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.

Recently reinstalled my computer with the final release of Windows 8, and to celebrate this, moved all the important data from my 2 pieces of 1TB HDDs (a Samsung HD103SI and a Western Digital WD10EARS-005B1 Green Caviar), converted both to GPT and Dynamic, and created a striped volume across the two disks. After this, I created an overlapping 2TB partition, for my downloads (storing Ubuntu images, my school DreamSpark installers, etc.).

It was working fine, until today, during a heavy P2P download of a Debian collection, the drive suddenly became unavailable. First I thought maybe some Windows bug, as it used to happen, so I rebooted. No, the drive was still unavailable, so checked it under Disk Manager. The so far perfectly working NTFS partition somehow is now unrecognized, RAW, and I cannot access my data.

Is there any way to restore the partition somehow? I have some very important data there, and it would be quite a cutback if I lost all those documents, not to mention the mass amount of (kinda retro or vintage) collection of old Linux installers.

share|improve this question
    
You understand using a stripped raid doesn't allow for a single device to fail right? If the a drive failed then your data is gone. This is what a backup routine is for. Besides using "Green" energy drives within a raid is a horrible idea. –  Ramhound Feb 19 '13 at 20:58
    
None of the drives failed, they all check back with perfect condition, diskpart shows the dynamic definition partitions, and yes I know that using a green drive is a bad idea, but I needed a united way to store my files. Using striped was a bad idea also. –  fonix232 Feb 19 '13 at 21:01
    
Doesn't matter if they are checking out "today" as good something happen that caused the RAID to fail. As I point out if one of the drives failed while data was being written to the drive, then the entire raid fails, even if the drives report as having nothing wrong. –  Ramhound Feb 20 '13 at 11:53

1 Answer 1

It's too late for you now (obviously), but for anyone in the future:

Try using a Linux boot disc to recover the contents.

While Windows is a bit unforgiving when it encounters "bad stuff", Linux thrives on chaos.

Ramhound is oversimplifying the issue as a complete loss...

If the hard drive is completely dead (clickity-click, etc.) then you are likely out of luck.

However, if the drive is failing but still responds to the underlying hardware commands (ATA, etc.) then a Linux boot disc may be able to recover a relatively large amount of your contents.

You'll probably need an attached USB to use as a destination (because you can't trust the failing drives).

Only the file segments that happen to overlap with the failed sectors will be lost. The files that are adjacent or even far away from the bad sectors will be unaffected by the failure and their stripes will be accessible/recoverable.

share|improve this answer
    
There were no failing sectors, the load of the P2P download was too much for either Windows or the disks to handle, and damaged the partition table. I used those disks for a year or so, up until the Samsung one gave up, without any issues. Dead block search (just ran after recovery) showed no more dead blocks than previous check. –  fonix232 Mar 4 at 13:34
    
Interesting. Out of curiosity... were your disks partitioned as NTFS? I can see the journaling aspect of the FS possibly being unable to handle the load... however, unless you have a huge download pipe (or a really really old system), the performance of the hard drives is much greater than the performance of a network service. A 1Gb network connection is still just 128 MiB of data per second, which should be consumable on both the Samsung (~113MiB/sec) and the WD drive (~110MiB/sec) since they are stripped lets say you lose 20% performance: 113+110 = 223MiB: 223MiB*0.8 = 178.4MiB sustained. –  Nick Mar 14 at 15:30
    
I know it should be able to handle such load, but it wasn't. Most of my hard drives actually can handle up to 4-6MB/s downloads using torrent - I'm pretty sure it is uTorrent's crappy caching and writing method that overloads the disks (any time I checked, uTorrent was using up to 200MB/s of disk I/O!). –  fonix232 Mar 14 at 18:13
    
Something you could check is Task Manager. If you are running windows 7/8 (or equivalent server versions) then you can see advanced stats and see how many file handles are open by and assigned to uTorrent... if there are hundreds then yeah that could be your bottleneck (not to mention a bad design by the uTorrent folks). –  Nick Mar 16 at 15:03

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.