[PLUG] Failed Go-Flex hard drive

[Dale Snell]

On Sun, 27 Jan 2013 17:50:09 -0800
John Jason Jordan <johnxj at comcast.net> wrote:

> I have a Seagate 3TB Go-Flex USB 3.0 hard drive that has apparently
> failed. It still spins up, the light comes on, but none of the GUI
> file managers on Fedora or Ubuntu can display the contents. They give
> me the error message "Failed to open directory 'Movies'. Error when
> getting information for /media/Movies/moviefilename."
> 
> The contents are old movies that I continue to seed. I noticed the
> other day that Ktorrent was complaining about a couple of them missing
> pieces, and I started worrying then. But otherwise it continued to
> function until just now when I returned home after taking my computer
> to PSU. The Go-Flex is connected to a USB port on the laptop dock, so
> as soon as I place the computer in the dock it is recognized. I have
> to mount it before I start Ktorrent, and now after mounting it appears
> mounted, but I can't read the contents. This happens on both my laptop
> and my desktop, using various USB ports, and after wiggling
> connections.
> 
> I opened Palimpsest, which reported the Smart information as "Disk has
> a few bad sectors." I unmounted it and ran "e2fsck -p /dev/sdb" which
> reported "Bad magic number in superblock while trying to
> open /dev/sdb: The superblock could not be read or does not describe
> a correct ext2 filesystem." (It is ext4.) The error message further
> suggested to run "e2fsck -b 8193 <device>." I have not tried this as
> I have no idea what this will do. 

The -b option to e2fsck tells it to use an alternate superblock,
this one at location 8193.  When an ext2/3/4 filesystem is
created, is builds several spare superblocks, just in case one
goes bad.  This sounds like such a situation.  In fact, I would do

    # e2fsck -b 8193 -fpDC 0 /dev/sdb

The options tell it to: use the alternate super(b)lock at 8193;
(f)orce the filesystem check, just in case the system thinks the
filesystem is clean; "(p)reen" the filesystem (i.e., correct what
it can without manual intervention); check and index the
(D)irectories; and (C) display a progress bar to tell you how far
it's gone.  The -D option rarely needs to be done, but in this
case, I think it might be a good idea.

You should check out the man pages for e2fsck(8), mkext2(8), and
tune2fs(8).  You might find them a bit dense, but it's good
information to know.  If you have questions, well, you know where
to ask.  :-)

I like to run fsck on all my filesystems every once in a while.
Since my desktop system is up 24/7, it's run automatically
whenever the system is booted.  Given that I only reboot when the
kernel is replaced, or some other major change, it doesn't happen
very often.  I do the same with my laptop, mostly because I don't
trust the thing farther than I could dropkick my house.  My USB
drives are not added at bootup, since I take them from place to
place (especially my music drive).  So I check them manually, once
every week or so.  The command I use is

    # e2fsck -fpC 0 /dev/sdn

where "n" is the drive number.  Very seldom does anything go
wrong, but every once in a great while, something does, as you've
just found out.

Hope this helps.

--Dale

--
Eye halve a spelling chequer           I ran this poem thru it
It came with my pea sea                I'm sure your pleased to no
It plainly marques four my revue       Its letter perfect in it's weigh
Miss steaks eye kin knot sea.          My checker told me sew.
    -- based on a poem by Mark Eckman