[Techtalk] Disk recovery?

[Rudy L. Zijlstra]

Maria Blackmore wrote:

>On Mon, 28 Jul 2003, Rudy L. Zijlstra wrote:
>
>  
>
>>I've a 60G disk from a friend that he could'nt read anymore. Linux 
>>basically tells me that lba sectors 0 up to and including 6 are 
>>unreadable. In other words, the most important sectors have gone to 
>>bitheaven.
>>    
>>
>
>can you define "unreadable" please? what is the error message?
>
OK, here comes:

Jul 28 03:00:04 barak kernel: hdd: WDC WD600BB-75CCB0, ATA DISK drive
Jul 28 03:00:04 barak kernel: ide0 at 0x1f0-0x1f7,0x3f6 on irq 14
Jul 28 03:00:04 barak kernel: ide1 at 0x170-0x177,0x376 on irq 15
Jul 28 03:00:04 barak kernel: blk: queue c0374a84, I/O limit 4095Mb (mask 0xffffffff)
Jul 28 03:00:04 barak kernel: blk: queue c0374bd0, I/O limit 4095Mb (mask 0xffffffff)
Jul 28 03:00:04 barak kernel: blk: queue c0374f34, I/O limit 4095Mb (mask 0xffffffff)
Jul 28 03:00:04 barak kernel: spurious 8259A interrupt: IRQ7.
Jul 28 03:00:05 barak kernel: hdd: dma_intr: error=0x10 { SectorIdNotFound }, LBAsect=0, sector=0
Jul 28 03:00:05 barak kernel: hdd: dma_intr: status=0x51 { DriveReady SeekComplete Error }
Jul 28 03:00:05 barak kernel: hdd: dma_intr: error=0x10 { SectorIdNotFound }, LBAsect=0, sector=0
Jul 28 03:00:05 barak kernel: hdd: dma_intr: status=0x51 { DriveReady SeekComplete Error }
Jul 28 03:00:05 barak kernel: hdd: dma_intr: error=0x10 { SectorIdNotFound }, LBAsect=0, sector=0
Jul 28 03:00:05 barak kernel: hdd: dma_intr: status=0x51 { DriveReady SeekComplete Error }
Jul 28 03:00:05 barak kernel: hdd: dma_intr: error=0x10 { SectorIdNotFound }, LBAsect=0, sector=0
Jul 28 03:00:05 barak kernel: hdd: DMA disabled
Jul 28 03:00:05 barak kernel: ide1: reset: master: error (0x0a?)
Jul 28 03:00:05 barak kernel: hdd: read_intr: status=0x59 { DriveReady SeekComplete DataRequest Error }
Jul 28 03:00:05 barak kernel: hdd: read_intr: error=0x10 { SectorIdNotFound }, LBAsect=0, sector=0
Jul 28 03:00:05 barak kernel: hdd: read_intr: status=0x59 { DriveReady SeekComplete DataRequest Error }
Jul 28 03:00:05 barak kernel: hdd: read_intr: error=0x10 { SectorIdNotFound }, LBAsect=0, sector=0
Jul 28 03:00:06 barak kernel: hdd: read_intr: status=0x59 { DriveReady SeekComplete DataRequest Error }
Jul 28 03:00:06 barak kernel: hdd: read_intr: error=0x10 { SectorIdNotFound }, LBAsect=0, sector=0
Jul 28 03:00:06 barak kernel: hdd: read_intr: status=0x59 { DriveReady SeekComplete DataRequest Error }
Jul 28 03:00:06 barak kernel: hdd: read_intr: error=0x10 { SectorIdNotFound }, LBAsect=0, sector=0
Jul 28 03:00:06 barak kernel: ide1: reset: master: error (0x0a?)
Jul 28 03:00:06 barak kernel: end_request: I/O error, dev 16:40 (hdd), sector 0
Jul 28 03:00:06 barak kernel: end_request: I/O error, dev 16:40 (hdd), sector 2
Jul 28 03:00:06 barak kernel: end_request: I/O error, dev 16:40 (hdd), sector 4
Jul 28 03:00:06 barak kernel: end_request: I/O error, dev 16:40 (hdd), sector 6
Jul 28 03:00:06 barak kernel: end_request: I/O error, dev 16:40 (hdd), sector 0
Jul 28 03:00:06 barak kernel: end_request: I/O error, dev 16:40 (hdd), sector 2
Jul 28 03:00:06 barak kernel: end_request: I/O error, dev 16:40 (hdd), sector 4
Jul 28 03:00:06 barak kernel: end_request: I/O error, dev 16:40 (hdd), sector 6



Note that system time is (luckily) too late by a few hours... ;-) 

[snip]

>It's probably still there, but I believe your friend has no learned a very
>important lesson, which is the vital importance of keeping backups.
>
I just hope you are right, he's a bit - how to say this - stubborn at 
times?

>>(1) Any suggestions on how to get data from that disk with the following 
>>provisos:
>>      - its circuitry is at least partially to HW heaven as well, i can 
>>only get it working as slave, not as master (not recognised on the bus 
>>with no other device present)
>>    
>>
>
>This does not bode well at all, but does the drive identify it's
>make/model correctly when being run like this?
>  
>
See above
With no jumpers in it seemed to kill the other IDE channel as well. Its 
a SIS chipset in that system (only used for testing and this kind of 
work), with jumper on master it was not visible and jumpered as slave it 
worked at least as far as this. /proc/ide contained all other information.

>It may be the case that you cannot trust the data coming off the drive.
>  
>
Alas, yup, its why i started asking opinions.

>  
>
>>      - So it may well be that the platter is OK, but the circuitry is 
>>not, resulting in the found problems
>>    
>>
>
>This is a distinct possibility, yes.  You may find milage in obtaining a
>hard drive of the exact same model, and swapping the drive electronics
>from the good drive to the bad drive.
>
I have demolished some known bad drives, so this should be possible. 
Will ask friend in question about this. His call to pay for the other drive.

[snip]

>>(2) Any suggestions for a european data recovery center / company?
>>    
>>
>
>I would have suggested Vogon, until they recently tried to defraud the
>serious fraud office here in the UK.  Strangely they were taken to court,
>and lost.
>
>http://theregister.co.uk/content/63/31870.html
>
Not very trustworthy, no

>
>What I would suggest is you approach a few firms you find in the vicinity,
>and ask them for quotes.  Data recovery can approach quite reasonable
>rates when you do not have time as a limiting factor, so be sure to
>mention this when you ask them.  The easiest way to find such companies
>would be through the application of google, I would suggest.
>
Was hoping for experience points. Tells way more than websites do..


Thanks,

Rudy