Copying a messed up directory off a windows drive
|
| Author | Content |
|---|---|
| techiem2 Apr 12, 2010 2:02 PM EDT |
I've got a laptop here with a nicely crashed hard disk.
I generally have good success copying the readable data off with mc, since it just errors when it gets to a file it can't fully read, and then continues on when you acknowledge.
However this particular machine seems to have a corrupted file entry in one of the directories.
It will start copying the stuff from the directory, and then hit a certain file, and keep trying to copy that file indefinitely (i.e. it tries, fails, tries again, fails, etc.), thus never getting past it. Is there another way to: recursively copy the directory structure don't follow symlinks (it's a Vista drive...so has all sorts of lovely recursive symlinks) ignore the file that is causing the problems fail files that can't be totally read but keep going I tried rsync, but that never started copying. It just sat preparing forever (i.e. it probably hit that file and was trying to read it...forever...) Thanks! On a side note, has anyone else noticed a rash of dying/crashing hard disks lately? I think I've had to deal with at least 5 or so in the past few months for people.... |
| jdixon Apr 12, 2010 2:16 PM EDT |
If you know what file it is, why not just move that file out of the directory? That may not work either, but it's worth a try. > ...has anyone else noticed a rash of dying/crashing hard disks lately? Not particularly. I've had a raft of power supply failures lately though (3 in the past 3 months). Probably just coincidence in both cases, unless your's are from the same manufacturer. |
| techiem2 Apr 12, 2010 3:53 PM EDT |
When I mount the partition ntfs-3g, I can't even access the directory.
I get I/O error.
When I mount it normally (i.e. readonly default kernel access mode), normal access commands like ls sit indefinitely. So at this point mc is the only think that has had any success whatsoever reading it. I guess maybe the rest of that directory is just lost. Hopefully the new data he mentioned isn't in that part of the directory. heh. I've tried running fsck.ntfs on it a few times, but that hasn't helped any. |
| TxtEdMacs Apr 12, 2010 4:18 PM EDT |
Quoting: [...] has anyone else noticed a rash of dying/crashing hard disks lately?[serious] Yes my son told me yesterday that Samsung, Western Digital and Seagate have about a 3%, 9% and 43%* failure rate**. This is shocking, at one time Seagate was the best choice in hard drives. My son's SOP is to load a limited set of data on a new drive where the loss would not be a problem. Moreover, he does this in advance of the actual need for a new drive. His experience has been if it survives a month it will probably be long lived. * May be understated, since I thought he said over half at one point. ** That may not include DOA (Dead On Arrival) failures. [/serious] YBT |
| tuxchick Apr 12, 2010 4:23 PM EDT |
Have you tried GNU ddrescue command? It is designed to copy data off failing drives.
http://www.gnu.org/software/ddrescue/ddrescue.html It is a great little app that can handle bad blocks without freaking out. It is on SystemRescue CD/USB, my fave rescue distro, and in most repos. This is different from dd_rescue, which is similar, but I think ddrescue is faster and better. |
| techiem2 Apr 12, 2010 4:24 PM EDT |
Interesting. The ones I've seen are varying brands/varying ages/etc. Old machines, newer machines, laptops, desktops, netbooks.... It's just been weird. |
You cannot post until you login.