4TB drives moved from USB enclosure have RAW & Unpartitioned space?

[silverdtvw]

Curious one this - hopefully someone can assist.

I recently picked up 3x Hitachi Touro Pro 4TB USB drives. I hooked up 2 via USB to a Win7 SP1 x64 system - they come ready-to-use pre-formatted in NTFS with a single partition (with some free Hitachi backup software on it).

I copied 2-3 TB's data to each drive, all tested and working.

I then removed the drives from their enclosures (they are SATA 7k4000 Advanceformat (512e) drives inside the enclosure, with a USB-SATA bridge) and connected them to Z68 based motherboard (via onboard SataII ports). Setup as ACHI in bios with Intel drivers in Win7.

All 3 drives - included one that has never been connected via USB - all show up incorrectly. Disk Manager shows them as split into 3 partitions, 1st one is Primary/Active/RAW (with a drive letter assigned), next 2 partitions are Unallocated space. 1st partition is ~1TB, then 1.5TB for the two Unallocated Spaces.

The RAW partition is unreadable to Windows - which obviously wants to format the RAW partition, which I've not done. No request to initialise.


Some googling suggests I'm not the first to encounter the RAW partition issue after removing a drive from USB enclosure - some reports on large Seagate and WD drives, although every assumes it's become corrupt and needs recovery software to fix.

This is clearly not the case here - I'm assuming that either:

1) Win7/Intel drivers are not liking the 4TB partition information - although it's shown as 4TB drive in Diskmanager.
2) Hitachi set the drives up like this, the USB enclosure chip is performing some form of emulation beyond the official 512e standard (very unlikely) to present a NTFS 4TB partition
3) USB drives need less in the way of MBR/GPT info, and the drives are missing some key info here.


Any suggestions about what is happening here, how it can be corrected? Yes, I could delete and reformat - but there should be no need here...

 

[fzabkar]

The USB-SATA bridge firmware in Seagate's 3TB GoFlex externals and WD's 3TB My Books is configured with 4KB LBAs rather than 512-byte LBAs. If you remove the HDDs from their enclosures and connect them directly to a SATA port on your computer's motherboard, you will be exposing the drive's 512e sectors. You will still see the partition information in sector 0, but everything else will be in the wrong place.

You can confirm whether this is the case for your Hitachis by examining sector 0 with a disc editor. Compare the reported number of sectors inside and outside the enclosure. If you could upload the sector dumps, that would be very helpful.

You could also use Microsoft's FSUTIL to determine the physical sector sizes:

fsutil fsinfo ntfsinfo C:

Replace the C with the letter of the drive you want to check.

Here are 3 freeware disc editors:

DMDE (DM Disk Editor and Data Recovery):
http://softdm.com/download.html

Roadkil's Sector Editor:
http://www.roadkil.net/program.php/P24/Sector%20Editor

HxD - Freeware Hex Editor and Disk Editor:
http://mh-nexus.de/en/hxd

BTW, some WD My Books (eg Essentials) are encrypted, so this can be an additional complication. However, this does not seem to be an issue in your case.

 

[silverdtvw]

Thanks for the info - will take a look.

I believe you are correct - I note the external drive is happy to be used directly with XP, which would not normally support a 4TB partition.

Unless there is a extremely unlikely program that can rebuild the block layout I'll reformat when directly connected spend a few more days copying all the data across....

Cheers

 

[sbpetrack]

I have what appears to be an identical problem, but I'd like to ask if there is a better possible conclusion than the unconfirmed one reached in this thread so far:

I have a Hitachi Truro 3TB External USB drive, which I have happily used for a few years now. I recently got a new PC, removed the drive from its enclosure, and installed it internally, directly connecting it via SATA. The result is that what was a single healthy NTFS partition of 3TB now shows up as three partitions: the first a RAW partition of 349.31 GB, followed by two "Unallocated" parttions (of 1698.69GB and 746.52GB each).
I know that the disk is fine, because (with some pain) I took the disk out again and attached the old controller (the case didn't really survive the transformation) and I have my old disk back with zero errors.

