Take the 2-minute tour ×
Super User is a question and answer site for computer enthusiasts and power users. It's 100% free, no registration required.

I've got a dying disk, it's an emotional time, but I've managed to either recover or retrieve backups for most of the data. I've got a few questions though.

When the disk first started dying, boot times went through the roof because it wasn't responding properly, at POST and windows splash. It would show up in disk management in Win 7 sporadically, and it seemed too much activity killed it and it disappeared again.

When it was available it would transfer it would transfer at Kb/s rather than Mb/s.

I managed to get Ubuntu 11.10 installed on my portable disk and boot into that, the dying disk was slow at first and ubuntu threw all sorts of tantrums because it wouldn't respond properly. But when it played nice it did it for a long time and speeds where sometimes in the region of 30-50Mb/s. I managed to get all my important stuff off this way (very happy) but booting back into windows and it still can't see it at all, back into ubuntu and there it is, ready to use.

Question

It's a few tiered question:

  • What could have gone wrong to the drive for the strange behaviour (no bad sectors, but still dying)
  • Does Linux really deal with [insert drive error here] that much better than windows that it was the difference between life and death of my data
  • Why does [insert drive error here] seem to deteriorate over time / useage

With all that's gone on I'm thinking more and more mechanical, but mechanical disk errors are commonplace for OSs and are (in my experience) nicely dealt with. Not this farce.

Specs:

Win 7 Pro x64 on seperate unaffected disk.
Ubuntu 11.10 x64 on portable USB disk.
Dying drive: WD 2TB Caviar Green

Thanks

[update]

Oddness continues, windows now refuses to boot with the dying drive plugged in. I can here a clunk from it every 2 seconds while windows sits at the splash screen, doesn't sound like what I've heard from a dying drive before, but that was a while ago. It sounds like it's trying to start the disk every two seconds and it's refusing. The confusing thing is windows isn't giving up, am I missing something? Shouldn't windows accept defeat and just boot without the drive, or at least show some sort of error, instead of sitting polling it every 2 seconds for 20 mins until I got bored of waiting and killed it?

[update 2]

OK I found something else a little odd last night that makes me not want to give up on this drive just yet. In the Grub prompt at boot (separate issue) even that can read files on the device with next to no lag, and when booting from an Ubuntu live CD that has seemingly good access to the device. Windows however still refuses to boot at all when this device is plugged in. I've seen a few mechanical drive errors before but never inconsistently across OSs.

share|improve this question

migrated from serverfault.com Apr 21 '12 at 14:53

This question came from our site for professional system and network administrators.

    
what filesystem do you have formatted on the USB disk? windows can't read linux filesystems like ext3, ext4 etc. –  ncdownpat Apr 21 '12 at 16:18
    
Sorry I wasn't clear, ive updated the question. The portable disk was fine but the dying disk was working in ubuntu but slowly. Thanks –  Paystey Apr 21 '12 at 17:31
    
Are you able to use the WD advanced diagnostic tools on the drive? Have you tried freezing it? –  Dave M Apr 22 '12 at 18:37
    
Ill give the tools a go, but the drive doesn't show up in windows so im not holding out hope. And the tools dont run on linux –  Paystey Apr 22 '12 at 18:46
    
OK just tried the data life guard tools form the WD site. They can't see the drive either, guessing it just piggy backs on windows detection of drives. Can't find any WD specific tools for linux. –  Paystey Apr 22 '12 at 19:43

4 Answers 4

up vote 2 down vote accepted
+50

There are many things that could be going wrong with your drive. From simple head faults to overheating motor or even overheating or dying controller chips.

To me your fault sounds a lot like the either the controller or motor is failing in some way.

Integrated circuits, such as the controller that takes magnetically read data from the head mechanism and converts it to the electrical data your computer expects, are very delicate and with age and use can wear out the same as mechanical components wear out. Either it got hot too many times, has a small imperfection that didn't affect it to begin with but time hasn't done it any good, or it's simply old.

When electronic parts get old they fail in surprisingly similar ways to mechanical parts. Sometimes they "kinda" work and you can get them to move data, but in a slow and painful way as it has a lot of error correcting logic that is having to kick in to do anything. When they're in that state extra heat can make the chip get into an illegal state and require a power cycle (just like your computer processor) or if you are lucky it may soft-boot itself without the OS knowing.

Similarly the drive motor failing may cause somewhat similar problems to what you are experiencing. The motor tries to spin up on boot up, but because the bearing or something is going then it either draws too much power causing a heavy load on the power regulators and controller and making them overheat and effectively "reboot" again or it simply takes too long to spin up causing the controller to think it has a fault and to "let it rest" a moment before trying again.

