I have a RAID card (a RocketRAID 2310) which I was using as a second hard drive controller, rather than a RAID controller. Connected to it were two unrelated 3TB drives each in it's own separate JBOD array. Following a PSU and Graphics Card upgrade, my MB has decided it no longer likes the RAID card and is refusing to boot with it in the machine. So I reconnected the two drives to spare sockets on the MB (an Asus p9x79 ws).

On booting into Windows (Win7 64 bit Ultimate), Windows can see one drive fine, but the other it thinks is uninitialised. I would like to recover the data on the uninitialised drive.

The uninitialised drive is a 3TB Hitachi Deskstar 7K3000 (0F12450, HDS723030ALA640). The drive is fine physically; it spins up fine and does not click or have any other worrying noises. It is/was formatted as a single GPT/NTFS 3TB partition.

I have attempted nothing with the drive so far. When Windows Disk Manager prompted me to initialise the drive, I clicked Cancel knowing better than to do anything that might cause anything to try writing to the drive. I have disconnected the drive for the time being until I have determined an appropriate course of action.

The data on it is not so valuable as to be worth going to a data recovery company. On the other hand, ideally I would not like to lose it, so I am happy to attempt a careful DIY recovery using appropriate software.

The drive is probably about two thirds full, so we're looking at several thousand files totalling around 2TB. I can soon get an extra drive to aid in any recovery strategy as I was thinking of adding another 4TB HDD to my system anyway.

Any help or advice would be gratefully received.

When you say that "Windows can see one drive fine", do you mean that Windows can access the files on this "good" drive, or do you just mean that the drive is visible in Device Manager or Disk Management but not necessarily in Explorer? I ask this because the latter symptom would be consistent with a broken RAID 0.

_________________
A backup a day keeps DR away.

The good drive it can see, which is a 3TB Seagate, can be accessed and used fully as normal. I can explore it with Windows Explorer and access files on it using whatever application. e.g. VLC to play video files.

Neither drive was ever part of a RAID array. They were set up as two entirely separate and independent JBOD arrays. Each drive was the sole physical device in it's own JBOD. They each presented in Windows as separate drives each with their own drive letter. No spanning, each unrelated to the other and certainly no RAID 0, 1 or anything.

Thanks for the clarification. The term "array" implies that the JBOD consists of more than just a single drive (at least to me), hence my confusion.

When Windows complains that a drive is not initialised, it means that it can't make sense of the contents of sector 0. That's where the partition table (or dummy MBR) is located. If it were my drive, I would examine sectors 0, 1, and 2. I would then locate the boot sector(s) of your data/OS partitions. My next step would be to locate the MFTs. That should give you some idea of the state of the file system(s).

You state that the drive had approximately 2TB of data. I'm wondering whether your present configuration is subject to a 32-bit LBA limitation (it would possibly be using a different driver), in which case any new data that is written beyond the 2TiB point would wrap around to sector 0, thereby trashing the file system.

You could use a free disc editor (eg DMDE) to read sector 0.

_________________
A backup a day keeps DR away.

No, this isn't the case. It was originally configured with GPT and could use the full 3TB and has had much more than 2TB on it in the past. I was merely giving an indication of the approximate amount of data that was on the drive.


While I could examine the first few sectors as you indicate, I would have no idea what to expect or what it should look like.

Try to connect the drive the an external usb rack and use a software recovery tool like R-studio or Easuse

The drive was originally initialized as a GPT volume and the full 3TB was formatted as a single NTFS partition. It has been fuller than 2TB in the past. It's just that now its approximately two thirds full.



I've no need to connect it externally. I have a case with backplanes. All I need do is plug it back into it's drawer. I will look into R-studio and Easuse, so thank you for those suggestions.

Part of the reason for my posting is to be pointed toward reliable recovery software (preferably free) and to avoid any dodgy fly-by-night stuff.

Right now I'm tempted to plug the drive back in and see if Partition Find and Mount can see it and copy the files to a new drive (I've ordered one already).

Oh dear. It seems from my reading that 'Partition Find and Mount' is 32 bit only and is supposed to crash on 64 bit O/Ss. Is this the case? I'm on Win7 Ultimate 64bit.

I'll have to look for alternatives. I'm checking out EaseUS Data Recovery, but I'd have to buy that to recover more than 2GB and it's quite expensive.

