctl -il error /dev/hdb
smartctl version 5.36 [i686-pc-linux-gnu] Copyright (C) 2002-6 Bruce Allen
Home page is http://smartmontools.sourceforge.net/

=== START OF INFORMATION SECTION ===
Model Family: IBM Deskstar 60GXP series
Device Model: IC35L060AVER07-0
Serial Number: SZPTZ1C6563
Firmware Version: ER6OA46A
User Capacity: 61,492,838,400 bytes
Device is: In smartctl database [for details use: -P show]
ATA Version is: 5
ATA Standard is: ATA/ATAPI-5 T13 1321D revision 1
Local Time is: Mon Nov 20 21:21:53 2006 EST

==> WARNING: IBM Deskstar 60GXP drives may need upgraded SMART firmware.
Please see http://www.geocities.com/dtla_update/index.html#rel and
http://www-3.ibm.com/pc/support/site...cid=MIGR-42215 or
http://www-1.ibm.com/support/docview...psg1MIGR-42215

SMART support is: Available - device has SMART capability.
SMART support is: Disabled

SMART Disabled. Use option -s with argument 'on' to enable it.
root@1[knoppix]# smartctl -il error /dev/hda
smartctl version 5.36 [i686-pc-linux-gnu] Copyright (C) 2002-6 Bruce Allen
Home page is http://smartmontools.sourceforge.net/

=== START OF INFORMATION SECTION ===
Model Family: IBM/Hitachi Deskstar GXP-180 family
Device Model: IC35L060AVV207-0
Serial Number: VNVB30G8THN8VH
Firmware Version: V22OA66A
User Capacity: 61,492,838,400 bytes
Device is: In smartctl database [for details use: -P show]
ATA Version is: 6
ATA Standard is: ATA/ATAPI-6 T13 1410D revision 3a
Local Time is: Mon Nov 20 21:24:09 2006 EST
SMART support is: Available - device has SMART capability.
SMART support is: Disabled

SMART Disabled. Use option -s with argument 'on' to enable it.


###################################

*** I remember reading some info spinrite mentioned. not sure if it actually turn of the 'SMART'. It also mentioned sth about the firmware. Should i go ahead to update/upgrade firmware?

sparkbot 11-20-2006 04:47 AM

I also have had good luck with manuf. floppy repair disks from Maxtor, Seagate and IBM( I think you can still find a floppy image of an " IBM Harddisk tool" if there isn't one on the Hitachi site--I now keep all 4 floppys I downloaded). They did fix bad sectors when i thought things were gone. Once I got the data I used the floppy to do a low level format of the drive and then a Smart test and it is still working fine and SMART says it is ok(6yrs old Maxtor fixed 8mo. ago with the maxblast floopy on the latest hard drive hiccup I had --smart showed at boot--couldn't read/ write to drive without errors)
I'd try the manuf repair disk before you do too much more with it.

archtoad6 11-20-2006 10:03 AM

1. I assume you want to recover your data, so don't do any re-formatting yet. :)

2. The reason you aren't getting good results from smartctl is in the "SMART Disabled. Use option -s with argument 'on' to enable it." error msg. you keep getting. Add -s on or even -son to your smartctl options & see what happens. I think you only need to do this once. You may need to check the BIOS to be sure S.M.A.R.T Is enabled there.

3. As you can see, the drive doesn't have to be mounted to run smartctl. In fact there is a captive mode, -C, for unmounted drives that runs faster.

4. I recommend the man page, it's long but thorough.

BTW, which ver. of SpinRite were you using?

taiwf 11-20-2006 11:30 PM

version 6 from 'X'mule ....

poweredbydodge 11-21-2006 02:27 AM

dead hard disks will often still spin... i'm not going to get into the nuts and bolts of it, but there are more than one reason a drive will physically fail. physical failure simply means that it is not a file system error, not an operating system error, not an external controller error, not a cable error, but - rather - something / anything physically within the confines of the drive's casing. this includes the control board, which is #1 on my shit-list of things to take a dive over the years.

anyway - if its physically shot, and you want the data off that bad, then expect to pay between 800 and 3,000 dollars (US funds) for a clean room job.

if its software shot, then you can use something that scans sector by sector, pulling files off one by one. the bad part is that all your files will be named "file1.*" "file2.*" file3.* -- etc etc, rather than haveing their actual filenames or directories... there will actually be no directories created, just a shitload of files dumped out -- and you have to have another hard drive of equal size available also.

such procedures take roughly 2 hours for about 200 gigs on a 2 GHz machine with ATA-100 drives that spin 7200 rpm. how do i know? --- www.spinellicreations.com (notice the catastropic data failure thingy at the top :) )

taiwf 11-21-2006 04:20 PM

Quote:

Originally Posted by sparkbot
I also have had good luck with manuf. floppy repair disks from Maxtor, Seagate and IBM( I think you can still find a floppy image of an " IBM Harddisk tool" if there isn't one on the Hitachi site--I now keep all 4 floppys I downloaded). They did fix bad sectors when i thought things were gone. Once I got the data I used the floppy to do a low level format of the drive and then a Smart test and it is still working fine and SMART says it is ok(6yrs old Maxtor fixed 8mo. ago with the maxblast floopy on the latest hard drive hiccup I had --smart showed at boot--couldn't read/ write to drive without errors)
I'd try the manuf repair disk before you do too much more with it.

is this the one you mean:
http://storage.free-driver-download....00-XP-DOS.html

But it seems to be more a diagnostic tool the recovery one. Can you give me more detail about the floppy image?

besides, i will post the smart-enabled result later. I didn't enable it the first place coz i dono if that will further complicated the recovery process.


All times are GMT -5. The time now is 08:03 AM.