Before I thought to look in this forum, I moved the disk back internally, thinking (or maybe it wasn't thinking, it was emotionally hoping ) that I was not careful enough when I connected the disk the first time.

To be honest, it's really a pain to connect this drive internally (I have very large clumsy hands I'm afraid, and there's not much space in there). So the prospect of taking it out to back up the files, then putting it back in a THIRD time to reformat it fills me with dread.

So before I do that, I would be very grateful for fzabkar's or another's answer to the question:

Is there any software (application/utility/driver) that would allow my Windows 8.1 box to deal with the apparently non-standard LBA size of this disk? Or do I have no choice but to take the disk out (again), attach the old controller (again), make certain that all 2.5TB are backed up -- and then (now for the third time!) install the disk internally, format it anew, and then restore from backup?

I know that silverdtvw said here last time (in 2012)....

>Unless there is a extremely unlikely program that can rebuild the block
> layout

but I'm hoping that this extremely unlikely software exists (or perhaps there is a file-system driver that can be configured with weird LBA sizes etc.)

Forgive my ignorance of hardware -- I thank you very much in advance for either a confirmation that I need another cycle of taking the disk out and then putting it back in, or a pointer to the software/method that will enable me to avoid doing this.

scott

The USB-SATA bridge firmware in Seagate's 3TB GoFlex externals and WD's 3TB My Books is configured with 4KB LBAs rather than 512-byte LBAs. If you remove the HDDs from their enclosures and connect them directly to a SATA port on your computer's motherboard, you will be exposing the drive's 512e sectors. You will still see the partition information in sector 0, but everything else will be in the wrong place.

You can confirm whether this is the case for your Hitachis by examining sector 0 with a disc editor. Compare the reported number of sectors inside and outside the enclosure. If you could upload the sector dumps, that would be very helpful.

You could also use Microsoft's FSUTIL to determine the physical sector sizes:

fsutil fsinfo ntfsinfo C:

Replace the C with the letter of the drive you want to check.

Here are 3 freeware disc editors:

DMDE (DM Disk Editor and Data Recovery):
http://softdm.com/download.html

Roadkil's Sector Editor:
http://www.roadkil.net/program.php/P24/Sector%20Editor

HxD - Freeware Hex Editor and Disk Editor:
http://mh-nexus.de/en/hxd

BTW, some WD My Books (eg Essentials) are encrypted, so this can be an additional complication. However, this does not seem to be an issue in your case.

 

[fzabkar]

@sbpetrack, IMHO, the least painful solution by far would be to reinstall the drive in its external enclosure. Otherwise you would need a data recovery tool (R-Studio?) to "recover" your data. That's because Windows is now seeing a 4Kn file system on a 512e physical drive.

The current partition structure is showing a single 349.31 GB partition because Windows thinks that the sector size is 512 bytes rather than 4096.

349.31 GiB x 8 = 3.00 TB

The reason that the volume is "raw" is because its boot sector is in the wrong place, ie it is out by a factor of 8.

Furthermore, the maximum size of an MBR partition is 2TiB, so ...

2 TiB - 349.31 GiB = 1 698.69 GiB

http://www.google.com/search?q=2+TiB+-+349.31+GiB++in+GiB

That accounts for the first unallocated area. The remaining 746.52GiB is that area of the drive beyond 2TiB.

 

[sbpetrack]

Totally get it; thanks for the explanation and the calculations. I guess I was hoping that this situation was common enough that someone clever might have written a low-level disk driver that lies to the OS in the same way that the firmware does (ok, maybe "lies" is too extreme; if I understand, it would just need to ssssttttretch the truth (by a factor of 8).

I will mark this thread as SOLVED (if i can do that) and you last post as "the answer" (if I can do that).