Barracuda XT, Barracuda, Barracuda Green, DiamondMax and other desktop SATA and IDE drives

[reason-8]

I have a Barracuda 7200.7 ST3200822AS (internal SATA) that I've been using for about 6 years as my main boot drive with four NTFS partitions on it. When it started failing I added a new drive with a fresh OS install to diagnose, repair and/or retrieve as much of the data as I could. I also have a 1 TB SATA drive that I use for storage.

The problem drive shows up in the BIOS and in Windows, but trying to access files (or run chkdsk) on it causes it to freeze, throw a "The device is not ready" error, then disappears from Windows. Occasionally it will fail to be seen by the BIOS, but this is rare. There's no audible clicking or any indication that something is physically wrong with it.

The drive is connected via an onboard Promise Fasttrak 376 controller, though the two SATA drives not in a RAID array, they are just two separate disks. I've tried swapping the SATA ribbons and ports (primary/secondary) but the 1TB drive always works and the Seagate always fails.

SeaTools for DOS reports error code 96DECC55

I was able to run the Long DST once and it successfully repaired about 24 sectors.

If I try the Long DST now, it will usually crash to a bright green screen before it finishes. The green screen shows a command line interface (on top of the SeaTools GUI) and is usually functional for a while before it freezes and the motherboards starts beeping continuously. Sometimes it can finish the Long DST and start repairing before it crashes.

I was able to boot up Windows in Safe Mode with Command Prompt and managed to robocopy all the files from partitions 3 and 4 to another drive. Whenever I try this with partitions 1 or 2, I can load them, switch to them, view and navigate the file system a bit (via DIR and CD), but robocopy always fails on the first file or directory copy, with the "Device is not ready" error.

At this point, I'm trying to figure out the best way to try to rescue the data. Do any of you have any advice for this? I'm tempted to boot up linux to try to create an image of the problem partitions, using either ntfsclone, ddrescue or something similar. Is there anyone with experience with sector-level or raw disk copying that can recommend something?

[fzabkar]

I would do exactly as you are doing. That is, I would clone the problem partitions using ddrescue. I would determine the starting and ending LBAs, and then clone the damaged area in reverse. This effectively disables look ahead buffering. Ddrescue understands how to skip over bad sectors, so you could configure it to clone the easy sectors on the first pass, and the more difficult ones on subsequent passes.

As for your SeaTools error code, each code is specific to a particular serial number. However, the second last digit identifies the error type.

See http://forums.seagate.com/t5/SeaTools/SeaTools-Test-Code-96A5D1E9/m-p/66946#M807

[reason-8]

Thanks for the advice. I tried using both dd_rescue and GNU ddrescue but was not successful. The drive seems to be in such a state now that it fails immediately upon any read, whether I read in reverse, or in direct mode. It is still recognized by the BIOS and by 'fdisk -l', but even if I try to read a single block (e.g. the MBR) it will fail, and once it fails, it never recovers. I haven't tried mounting a partition first, but I don't think that would help.

Interestingly, if I boot into DOS (or Windows XP command line), I can still change to one of its drive letters, and if I run DIR, it will return the first file/directory in the list before failing with 'Device not ready'. Unfortunately this happens on all of the partitions, even the ones I was able to read before. Still, this seems to indicate that it can read something, right? Why would I not be able to read any bytes with dd, but be able to get some filesystem info with DIR?

Do I have options left for this drive or is it completely dead? Does this sound like a mechanical problem that can be repaired or is it something else?

[fzabkar]

From reading the HDD Guru forums, my understanding is that Seagate drives develop weak heads. When software imaging fails, the preferred method of data recovery involves hardware imagers such as DeepSpar, Atola, or Data Compass. These can control the power supply to the drive, and can image by head rather than by LBA. This allows a weak head to be avoided. A subsequent head swap can then recover the remaining data surface. I've read that it may sometimes take several weeks for a drive to be imaged in this way.