4

Situation is this. A PC with 2 hard disks, in an RAID 0 Array.

The electronics on one of the disks has failed. I can not find the same board for the disk (I've tried this, removed board from the OK disk, and the second, the damaged one, works fine).

I've made an image with "dd" on linux on a new hard drive (same size, not same model) and now I get "Offline member" in the RAID config screen.

Will I succeed to recover the data which is stored on the drives, any help, any experience with this kind of problem.

And surly, I know it was stupid to put the disks in RAID 0 and store data on them :(

6 Answers 6

11

You cannot.

RAID 0 caused half of data to be on one disk and half on other. You literally have block of data interleaving one another (in order to maximize performance). I am not aware of any tool that can come even close to untangling that mess and even if there was such tool, you can hope for 64K block recovery topmost.

Best chance is either finding same electronics (and hoping that electronics is all that failed) or recovering data from backup.

2
  • 1
    huh, I hoped that there is some tool to fix the mbr records on the disks, and "join" the ghost disk in the raid again without data loss
    – IgorS
    Sep 2, 2009 at 12:46
  • 2
    I think this is getting votes as lots of people are seeing "Missing a disk of a raid 0 partition? Data is gone". As Igor points out he had both halves of the data; the physical drive (1) and a drive image (2). It seems reasonable that his could be recovered.
    – RJFalconer
    Feb 1, 2016 at 16:27
3

This is a common issue and you have uncommonly good potential for recovery. Most modern drives have PCB locked to serial number. If you can operate either drive from the samePCB then make images of each drive using Quetek File Scavenger.

Once both images are made you should be able to mount both images and recover data using professional version. This costs money but is well worth while if needed.

2

I guess you understand that RAID-0 is just striping the disks, there is no redundancy.

You could try RAID Reconstructor.

It will not try to "fix" your RAID. It will merely create a copy of your RAID at another location. It will collect sector by sector from each single drive involved and write these sectors in the correct order to the designated destination. This process is also called "de-striping".

For a RAID-0 (striped) array you will need all drives.

The RAID Reconstructor will recover both, hardware and software RAIDs. It will recover from broken Windows Dynamic Disk sets.

The part in bold may be a problem for you :-(.
I don't know if your copied disk will work...

1
  • Yeah, the bolded part is kicking me in the a.s :( But good to know about the mentioned utility.
    – IgorS
    Sep 2, 2009 at 13:10
2

Old question, but it sounds like you can read both of the disks, just not at the same time. If this is the case, I suspect that mdadm would be happy to run your array if one or both disks were an image file instead of a hardware disk, assuming that mdadm could run the array if both disks were hardware.

0

I think it depends what kind of RAID was used. Is it a mdadm linux software RAID?

If this is the case and both drives are exaxctly the same model and same firmware then you can try to the can use the board from the intact drive to recover the data from the defective drive. You should be able to use the images of the harddrives and use them as normal RAID members.

Note that RAID 0 is not normal RAID where you have redundant information. RAID0 stripes the information across drives/partitions. If you have a file which is larger than the used stripe size (e.g. >64KB, which is a typical value) then part of that file will be on one disk and part on another disk. This makes full recovery impossible if any drive fails.

3
  • "you should be able to use the images of the harddrives and use them as normal raid members" Sorry, but RAID 0 does not work that way. You're probably thinking of other RAID levels, including RAID 1 and RAID 5, which are able to survive partial array loss. RAID 0, by definition, is striping without parity or redundancy and thus every array member is needed for the array to be functional.
    – user
    Sep 6, 2014 at 10:29
  • Yes but he wrote that he can create images from both drives as he can use the board from the intact drive to recover the defective.
    – Florian
    Sep 6, 2014 at 10:32
  • @Hennes Why did you significantly change the content of this answer? The original answer deserves an upvote, your edits to it deserve a downvote.
    – Jivan Pal
    Nov 5, 2023 at 21:43
-1

You cannot recover a RAID 0 drive that has failed. RAID 0 is striped, meaning the data is split. It's like having only half of a piece of paper that was ripped in two... there's no way to know what was on the other half without actually having it.

Next time, try RAID 1. You get less space, but the fault tolerance is n-1 drives, so you'll have a much less chance of losing data.

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .