Partially dead drive

[D. Hugh Redelmeier]

| From: Paul King <sciguy-Ja3L+HSX0kI at public.gmane.org>

| || I have a 1TB Seagate drive which, I admit, was partitioned when the NT 
| || drive was less than half full, and I didn't defrag the data to the 
| || start of the drive.
| 
| | Why do you have to "admit" that?  Are there any consequences?
| 
| I felt I needed to mention that because I may have set a partition boundary
| during resizing in a way that closed off a file or cut it off. I could be
| wrong.

How did you resize?  I hope it wasn't just fdisking to set another 
boundary.

Programs meant to do resizing must understand the filesystem in question 
and manipulate it to allow resizing.  gparted does this, for example.  
Partition Magic was the first popular commercial tool to do this.  The 
task is tricky enough that PM was considered, umm, magical when it first 
came out.

If you used a suitable tool, you would not have cut off a file.
Unless it was buggy: then all bets are off.

| FWIW, EMusic no longer allowed me to re-download all of my purchased music
| (they used to).

Wow.  But on the other hand, it wasn't DRMed, if I remember correctly.
What I really hate is a DRMed platform disappearing (eg. MS Plays For
Sure -- what an inappropriate title!).

I always thought that an argument for the cloud was that they were
safer than leaving things in your own hands.

| | <Superstitious mode> When I've used Linux tools to resize a Windows 
| | boot partition, I've found it important to immediately boot Windows 
| | and do a repair.  This is before installing Linux.  I think that 
| | the resizing might be moving unmoveable files and the Windows 
| | repair tools can fix that. </superstitious mode>
| 
| The hard drive in question was for data. It wasn't a system drive.
| <snip>

It might still have unmoveable files -- I don't know enough about Windows 
to say for sure.

Hmm...  At one time suspend-to-disk was managed by the BIOS and it had to 
know where to write stuff.  It wrote to the Windows swap file (or a 
special partition).  If the swapfile was moved or deleted, the BIOS needed 
to be "told".  I *think* that this mechanism disappeared with APM.  Not 
sure.

| The errors happened suddenly. It was simply that, one day, I could not
| reboot the machine. And this was because Grub was on the failed drive.
| 
| <snip - but all good reasons for doing backups>

No, not doing backups, doing recovery carefully.  It's not too late!
Take an image!

I know that it is pointless to ask someone to get in their time
machine and do a backup in the past.

| But since I suspected that reducing a partition size and creating new
| partitions was going to "do something" unstable, I probably should have 
| made a backup before doing that.

I often takes my chances.  Only once was that a problem.  Vista became
unbootable.  So I had to buy restore CDs from Acer (the machine had a
1 year warranty but apparently installation problems were only covered
for 90 days).  Later I learned that the bundled a Window upgrade
disk could do the repairs (thanks, Lennart!).

| Thanks for your valuable advice, Hugh. This can prove helpful, although 
| things are looking pretty dim at this point.

Have you tried any of the advice?  To re-iterate:

- start by grabbing an image using GNU ddrescue.

- For a Linux filesystem, you can loopback mount an image file and
  hack on that.

- For ntfs, I'd recommend plopping the image, raw, as the contents of
  a partition and whacking on it in the new place.  The new partition
  doesn't have to be exactly the same size as the original: it can be
  larger.

- I pointed at an article which talked about programs that can root
  through a busted filesystem for valuables.

- I also mentioned Smartmon Tools to know disk hardware status.

- I forgot to mention the the output of dmesg might give some clues
  about I/O errors.
--
The Toronto Linux Users Group.      Meetings: http://gtalug.org/
TLUG requests: Linux topics, No HTML, wrap text below 80 columns
How to UNSUBSCRIBE: http://gtalug.org/wiki/Mailing_lists