Toshiba failure x2

[lsn]

Background story first, as I am new here...

I've been tinkering with computers for about 8 years now.
I'm very familiar with the sound of a dead hard drive. "clicking noise of death."

But that was the extent of my knowledge until recently..

I'm now an electrical engineering technology major in my second/ third year of classes.
I've read the page about the internals of an HDD, which was a great! I had a generalized idea of the workings.... Platters, head, arm... still a little off on the acronyms...
I've been lurking around these forums for about two hours now and I feel like this is the place to post my predicament.

So here it goes.

I bought a Dell XPS M1530 with a Toshiba 320GB Hard Drive about 4 years ago.
The original died three-ish years in and Dell replaced it with a Toshiba MK3265GSX 320GB 5400RPM
To be honest I don't remember how the first one died, all I knew was that I had it backed up and Dell was replacing the drive under warranty.

So last fall the replacement went out on me.
I came home one day to find my laptop completely unresponsive, not even mouse movement.
I gave the laptop a hard shut down and it attempted to reboot, getting though the boot screen and into Windows... that lasted about a minute until the OS became unresponsive once again. Before shutting it off, I figured it was my hard drive and sure enough it sounded like it was stuck in a read/ write sequence. The drive remains spinning but will not read properly. When I first tried plugging it in via Sata Usb bridge, it would read for a minute then Explorer would freeze. So I left it alone again.
Paranoid because I am a photographer and have rather personal and sentimental pictures stored on this drive.
A fellow Engineer told me there was a way to switch out the boards on the Hard Drives, which I now know is called the PCB.
So I went on to Ebay and did my best to match the exact part number and every other number I could find on the label.
Sadly, the picture of the drive was not what they sent me. The drive they sent me was off by one or two digits and what I know believe was the memory module, was different.



Not knowing exactly what I was doing, I made the switch anyway... on both of the drives.
Neither of them worked.
My original hard drive was still making the same read/write pattern but the "donor" drive I had just bought flat out did not work. It spun and read, but was clicking like it just could not read.
I switched the boards back and the "donor" drive worked correctly again. My drive was still toast.

Just recently, I finally got around to installing Linux on an old Dell Dimension 4300.
Upon plugging my drive into the Linux machine, Linux mounted the system reserved filesystem and the 320 partition.
I was really surprised.
I was able to navigate to My Pictures and grabbed some things I really wanted
approx 2GB worth of pictures was attempted to be copied but it got about to 1.6GB before Linux told me there was a read error.

It took a little more fiddling around with Disk Utility to figure out how to unmount the drive to check it... and to understand you can't read SMART data though USB. All the default Disk app. was able to tell me was that both file systems were unclean. Surprise.

I poked around on the Linux forums and found checkdisk... which was no help at all. and I managed to mess up the MBR where I could no longer attempt to boot the drive past Grub (I had dual-booted Linux and Win7 on that drive). However, after messing with HBCD, I can still access the partition where my files are located but I have no clue what to use off that CD so I stopped... Getting off-track....

Something else I feel I need to tell is that just recently, the Donor drive I bought stopped working and has started to make the same noise pattern of my already broken drive... which lead me to believe that I may have a mobo issue on my XPS... then again, as with most hard drives I've dealt with... I let that one sit for about a week and just the other day I got about 5 hours of use out of it... the donor drive I don't care about. It it just interesting that it has failed in the same manner as the old one...

Summation:
The drive in question has failed due to unknown reasons and sound diagnoses suggests that something may just be off-track.
I've made a Linux 12 boot CD and the HBCD. Upon using those, BIOS will not list the drive as a boot option, yet the utilities will read the drive.
SMART data read from Linux:
-1301 Power cycles
-359 Bad Sectors
Passed Self Assessment.
Good spinup time
Reallocated Sector Count : 359

I will attempt to post an audio sound-byte later.

any insight and help would be greatly appreciated.

Thank you very much for your time,
-Ellison

[northwind]

Hello

As I understand it, initially there was absolutely no problem with your pcb. Your pcb swap was a waste of time and money.

I think there is problem with bad sectors on your drive and maybe also some problem with heads. One must accurately diagnose drive to tell for sure.

At this point, it seems pretty straightforward for a pro, but i'm afraid there is not much you can do yourself.

If you don't want to consult a pro on this, and as a last resort, and assuming you accept the risk of DIY (ie. causing further damage to the drive to a point it would be more complex, expensive or even impossible to recover if you decide to go to a pro), I would try to image this drive using ddrescue to a new drive and then run R-Studio on the image.