Linux may simply be more resilient about dealing with failing hardware. There are different ideologies behind the operating systems and how Linux interacts with the hardware could well make it more tolerant of hardware that is spewing errors at it while Windows expects the hardware to just do its job.

Lastly:

Why does [insert drive error here] seem to deteriorate over time / useage

The same can be asked of any piece of equipment in the world, be it cars, trains, microchips or even biological organisms and the answer would always be along the lines of "because everything experiences some kind of deterioration over time."

We can't make frictionless bearings for motors, we can't make make a microprocessor that doesn't waste energy as heat and we can't make cells that live forever. The work that these items do costs them a tiny fraction of their lifespan a lot of the time it's just a guessing game as to which part in any given item will die first.

share|improve this answer
    
OK That would mostly make sense, if it was failing a little and doing lots of error correcting. I would have thought the SMART data would be more definitive with problems like that but hey ho. I've updated my question with something I found last night that I'm not sure fits. –  Paystey Apr 25 '12 at 7:39

If you want a possible solution to the drive, if it can be saved, then SpinRite can do it.

I have never found a better program to solve HDD problems like this. I don't actually see a question so I assume you want a solution so you can continue to use the HDD at least for a little awhile.

share|improve this answer
    
Thanks I'll give it a go see if it can be saved, but as of right now windows won't even boot with it plugged in. –  Paystey Apr 24 '12 at 13:10
    
Stop booting windows on damaged drive, each read or write to disk just makes scratch on disk surface bigger... –  ZaB Apr 24 '12 at 13:39
    
SpinRite is used with a boot CD, not from Windows. –  harrymc Apr 24 '12 at 15:04
    
SpinRite tries to read every sector on your drive up to 2000 times, then copies the data elsewhere on the disk. That is fine on small drives, but if you are running this on a drive larger than 100Gb then you can forget about using your computer for anything else for a week! I tried to use SpinRite 6.0 to recover a 150Gb drive a few years ago - it was not worth the effort or expense. –  dunxd Apr 24 '12 at 15:31

You say that you are booting Ubuntu from another drive than the dying one, so Ubuntu doesn't need much from the bad drive except to be a little bit readable. Windows on the other hand needs to boot from this disk, so it has many more requirements and needs the disk to be in a much better shape.

It is impossible to really know why there is a clunk coming every 2 seconds in Windows boot. If I had to guess I would say that Windows is issuing a seek on the drive. Probably some badly-programmed Windows boot program doesn't know that this is a lost cause, but whatever it is doing is not succeeding. It might be spending its time re-reading the disk a certain number of times, which might take two seconds for the retries, then reissuing a seek and trying again.

You could try a chkdsk on the drive, but the chances of it succeeding are very slim. The clunk could mean a mechanical problem, so you would be damaging the disk some more by trying to use it.

share|improve this answer
    
Windows boots from a separate drive. When it did pop up in windows and I got to run some simple self test, everything came back fine. –  Paystey Apr 24 '12 at 15:38
    
I understand from this that the bad drive is not the system drive. Did you try a chkdsk ? –  harrymc Apr 24 '12 at 15:45
1  
Yes I did very early on when the drive was popping up sporadically in windows and all returned fine apart from it saying it fixed a couple of bad sectors, but that was evident from the SMART data saying that there were reallocations. –  Paystey Apr 24 '12 at 15:48
    
What you are describing is a random error event, that if it doesn't arrive once then the drive continues on working correctly. This may be some bad contact, maybe connected to the temperature. Something like that: when the disk drive is hot then some part expands and the contact improves. –  harrymc Apr 24 '12 at 16:12

You have your data. Disks are cheap. I wouldn't spend too much time worrying about what went wrong. Be happy you took backups and made the effort to recover, and move on.

If the disk is new, see whether it is covered by the manufacturers guarantee, and if so take them up on it. If it is old, accept the fact that disks wear out, and replace it.

There are a lot of reasons why different OS might get different performance out of a disk, but I would doubt that for a non-server computer this is a worthwhile factor when choosing an OS.

Unless of course you are a student of OS/disk technology, in which case someone else may answer your question :-)

share|improve this answer
    
Yup, WD had a stern email from me and an RMA request as the disk is only 6 months old and was an RMA for a previous crapped out drive. Both issued before the Thailand flooding and therefore not affected by the disaster. –  Paystey Apr 24 '12 at 15:39
    
There you go then - I don't see any reason to worry about which OS is better at handling a broken drive. Maybe get a refund and buy Seagate instead, but I think you are probably unlucky, or have some sort of environmental factor that is breaking drives in your computer. –  dunxd Apr 24 '12 at 16:20

Your Answer

 
discard

By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.