The point I'm making is that, when the drive was attached to the RAID card, it would probably have been using a 48-bit capable driver. Now that it is attached directly to your motherboard, it would be using a different driver which may be limited to 32-bit LBAs (eg earlier Intel RST driver).

_________________
A backup a day keeps DR away.

Point taken. But it is Windows 7 which is capable of dealing with drive sizes larger than 2TB and I do have a 4TB drive on there already. None-the-less I note what you are saying and I shall be mindful of the possibility.

I mentioned my issue to a colleague who works in the storage team where I work just to seek advice. It turns out they use R-studio. Long story short he lent me a laptop with a copy of R-studio network on it. Took it home and after a bit of mucking about connected remotely to my home machine.

It saw the partition on the affected drive no problem at all. As I suspected, it was completely intact, directory structure and everything. I recovered everything I wanted to my brand spanking new 4TB HGST. No bad sectors or anything.

Once I was sure everything was accessible and working properly, I reinitialised the old drive with GPT and formatted a new NTFS partition.

So all sorted.

bit late, but just a note - I have Partition Find and Mount on my 64-bit win7 laptop, and am able to scan for partitions, create images no problem. Never seen any issues so far.

Thanks for replying. I shall remember that just in case (God forbid) there's ever a next time.

In fact, thank you to everyone that replied. Your help has been most appreciated.

I think the issue stemmed from the way my Highpoint RAID card/drivers/utility had set up that disk rather than any kind of catastrophic failure. Wiindows clearly didn't understand the setup once the Rocketraid drivers and card were out of the equation, though why that should be I'm not sure; the Windows understood the other drive that was on there just fine.

Anyway, as a result, everything was there and intact, as I suspected it would be. I just needed something that could bypass Windows, see the data partition and the file structure therein.

Thanks once again.

I think you need to do more troubleshooting to narrow down the cause of the problem, otherwise your "good" drive might become a time bomb. Typically a drive that has been formatted in GPT mode will have a data partition that begins at the 100MB point, with a small Microsoft "reserved" partition at the beginning. It could be that wraparound has occurred, wiping out the GPT partitioning information plus the beginning of the first partition, but leaving the data partition untouched. If that is the case, then the same problem could happen to your "good" drive once you write data beyond the 2TiB point.

_________________
A backup a day keeps DR away.

Ok. I need to understand what it is I'm looking for that might cause this wrap round.

I have four disks that have always run from the MB sata ports. Only two were formerly running under the control of a RAID card. 6 total. They are all GPT disks, every single one. They all have a 128MB Microsoft reserved partition at the start of the disk followed by a data partition.

Of the two formerly running under the control of a RAID card, before I reinitialised it the bad disk had three partitions. One strange one at the front marked something like "GPTpart2" or something like that, then a Microsoft reserved partition and then the data partition. But I didn't note it all down and I don't remember the sizes.

The "good" disk just looks like all the rest of the disks, a 128MB Microsoft reserved partition and a data partition. As far as I can tell it's set up properly for a GPT disk. Comparing it to the others, it looks the same.

You mentioned drivers, I've checked the drivers Windows is using to access all my disks. They all use the same Windows driver: it's version 6.1.7600.16385 (win7_rtm.090713-1255). Unless you mean something else in which case I need to know what I'm looking for and where to find it.

So, what do I need to look for on this "good" disk, to see if there might be an issue? Or do I play it safe and use R-Studio to recover all the data from that one just in case and then reinitialise and reformat the "good" disk and move all the data back?

Thank you for your concern and patience.

Sorry, I'm confused. Did you reinitialise the drive while it was still attached to the RAID card, or after attaching it to your motherboard?

Anyway, to verify that you don't have some strange 32-bit limitation, I would first examine sectors 0, 1, and 2. Then I would examine sectors 2TiB, 2TiB + 1, and 2TiB + 2. If the latter three sectors are identical to the first three, then this would confirm that wraparound was occurring. If you get an error when trying to read beyond 2TiB, then this would also indicate a problem.

In DMDE you would ...

uncheck Show Partitions
select your Physical Device
Editor -> Goto Offset
Sector = 0 or 1 or 2
Sector Offset = 0
From Start/End
Hex

Then ...