Good Luck.

[Vulcan]

@lsn:

Thanks for the clear description of events. There are some incorrect comments / guesses about what has been happening at different points in your posting (e.g. no, chkdsk did not alter your MBR, but it's never a good idea to run chkdsk on a failing drive as it can sometimes make things worse, and it's not clear to me why you mention Linux and chkdsk in the same sentance, as chkdsk is a Windows program) - but that's all in the past.

Based on your description, I completely agree with the comments from northwind. It's your choice how much risk you want to take by attempting a DIY recovery, or instead to use the services of a reputable DR company (members here would be able to give recommendations, if you ask).

If you decide to take the risks of DIY, then I would: Stop trying to read files from that failing disk, as you have been doing (i.e. don't mount its filesystem(s)) and urgently clone the whole of the failing disk onto working storage (you have several options on these details), before it deteriorates further. Personally I would use GNU ddrescue to do that cloning, but as you'll see in other threads on the forum, some people dislike that. Whatever cloning tool you choose, consider that you might need to restart the recovery part way through, depending on how the disk behaves. My experiences (and reports here) show that you'll typically get more successful recovery if the faulty disk is attached via SATA. Your storage to hold the clone can be via USB if needed.

Given that the disk is still readable (at least partially), I would not expect a reputable DR company to charge an arm and a leg, as they would likely use dedicated cloning equipment, making that part of the process a quicker, less-risky, and perhaps more successful operation than it would be for you. Your disk, your choice

[lsn]

Thank you so much for the insightful responses!

For clarification, I used a Linux application called testDisk, not check disk... sorry for the confusion. That was the program I used a "fix MBR" command.

I just attempted to use SpinRite on the donor drive that failed. It got to about 85 percent when it started to have to read out every bad sector bit by bit... then I think I must have tripped over the power cord because when i came back, my laptop was dead. -_- I was quite unhappy.... so now I don't know if SpinRite would have even worked...

I also understand that if I continue to use bad heads to read the disk, then I run the risk of damaging the platters, correct?
And no, I'm not thinking about replacing the heads myself.

If anything, I'd like to accomplish a DIY diagnosis using MHDD as long as that does not cause additional problems.

However, I'd like to take up a service that will use a dedicated disk imaging tool unless I can get my hands on my own. I forgot where I was looking at them the other day, but I remember reading about them a long time ago... Then again, I understand I shouldn't go about attempting a solution to a vague problem...

Any insight as to why both hard drives failed in the same manner? This is another major concern because I don't want to waste money on a new HDD for my laptop if it's defective and "eating" my hard drives.

Thank you again!

[lcoughey]

I just attempted to use SpinRite on the donor drive that failed. It got to about 85 percent when it started to have to read out every bad sector bit by bit... then I think I must have tripped over the power cord because when i came back, my laptop was dead. -_- I was quite unhappy.... so now I don't know if SpinRite would have even worked...

Don't use spinrite on a failing drive. It is NOT a data recovery program and can cause some serious damage to your drive if the heads are failing.

I also understand that if I continue to use bad heads to read the disk, then I run the risk of damaging the platters, correct?

Yes, you are correct.

And no, I'm not thinking about replacing the heads myself.

If anything, I'd like to accomplish a DIY diagnosis using MHDD as long as that does not cause additional problems.

This cannot be guaranteed without proper diagnosis first.

However, I'd like to take up a service that will use a dedicated disk imaging tool unless I can get my hands on my own. I forgot where I was looking at them the other day, but I remember reading about them a long time ago... Then again, I understand I shouldn't go about attempting a solution to a vague problem...

DeepSpar Disk Imager is the best hardware disk imager, but you are still making the assumption that all the heads are reading. You can contact them about purchasing the tool, it is about $3,500. At the very least, it will help you know very quickly if you have a failing head...then you are definitely over your head. But, if you are willing to spend thousands of dollars for this single case, you are probably smarter to get the drive assessed by a pro and pay a fraction of the cost with a much higher chance that you will actually recover your data.

Any insight as to why both hard drives failed in the same manner? This is another major concern because I don't want to waste money on a new HDD for my laptop if it's defective and "eating" my hard drives.

The "portable" function of a laptop is that you can easily transport the device WHILE POWERED OFF. Many people move their laptops from room to room and don't take any special care for the moving parts inside the drive that are very sensitive to impact while spinning.

So, again, get it assessed by a pro before you make a simple job and difficult job that will be much more expensive.

