home | list info | list archive | date index | thread index

Re: [OCLUG-Tech] Help recovering a (software-dead, not hardware dead) hard-drive

  • Subject: Re: [OCLUG-Tech] Help recovering a (software-dead, not hardware dead) hard-drive
  • From: Bill Strosberg <oclug [ at ] strosberg [ dot ] com>
  • Date: Wed, 27 Feb 2013 08:48:49 -0500
On 13-02-27 08:21 AM, ed stuckems wrote:
My test drive is a 1T drive that's connected to the computer via usb
and I'm trying to format the drive so that it'll boot in a (u)efi
machine.  The last thing I did with the drive was to try and create
one large partition formatted with a fat32 filesystem.  When I last
used the machine and drive, I was still able to mount, umount, and
access the drive with parted but when I turned off the machine and
rebooted, the device wouldn't attach to a device file in /dev thus
preventing me from either mounting the drive or using parted/fdisk to
repartition it.  If I were to ask a specific question it would be "How
to I coerce linux into connecting the drive to a device (in /dev) so
that I can run something like fdisk or parted on the device be able to
create a filesystem with mkfs"?

Ed:

There really isn't enough information here to provide perfect instructions on how to proceed. I'm tempted to say all the obvious things about disaster prep here and mitigating risk but won't.

If the data on the drive isn't important:

Each manufacturer of drive typically has website available bootable disk images that have their own particular low level format, testing and drive recovery software. If the drive is in warranty, they generally require you run this software to generate a failure report identifying the failure. In the process of testing and low level formatting a drive, it quite often will cleanly reformat and become functional again. That being said, the test and format process usually ends up marking bad disk sectors and formatting around them. I'm pretty leery about using a drive like this for anything I care about. Replacement drives you get on warranty exchange are usually just recovered and reformatted (remanufactured?) field failures.

Disks all have a lifetime, and once errors start creeping in, I generally just replace the drive. It isn't worth burning $150 in billable time to recover a used and suspect $150 drive. If I do that I'm down $300 and only get a undependable used drive I can't trust out of the deal.

Now that the flood plains in Thailand have dried out and the drive manufacturers have stuffed the supply chain, drive prices are dirt cheap once again.

If you want to get data off the drive:

If you've got time to spare and losing billable hours isn't a problem GNU ddrescue is pretty amazing at bringing dead or dying drives back to consciousness, even if only long enough to get a backup off.

--
Bill