Editor -> Goto Offset
Sector = 100000000 or 100000001 or 100000002
Sector Offset = 0
From Start/End
Hex

_________________
A backup a day keeps DR away.

Right let me explain the sequence of events.

1. I have a server case with two drive cages with backplanes each able to take four 3.5 inch hard drives. At the time I had a MB with not enough SATA connectors to connect them all. So I connected the four SATA drives from one backplane to the MB and I bought a PCI-E RAID card to run the second backplane of four drives. (Everyone seemed to have stopped making PCI-E SATA controller cards with four internal SATA sockets, so I had to buy a cheap RAID card instead.)
2. I bought a 3TB Seagate drive and put it in a slot that connected it to the RAID card. With the RAID utility I created a JBOD array and added that 3TB Seagate device to it, because that was the only way it would let me run a single drive. I initialized the drive in Windows with GPT and created a single NTFS partition on it using all the space.
3. A year or so later, I bought another 3TB drive, this time a Hitachi one. I put it in another slot that also connected it to the RAID card. With the RAID utility I created a second JBOD array and added that 3TB Hitachi device to it. Once again I initialized the drive in Windows with GPT and created a single NTFS partition on it using all the space.
4. About eight months later, at the end of 2012, I upgrade my MB and CPU. The new MB has more SATA connectors, but I keep the RAID card. Everything is rosey and working fine. Those two drives are happily using all their capacity, no sign or corruption, wrap around or anything all under the control of that RAID card.
5. Fast forward to last weekend, when I upgrade my GPU and PSU. Having done the work inside the box, I discover it won't POST or boot with the RAID card connected. I could see the RAID card do it's own post and I could get into it's BIOS, but the motherboard was throwing an obscure error. No idea why.
6. So I decide that I didn't need the RAID card anymore anyway and remove it. So at this point: no more RAID card. It's gone.
7. I attach the SATA leads from my second backplane directly to the MB. I ended up with one drive that Windows recognised: the Seagate which seems to be working fully. And one drive that Windows saw as uninitialised: the Hitachi.
8. The partition list for the Seagate (good drive) looked like this in R-Studio:
   
   • Microsoft reserved partition 128MB
   • Data partition 3TB-ish
9. The partition list for the Hitachi (bad drive) looked like this in R-Studio:
   
   • GPTpart2 ???MB
   • Microsoft reserved partition ???MB
   • Data partition 3TB-ish
10. Using R-Studio I recovered all the data from the Hitachi. And I have now reinitialised and reformatted the drive.
11. Because of your warnings, as a precaution, I am recovering all the data from the Seagate (good drive) using R-Studio even though Windows seems to see everything on it fine.

Because R-Studio is still running that recovery, I cannot go check the sectors you specify on that drive.

I can't edit my previous post, so here is an addendum to step 5.

When I got the obscure MB error, I checked and rechecked everything I'd done; it was all good. I then posted the MB error I was getting on the Asus forums and a techie suggested it might be something to do with the SATA drives.

So I went through a process of elimination, disconnecting SATA drives, backplanes, etc. It wasn't until I removed the RAID card that suddenly everything started working again.

With the RAID card in the machine, I could only get at it's onboard BIOS setup, so what I could do with the drives was limited and I didn't want to risk causing any data loss. On top of that, I was no longer sure how much I could trust the RAID card. Hence I decided to ditch it and connect it's two hard drives directly to the motherboard.

Also, another addendum to No.1: the RAID card was a Highpoint RocketRAID 2310. When I bought it, part of my criteria was that it had to be able to manage drives larger than 2TB.

Well first of all I want to explain you what is RAID...

RAID is a storage technology that combines multiple disk drive components into a logical unit for the purposes of data redundancy and performance improvement.

You can find more on Wiki: http://en.wikipedia.org/wiki/RAID

When you construct the RAID between 2 or 3 hard drives it become one hdd... the read&write operation will perform like one hard drive. The data will be copied like one hdd so its mean you are using 6 TB of hdd. When one hdd fail customer always try to get some data from working hdd but it is no possible. The only possibility of data recovery is to repair fail hdd... when you successfully repair fail hdd just reconnect the RAID and you can get all of your data without using any software. If you write any data on RAID hdd sperately the all data lost forever.

I think you understand and have a Good Luck.