[lsn]

Don't use spinrite on a failing drive. It is NOT a data recovery program and can cause some serious damage to your drive if the heads are failing.

It is a recovery program as it scans sectors and will read out binary sector values in an attempt to recover data... So I'm hoping that I'm just tripping up on semantics with the assumption that you meant it's not a recovery program in the sense that I'm trying to recover data from a physically failed/ failing drive?


I do have a technical question I'd like answered:

Why is it that when a hard drive has failed or become inoperable once, it can work fine (for a short time) if left alone for a long period of time?
Would this phenomena not suggest an electrical issue with the PCB?
Or maybe wear on the material that expands and contracts to distance the heads above the platters?
This has occurred with every last hard drive I've owned.


Thanks again for the responses!

[drc]

it scans sectors and will read out binary sector values in an attempt to recover data

And what does it do with these "binary sector values" assuming it is successful in reading them?

[lsn]

it scans sectors and will read out binary sector values in an attempt to recover data

And what does it do with these "binary sector values" assuming it is successful in reading them?

I never found out as I accidentally unplugged the laptop before it finished...



from the report it would read something along the lines of "All data was successfully recovered from the sector"

[drc]

Well, presumably it has to put them somewhere...

[lsn]

...which it what I was wondering.
The program was originally written in the late 90's I think...
I had an external drive hooked up via USB just waiting to catch my recovered data...

I may try that again just to see what happens.
Note that I'm talking about my donor drive that I tested SpinRite on... I haven't touched/ abused my actual drive with the data I want.
They just happen to have the same problem from what I can hear.

[Vulcan]

...which it what I was wondering.
The program was originally written in the late 90's I think...
I had an external drive hooked up via USB just waiting to catch my recovered data...

SpinRite won't write the "recovered data" onto an external USB drive.

FYI, SpinRite rewrites data back onto the drive it has read it from, which can have all sorts of bad effects, on a faulty drive (e.g. G-list overflow)

[lsn]

SpinRite won't write the "recovered data" onto an external USB drive.

FYI, SpinRite rewrites data back onto the drive it has read it from, which can have all sorts of bad effects, on a faulty drive (e.g. G-list overflow)

Thank you for clearing that up!


Before I get more off-track...

Based on your description, I completely agree with the comments from northwind. It's your choice how much risk you want to take by attempting a DIY recovery, or instead to use the services of a reputable DR company (members here would be able to give recommendations, if you ask).

If you decide to take the risks of DIY, then I would: Stop trying to read files from that failing disk, as you have been doing (i.e. don't mount its filesystem(s)) and urgently clone the whole of the failing disk onto working storage (you have several options on these details), before it deteriorates further. Personally I would use GNU ddrescue to do that cloning, but as you'll see in other threads on the forum, some people dislike that. Whatever cloning tool you choose, consider that you might need to restart the recovery part way through, depending on how the disk behaves. My experiences (and reports here) show that you'll typically get more successful recovery if the faulty disk is attached via SATA. Your storage to hold the clone can be via USB if needed.

Given that the disk is still readable (at least partially), I would not expect a reputable DR company to charge an arm and a leg, as they would likely use dedicated cloning equipment, making that part of the process a quicker, less-risky, and perhaps more successful operation than it would be for you. Your disk, your choice

My original plan was to go straight to a Professional because I assumed the same thing about the data still being readable.
But because I wasn't getting a clicking noise from the drive, I wondered if it was something I could figure out mending long enough to get my pictures and maybe documents.

I may try the disk imaging on the donor drive first to see how well it goes.

As I mentioned earlier, most disks that have failed on me will become fully responsive again if the drive has been left alone and off for an extended period of time (weeks, months, or years). The drives would only remain responsive for an hour or more when booted back up... depending on the time left untouched.

My plan of attack right now, based on everyone's responses, is to try the disk imaging on my donor drive, then let the drive that I want to actually recover, sit for at least a month or more, then possibly try to image that drive based on how well the test recovery on the donor drive goes.

This is unless someone can explain why failing drives will become responsive when left untouched for extended amounts of time and can properly advise me to do otherwise.

[Vulcan]

My plan of attack right now, based on everyone's responses, is to try the disk imaging on my donor drive, then let the drive that I want to actually recover, sit for at least a month or more, then possibly try to image that drive based on how well the test recovery on the donor drive goes

You've already had replies mentioning that DIY attempts have risks. We don't know the value of your data - it's your data, your risk, your decision.

This is unless someone [...] can properly advise me to do otherwise.

That sounds like you're asking for someone to prove to you that a different approach is better. It depends on your definition of "properly advise".

