[TriLUG] Fwd: Drive recovery services

[David Burton]

---------- Forwarded message ----------
From: David Burton <ncdave4life at gmail.com>
Date: Mon, Oct 13, 2014 at 2:50 PM
Subject: Re: [TriLUG] Drive recovery services
To: Triangle Linux Users Group General Discussion <trilug at trilug.org>

You're welcome.

There's one other step that might possibly work for you. But don't try it
until *after* you have first tried ddrescue on all the drives, to rescue
whatever you can from them.

If you have a drive that is not recognized by the controller, and isn't
even spinning, and you have another drive of the same make and model (and
preferably same firmware version), you can try swapping printed circuit
boards. You might be able to run the dead drive long enough to recover the
lost data by temporarily substituting the good drive's PCB.

There's a chance that this process could go wrong, perhaps even destroy the
good drive's PCB, thus leaving you with two bad drives instead of one bad
and one good. So be sure to do the ddrescue data recovery step first!

Also, I've read that some drives keep surface defect mapping info about the
platters, or perhaps some other sort of platter-specific calibration info,
in flash storage on the drives' PCBs. I don't know how common that is, nor
even whether current model drives still do it. But if your drives do that
then I suspect that recovering data by swapping PCBs might result in some
mangled or misplaced sectors in the recovered data. YMMV.


One other thought: all those drives dying at once might not be coincidence.
You might have a bad power supply sending power spikes to the drives.

That doesn't happen very often with internal PSUs in PCs, but it can
happen. It happens more often with those crummy little external power
supplies which come with cheap SATA+IDE to USB adapters and cheap external
drive enclosures.

I have a nice WDC Caviar 1.5 TB Green drive that got fried by one of those
miserable pieces of junk. Fortunately, it turns out that it just blew an
overvoltage protection device (TVS diode), which crowbarred (shorted):
http://www.burtonsys.com/1.5TB_pcb_17pct_annotated.jpg
http://www.burtonsys.com/1.5TB_pcb_blown_D3_TVS_closeup.jpg
I cut it out of the circuit, and the drive is now working fine, albeit with
no overvoltage protection.

I've learned my lesson: I never use those little external PSUs that come
with the SATA+IDE USB adapters:
http://www.burtonsys.com/SATA_IDE_USB2_adapter.jpg
I just power the drives from a real PC power supply.

Dave



On Mon, Oct 13, 2014 at 11:58 AM, Wes Garrison <wes at xitechusa.com> wrote:

> That's an AWESOME post, thanks so much for the advice.
>
> At this point, I think 2 of the drives are so FUBAR that they aren't
> recognized by the controller, but I'll give it a shot.  I've already got a
> Parted Magic USB stick, so that's a step in the right direction.
>
> -Wes
>
> On Mon, Oct 13, 2014 at 11:49 AM, David Burton <ncdave4life at gmail.com>
> wrote:
>
> > Yes, pull the drives from the raid controller and connect them to
> > standalone machine(s) for the data recovery effort. But don't use
>
*...[snipped to get it under 20K]*



*Could we possibly get the message size limit increased, please?*


>
>
>
> *Your mail to 'TriLUG' with the subject    Re: [TriLUG] Drive recovery
> servicesIs being held until the list moderator can review it for
> approval.The reason it is being held:    Message body is too big: 24462
> bytes with a limit of 20 KB*