Sorry for the long post, but I'm trying to be descriptive about the situation and everything I've tried so far.

I have a seagate barracuda 80gb ultra ATA drive, model number ST380013A. Sometimes the BIOS doesn't detect it, but after a few reboots and about a 10 second pause it finally does detect it. I'm having a hard time trying to get data off of it.

I sent it to a recovery service, but they wanted $1100 which is more than the lost data is worth (last backup is february), but I'd still like to try and recover it myself. FWIW, they said it's "slow, but allows reading, with some damage to system areas" and that it would take a lot of time but they were confident of a successful recovery.

My goal was to try and copy the partition, ignoring bad sectors, to another drive. Then I could CHKDSK and save as much as possible. This is proving to be not as simple as I thought (hey, I was optimistic).

I booted from the linux system rescue CD, and I can see and mount the NTFS partition (although only with the regular ntfs driver, not the ntfs-3g driver). If I try to list the directory contents, it appears to lock up, but in reality it takes about 5-10 minutes to finally give an input/output error. I managed to copy a directory by specifying the path to it directly, and I guess I was lucky that directory wasn't using any bad sectors.

I attempted to use 'ntfsclone' and 'dd_rescue' but ntfsclone never seems to do anything, and I get a kernel panic with dd_resce after about 2 hours (something about invalid CPU state). Possibly not related to the disk issues, but who knows. Either way, the dd_rescue status after that time only gets through about 4096k and 88 bad sectors. I tried dd_rescue in reverse "-r" option and it gets through about 650000k before encountering a block of bad sectors which takes about 5 minutes per bad sector. Needless to say I can't get very far when the kernel panic keeps happening.

I'm pretty sure my computer is OK, I put the drive in another IDE system and it had the same issues. I also have my computer running Windows on a different drive, and it seems to be OK.

I don't hear any loud noises, but if I get close to the drive I can hear the usual seeking noise (normal HD read/write noise) like it keeps trying to read the same sector.

So here's my questions:

1) Is it worth trying to replace the PCB? I found an identical model on eBay (drive model and firmware rev). Could the current PCB be the problem or is this a dead end route?

2) Is there a better utility to copy the partition to another disk (skip/zero bad sectors)? The SeaTools CD seems to do much better at reading the disk when running the long test (doesn't spend as much time on bad sectors, doesn't crash). If only it had a utility to copy the sectors. Or if only the system rescue CD didn't panic on me I could leave it running for a few days and possibly let dd_rescue finish.

3) Should I try and let the SeaTools CD remap all the bad sectors? It got through about 12% of the drive before it gave up with 100 read errors. The bad sectors seem to be in consecutive sections. Like block 212800 through 212850. During one long test it actually had only the even blocks show up bad, but usually it's the whole range.

4) Any other ideas?

Well, thanks for any and all help. I feel like I'm close to being able to recover something but I need some expert advice.

1) Is it worth trying to replace the PCB? I found an identical model on eBay (drive model and firmware rev). Could the current PCB be the problem or is this a dead end route?

you have to gamble some money on this it might work or it might not work

what your saying sounds like the problem we had
and we replaced the logic board but had the same fault

turned out that it was the preamp control head that was breaking down

or the same problem can happen with the following below faults

1)amp control head is weak
2)firmware corruption
3)service track has got errors

once the hard drive has filled up with to many data errors in the factory list
it will simply shut down and wont work again
unless you cancel out all the errors it found to begin with.
only way of doing this is by going into factory mode
which you need special software for and its not cheap.

i used a sector repair program but it had way to many errors
more like 50,000 bad sectors


check out a program called HDD Regenerator see how many errors it picks up
if it picks up way to many like 50,000 its a bigger problem.

your have to becareful when using a bad sector repair program
as your data could be on one of those sectors and your lose it for good


we offer a lot better services for data recovery
our goal is no fee for checking your drive out (yep its total free)


we based in toronto canada

Check terminal for problems , then after getting the data recovered, selfscan to repair the surface (if no head problem). Selfscan will destroy ALL the data so don't do it until you have backed up all the data.

Dont run CHKDSK, may be Bad sectors, dont try it urself which u will mess it up more.
mension ur are where u r from...
someone can help you of the forum who is located in ur area...

You can also have a try to me if you wants.
my pricing started from $100 upto $700 only...

You can share your expected price that i can help you if possible.

my email address: shahij@gmail.com

_________________
__________
There is no substitute for education and experience
THANK YOU
SHAHI
shahi.mahbub@gmail.com

Thanks for the replies.

BlackST, if by "terminal" you mean building a terminal adapter and hooking the drive up to a COM port, that's way beyond my skill level.

craig/shahij, thanks for the offers but I already decided the data is not worth paying for recovery. This is really just a learning experience for me so I was just looking for a nudge in the right direction.

I ran ddrescue again last night, and it managed to get through 12GB worth of sectors with only 3MB of bad ones before the kernel panic happened again (took about 6 hours). So I figure I can eventually get through the whole disk with enough ddrescue sessions, each time starting where the last one crashed. Is there a risk of making things worse by doing this? At this rate it'll probably take another 7 days to get through the whole thing. But if the ratio stays the same and I only end up with 24MB of bad sectors that seems really good.

I've also read about putting the drive in the freezer, would that possibly help get the data off quicker?

I'll probably try the PCB change after I get through the whole thing with ddrescue. That seems like a risky thing for me to try.

If anyone has any other ideas I'd appreciate it.

If you're not paying for data recovery, then you have nothing to lose. Anything you get is gravy.

Putting the drive in the freezer is a bad idea, but keeping the drive cool is a good idea.

If the drive is taking a long time to appear to the operating system I would suggest there may be more than bad sectors as is likely the heads are weak and failing. A pcb swap is very unlikely to help!
Download a demo of R Studio and after installation boot up into windows with your drive on the secondary controller.
Leave the system the amount of time required so you can at least see the drive in my computer.
Open R Studio and leave the pc for as long as it takes for r studio to find the drive. If it finds it then attempt to explore the drive inside r studio programme.
I'm not sure what the demo limitations are but with a licensed version you would be able to recover just the required files and folders and nothing more.
By default theR Studio is set to 2 read retry attempts. If things are bad you can set this to 1.

I get the following Kernel Panic when trying to copy sectors off the drive in Linux:

CPU 0: Machine Check Exception: 0000000000000004
CPU 0: Bank 0: 2200000084010400
Kernel Panic - not syncing: CPU context corrupt

Now from what I can tell reading Intel docs, this means "FSB parity error". Could this actually be caused by the bad drive?

I don't think so.

pepe

_________________
Adatmentés - Data recovery