As I have been reminded by recent experiences here, I decline to get involved in any such request, as it can be impossible to explain things to a sceptic's level of required proof - especially since this is a hypothetical discussion, because you haven't got a proper diagnosis of what has happened to your drives, so any replies could only be educated guesswork. Other members may have a different view, of course.

[drc]

FYI, SpinRite rewrites data back onto the drive it has read it from, which can have all sorts of bad effects, on a faulty drive (e.g. G-list overflow)

Vulcan wins the prize

[labtech]

The drive off of Ebay, was it new or used?

[lsn]

Vulcan and other Data Recovery guru's out there,

I really do appreciate and understand your concern. I've read the mechanic metaphor here several times and I've read enough about the internals of a Disk to understand that if I screw things up even more, I may lose all my money and/or lose all chances of being able to recover any data.

[...]

I wanted an explanation to what you shortened there, then a conclusion based on that explanation... but upon further reflection, I realize my fault there and your decision to skip the details and further emphasize the risks and problems when making suggestions...

Yes, I am skeptical to an extent, but at the same time, I really and honestly just want to know what is going on in there.

Additionally, the more I read about recoveries, the more I realize that HDD data recovery is more of an engineering art than a systematic process... and potentially something I might like to peruse when I graduate.

The drive off of Ebay, was it new or used?

Used. And I honestly didn't expect the drive to last as long as it did...

[fzabkar]

...which it what I was wondering.
The program was originally written in the late 90's I think...
I had an external drive hooked up via USB just waiting to catch my recovered data...

SpinRite won't write the "recovered data" onto an external USB drive.

FYI, SpinRite rewrites data back onto the drive it has read it from, which can have all sorts of bad effects, on a faulty drive (e.g. G-list overflow) :(

SpinRite came to be in 1988. I was using it two years later.

http://en.wikipedia.org/wiki/SpinRite

As for how it handles reallocations, here is what Steve Gibson has to say:
http://www.grc.com/srfeatures.htm

Code:

Special Handling of Automatic Relocation:

All modern mass storage drives perform automatic relocation of defective sectors. This presents problems for all other "data recovery" programs because a defective sector which still contains unrecovered data may be automatically relocated by the drive before all of its data has been successfully recovered and removed. SpinRite deliberately disables this automatic sector relocation, when present, until all of the sector's data has been recovered, then deliberately re-enables relocation to allow the drive to replace the defective sector with a good spare. After the relocation has occurred SpinRite replaces the data into the newly relocated sector. No other utility has ever done this.

Clearly Steve Gibson has absolutely no idea how sector reallocation works. If the drive were to relocate the data in a defective sector (ie one that it can't read), then what data would it write to the spare? If the drive were to fill the spare with a data pattern of its own choosing, then how would the OS know that the file that owns that particular sector is now corrupt? In fact an LBA is only reallocated to a spare sector when the OS writes new data to it, thereby telling the drive that the original data are no longer of any value.

Apparently Steve Gibson has never heard of "pending" sectors.

[labtech]

The drive off of Ebay, was it new or used?

Used. And I honestly didn't expect the drive to last as long as it did...[/quote]

Although it could be crazy coincidence that both drives have the same intermittent type of failure, I honestly tend to believe that they are likely both behaving the same because of bad sectors which develop with age and usage conditions.

Nevertheless, can never tell accurately until a thorough analysis is complete on each.

[Vulcan]

@fzabkar,

I don't want to turn this into a "pick the holes in the SpinRite claims" thread (I might do that one in the future ), but we're certainly agreed that there are apparent problems with several of those claims, especially with modern drives.

In fact an LBA is only reallocated to a spare sector when the OS writes new data to it

I've seen additional reallocation behaviours which don't seem to fit that statement exactly as it's written (depending on the interpretation) but again, as they're not relevant to this thread, I'll avoid doing a thread hijack.

[lsn]

@fzabkar,

I don't want to turn this into a "pick the holes in the SpinRite claims" thread (I might do that one in the future ), but we're certainly agreed that there are apparent problems with several of those claims, especially with modern drives.

In fact an LBA is only reallocated to a spare sector when the OS writes new data to it

I've seen additional reallocation behaviours which don't seem to fit that statement exactly as it's written (depending on the interpretation) but again, as they're not relevant to this thread, I'll avoid doing a thread hijack.

Great! No more SpinRite!

So... can anyone explain to me why a hard drive, if left un-powered and untouched for an extended amount of time, will become operational again for a short amount of time?