[TriLUG] Hard drive recovery

[Kevin Faulkner via TriLUG]

On Sat, 13 Apr 2019 16:49:35 +0000 (GMT)
Joseph Mack NA3T via TriLUG <trilug at trilug.org> wrote:

> On Sat, 13 Apr 2019, shay walters via TriLUG wrote:
> 
> >    I've seen some videos of opening up a drive that had the head
> > stuck to the platter, and manually rotating the platter while
> > retracting the head, but once you break that seal, you're just
> > asking for trouble, unless you have some sort of clean environment
> > to do the work in.  
> 
> If you have no other option, it's worth it to try to get the data
> off. You only need to access the drive once. Hopefully any trouble
> doesn't start till later.
> 
> Apparently taking the lid off a drive isn't as traumatic as I
> thought. I once offered to backup someone's laptop drive, a process
> which involved taking the drive out and dd'ing it on another machine.
> I told him to take the drive out and give it to me. He assured me he
> could handle it. He handed me the drive with the lid off (platter and
> heads visible) saying "here you are". I didn't let him know that I
> thought he'd stewed his (un backed-up) drive. I had thought that the
> lid had a bearing for the top end of the spindle, but there was no
> bearing in the lid. The lid was just a cover. It seems that the drive
> is only supported at the bottom (at least for laptop drives). He put
> the lid back on, screwed it shut, we backed it up and it's been
> running ever since.
> 
> To the OP; if you get the drive running, back it up with ddrescue
> asap.
> 
> https://www.linux.com/learn/intro-to-linux/2017/3/gnu-ddrescue-best-damaged-drive-rescue
> 
> make sure you use a logfile so the second run of ddrescue only goes
> to the unrecovered sectors.
> 
> Joe
> 
If the drive is not initializing I would first look at the PCB first
instead of a head failure. Does the drive respond to any ATA
commands/queries? Perhaps look at smartctl/smartmontools to help
diagnose, since it isn't attempting to mount the drive you might have
more luck in diagnosing the actual issue. I used to work at a drive
repair shop in Phoenix once I got out of the army, I was doing Linux and
alternative filesystem recovery (riesier and QNX like CNC machines).
While it was interesting it wasn't where my heart was. Here are a few
points that I can remember:

If the spindle motor is bad, that will make the
repair much more difficult. The method I saw that worked fairly well
was to tape around the circumference of the drive platters keeping them
aligned then placing them in a working same model.

You can try to replace the PCB, I don't think it can get much worse. To
be honest I forget/don't know the ramifications of a PCB mismatch I
really don't believe they are serious, I think mismatched one and it
still didn't initialize. It sounds like the drive isn't initializing.
What you could consider doing is going on to PCB replacement sites or
places that sell the full drive and if that price is still too high
likely a full service is going to be too expensive too.
https://www.hdd-parts.com/seagate-barracuda-7200-10-sata.html
When looking for the replacement, pay attention to the site code and
firmware revision. I have a spare seagate 4tb sitting here and it says:

:: 

  Site: TK
  DOM: 15JUL2017
  FW: 0001
  PN: 2CV104-300

Just to let you know better what the site code and firmware revision
is exactly. The closer you can get the better. Again this might be to
just price things out because if it's a PCB issue, either the recovery
service has a PCB on hand (from say a file system recovery or a drive
that bad a ton of bad sectors, thus the PCB would still be ok) or
they'll be doing the same thing (going to a service to buy a specific
PCB).

If you (or others) are curious, there was a tool that we had called `PC
3K <https://www.acelaboratory.com/pc3000.udma.php>` It could do a
number of different cool things like empty out the bad sector map list,
or perform low level service commands and tell the drive to do specific
functions like calibrate the drive arm or something. I thought when
looking at it work, some of these commands had originated around
smart commands and maybe some leaked vendor information. Hard to say,
but it was really neat.

I agree with Joe, about, once the drive is up and you have a copy, 
keep it going no matter what. Here ddrescue and similar programs might
help. I have heard that others had good experience putting the drive in
the freezer. I (and others I worked with at the time) found that the
contrary (warmer drive) performed better. Hard to get actual data on
that since every failure is different. If I recall ddrescue was nice
because it would jump around trying to get as much as it could until it
ran into an error. There are many like them now, and I haven't looked
at them in a long time.

This is just my experience I didn't do this for years, and my
experience is now somewhat out of date too. Take it with a grain of
salt, but I wanted to at least offer some guidance. If I can be of
further help please let me know

-Kevin
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 488 bytes
Desc: OpenPGP digital signature
URL: <http://www.trilug.org/pipermail/trilug/attachments/20190413/a5dc26e4/attachment.pgp>