Walt Mankowski on 18 Jul 2015 18:36:13 -0700

[Date Prev] [Date Next] [Thread Prev] [Thread Next] [Date Index] [Thread Index]

Re: [PLUG] Cloning a dying hard drive

On Sat, Jul 18, 2015 at 08:39:45PM -0400, brent timothy saner wrote:
> Second though: if you have the space handy, it's actually better to
> first *image a drive* if it's failing, and then try cloning from the
> image. You can use the I/O loopback module and kpartx to interact with
> the image.

Maybe if I'd gotten a 2 TB drive... Otherwise, I don't have space to
image the entire drive.

> On that note, instead of dd with noerror,sync I'd instead recommend you
> use GNU ddrescue (*not* dd_rescue- it's a similar name, and they aim to
> do the same thing, but GNU does it fairly better- it's a bit more
> intelligent when it does encounter errors.
> See https://wiki.archlinux.org/index.php/File_recovery#Failing_drives
> for more info (I know I'm repeating myself, but I love the Arch wiki).

ddrescue looks like the way to go, thanks.

> I would avoid mounting the disk at all, to be honest, if at all
> possible. The SystemRescueCD[0], on that note, has ddrescue (along with
> testdisk and photorec and i BELIEVE foremost, sotware that may help with
> recovery from troublesome sectors).
> Or I can get you a download link for a build of my personal project,
> BDisk[1].

Thanks, that answers what was going to be my next question. :)
Running SystemRescueCD sounds safer than trying to do this from single
user mode.  I wonder if my ancient box will boot from a USB stick.  Or
if my CD burner still works...

> > * full restore from my backups
> > 
> Oh, you have backups! That's really, really good. You get a cookie. I'm
> not being sarcastic; too many people don't keep backups. Are they
> relatively recent? What does the payoff of doing a dd/ddrescue +
> filesystem/data corruption repair vs. an OS install + backup restore ( -
> missing data from backup) cost relative to each other? This is something
> only you can answer, as it all depends how much time you want to wait
> before having a box back up again.
> If the backups were relatively recent (no more than a day or two), I'd
> go that route.

This ain't my first rodeo.  I run nightly backups using rdiff-backup.
That's how I first noticed the errors.

That said, it's on a slow USB connection and it would take a long time
to do a full restore.  Also, this is like a 10 year old Debian testing
box.  I really don't want to deal with reinstalling the entire OS.
And as I said in my original post, I want to get a new server anyway.
I'd just like to limp along with this old box until I can get the new
one up and running.


Attachment: signature.asc
Description: Digital signature

Philadelphia Linux Users Group         --        http://www.phillylinux.org
Announcements - http://lists.phillylinux.org/mailman/listinfo/plug-announce
General Discussion  --   http://lists.phillylinux.org/mailman/listinfo/plug