XiliX | # 1 | 2013-08-20, 14:08 | Report

Does anyone here know if its possible, and if so, how to recover data from a corrupted RAID-0 array?
I have tried a linux live disk, (mandriva 2011) but the harddisk partitioning program told me the disks partitiontables are too corrupt to restore.
Besides that, one of the two disks seems to be in read only mode.

I dont have a lot of experience with raid, so any help is welcome.
Thanks!

juiceme | # 2 | 2013-08-20, 17:29 | Report

Well, usually using RAID-0 is asking for trouble.
Even a small error will render a lot of data unusable so my advice is always stay well clear of it

Usually even the speed increase is not that much when compared to a single disk or RAID-5 so there's no need ever really to use RAID-0.

XiliX | # 3 | 2013-08-20, 18:15 | Report

I couldnt agree more, but its not my own computer, and the damage is allready done.
So the question remains; is there any way to retrieve this data?

minimos | # 4 | 2013-08-20, 20:55 | Report

Do you mean that the program couldn't even read the MBR area?
In that case, I don't think you could do much to recover any significant data from the array: data is striped on the two disks and one of the halves is basically unreadable, so what's left is likely useless.
You could try using photorec & testdisk suite check testdisk package with Ubuntu/Debian distros), but I don't know if it can work with RAID arrays or only single disks.

wicket | # 5 | 2013-08-20, 21:29 | Report

RAID 0 provides no redundancy so you're pretty fsck'd (pardon the pun ).

I would first image the array to get the data onto reliable storage to avoid further corruption.

Can fdisk read the partition table? If not, do you have a backup of the partition table that you can restore? What file system are you dealing with here? Have you run fsck?

XiliX | # 6 | 2013-08-20, 23:10 | Report

I have;
-tried booting, the computer hangs after bios, and after initialising the raidcontroller.
- booted mandriva linux 2011 from dvd, and ran diskdrake, its main hdd utility. Upon starting tis it told me that the partitions on sda and sdb were too currupted, and could not be restored.

I have not tried fsck, as the drives arent recognised as beeing partitioned. I have used "dmraid -a y" command, but that didnt work either.
And im affraid i cannot use datarecovery tools as the data is stripped over two disks and they cannot be seen as one as long as raid isnt working.
So judging from here i guess the datas is lost than..
Or could running fsck still work?

Its a win7 system, i guess using ntfs.

wicket | # 7 | 2013-08-21, 05:10 | Report

So you've got a fake RAID setup. Make sure the controller driver is loaded before running dmraid.

I've never used diskdrake but it's likely that it would not be able to understand your fake RAID disk format even if it was in working order.

fsck won't be of any use on NTFS.

How soon after initialising the controller does it hang? If it's immediate, it could well be a controller failure rather than disk corruption.

juiceme | # 8 | 2013-08-21, 07:00 | Report

There are no generally available tools to fix this kind of problems.

I'd imagine data recovery companies like Norman Ibas have some in-house tools that could be used to do rescue operations on all kinds of NTFS systems, even striped. (as propably most of their customers run some windows variant)

How large are the filesystems?
If you really want to dig into it, and if you have the patience you could do a raw dump of the partitions to disk files and then depending on the structure of the data attempt to write your own tools to recover something.
(assumptions; the data must not be encrypted of course, and have some recognizable structure like plain text files...)

XiliX | # 9 | 2013-08-21, 07:08 | Report

The bios inits, than nice progressbar fills, telling me its checking the disks. After its done it tells me both disks are ready to use, begins booting the os (which isnt there) and hangs on a blinking cursor.
If i physically remove one drive, at this check point, it gives me a "disk array failure" message, and the option to press cntl-f to configure raid.
So i suspect the data on the disk is messed, not the raid controller.
Ah, i recall, before the system died it gave a number of BSOD's and eventually hang on a distorted screen...

edit:
its two 1TB drives. I also thought of using dd to backup the disks enterely, but we do not have enough empty diskspace available...

---

Last edited by XiliX; 2013-08-21 at 07:16.

juiceme | # 10 | 2013-08-21, 07:35 | Report

Well, what about sticking a third drive in the system, containing a booting winXP image and setting that as the boot drive from BIOS?

I'd imagine if you manage to initialize the RAID controller and run windows there it might be that the OS sees something on the striped set, even as it's not possible to boot from there