Re: [Bug-ddrescue] ddrescue 1.3 - questions from a newbie

[Matt Boge]

Thanks Ariel,
 
Yea, I've pretty much accepted the idea that this will not be a quick and easy 
process and I'm OK with that. Fortunately, I've got a secondary computer I can 
use for this purpose and am willing to be diligent and patient.
 
When I first ran ddrescue, I mistakenly copied the partition to a 
non-partitioned disk:
 
  ddrescue -B -n /dev/hda1 /dev/hdb rescued.log
 
(By the way, I know I should have copied it to /dev/hdb1, but should I have 
created that partition with a tool like cfdisk first or would including that in 
the ddrescue command have created the partition itself?)
 
Anyway, to capture the MBR, I then ran (with some helpful guidance from here):
 
  ddrescue -B -n -s 63 /dev/hda rescued_mbr.ima rescued_mbr.log
 
So, I've got a copy of my MBR in as an image om my USB drive, a rescued copy of 
my damaged partition on a new, unpartitioned drive (/dev/hdb).  Based on some 
more help here, I created a new partition on another new drive (/dev/sda1) with 
the IDENTICAL specs as the failed one (I used cfdisk to do this).  I then ran 
the following two commands:
 
  ddrescue -B rescued_mbr.ima /dev/sda1
  ddrescue -B /dev/hdb /dev/sda1
 
Does that look right to you? Is that all I really need to do to /dev/sda1?  Is 
it now what it would have looked like if I had run ddrescue correctly the first 
time?  Is this my "best-good-copy" now and should I duplicate this back to my 
/dev/hdb drive, overwriting the unpartitioned data from the first ddrescue run?
 
Also, when I rerun ddrescue with autosplit turned on can I run it to the new 
partition (i.e. /dev/sda1) or will the ddrescue log file be off by 63 sectors?
 
I'm sorry this is all so confusing, but I'm just not sure how much of a mistake 
I made originally by not creating that partition on /dev/hdb/ and if the 
relatively simple process above of merging the MBR and rescued data to a new 
partition on my /dev/sda drive was all I needed to do to correct that mistake.
 
Really, thanks again for your time and patience dealing with me on this... I'm 
sure you have much better things to do than to hand-hold a Linux newbie.  But 
you all have been such a lot of help and really appreciate it.
 
- Matt
 
 
 

----- Original Message ----
From: Ariel <address@hidden>
To: Matt Boge <address@hidden>
Cc: address@hidden; David Burton <address@hidden>
Sent: Thursday, January 25, 2007 5:29:04 PM
Subject: Re: [Bug-ddrescue] ddrescue 1.3 - questions from a newbie


On Thu, 25 Jan 2007, Matt Boge wrote:

> Hey guys,

> I don't want to be a pest, but if there is any chance someone could take 
> a few minutes and look at my questions below, I would definitely 
> appreciate it.  Again, I really appreciate what you have all done 
> already, but I'm kinda dead in the water right now and I'm a little 
> anxious about moving forward on my own.

> I have been checking out the rescued data using Knoppix, and while a lot 
> of data was saved, I noticed a huge glaring hole:  the My Pictures 
> folder was missing.  As you can probably imagine, this is one of the few 
> folders I really cared about rescuing and while I was expecting that 
> some of the data might be lost, I certainly wasn't expecting the entire 
> folder to not show up.

Seems that you had the disk failure where new writes fail, I guess you 
were writing a file to that directory?

Anyway the directory itself may be gone, but most likely all the files in 
it are still there. If you are using ntfs (not fat32) there is a change yu 
can save the file names, but otherwise you will lose the file names, but 
at least you will have the data.

Run checkdisk on it, and MAKE SURE you tell it to save lost clusters.

> Anyway, I think I'd like to give SpinRite a shot on the damaged drive to 
> see if I can eek out anymore data, but before I do (and possibly damage 
> the drive further), I want to be sure that the stuff I did pull off 
> already with ddrescue is safe.  In that vein, I want to have two copies 
> of the rescued data: one scratch drive (to add any potential new data 
> saved by SpinRite and updated by a second running of ddrescue) and one 
> "safe" copy that I can revert to if my scratch drive fiddlings go bad.

There is another reason to have a second disk: when you run checkdisk you 
may end up with less data then you would like, so if that happens you can 
revert back to your saved disk.

Also, before running spinrite, you may want to consider running ddrescue 
again, using the same logfile and settings etc (watch out if you moved the 
data because of your partition issue, I don't remember what you did).

It's possible ddrescue will be able to find more data - also this time 
turn on autosplit mode, I remember you had that off the first time.

And finally, after spinrite finishes, run ddrescue on the disk, again with 
the logfile - if spinrite managed to recover any bad sectors, then 
ddrescue will copy them, but it won't bother copying data that it already 
managed to get.

BTW: The first time I had a disk error it was 3weeks to a month before I 
was up and running again. The second time was quicker since I had raid 
(although it didn't help that I had 2 disks fail at the same time, and 
that Maxtor sent me a defective disk as a warranty replacement).

    -Ariel