I have a hard drive which has suddenly developed unstable sectors.

I am able to read it with dd_rescue, so I transferred it completely to another new drive of the same size.

The Windows partition is bootable after the transfer, however, the Mac partition behaves weird.

When I boot it for the first time, it boots just fine, but forces an FSCK on the next boot. The FSCK however fails.

If I boot in single-user mode and forcefully tell FSCK to rebuild the FS, then the following happens:

  • A lot of ‘invalid node’ errors appear
  • fsck restarts a couple of times
  • after one of the iterations it’s getting abort()’ed
  • if I run it once again my screen is filled with ‘Node unrecoverable’ errors
  • afterwards if I try to continue booting, it tells me to ** REBOOT NOW **
  • if I obey and reboot, the partition is rendered unbootable, in case I try to mount it while booting in single-user from the failing drive I get an error of being unable to find the root in the catalog

I am currently running the dd_rescue procedure the second time (and it will probably take a week again), but can I somehow forcefully mark the partition as clean? From the FSCK logs I saw that the damaged files are some of the drivers (kexts) I don’t use or calendar files from 2013 which I couldn’t care less about. Maybe somehow deleting them might work?

I don’t have any third drive of the same size to save just an image of the whole thing because they are too expensive :/

Any help is appreciated. Thanks in advance!

  • If there are such difficulties in fsck, the problem is not only on the content of the files, but in the content of the filesystem. Every write on such file system could cause catastrophic damages. I recommend to format the partition, and to copy (from a read-only partition) the files. [a "live/rescue" distribution would help]. – Giacomo Catenazzi Oct 18 '16 at 14:20
  • Boot a live CD/OS, and copy as many files off the drive as you can now. I'm reading between the lines that you have no backup? – Criggie Oct 23 '16 at 6:16

I interrupted the second dd session after about 100Gb in the disk. Then I booted off an external OSX drive with DiskWarrior lent from a friend. From there I got a list of the overlapping files which were mostly cache files, so I went ahead and deleted them from the terminal. Then let DW rebuild the disk directory. Afterwards I restored the files from a backup or a last-minute read from the failing drive. This helped resolve the issue and all files are now intact.

  • OK good work. I assume you have disposed of the failing drive instead of re-using it? – Criggie Oct 23 '16 at 6:16
  • This is also an excellent opportunity to set yourself up backups that work. You could extend that by creating a RAID for your critical data as well as a backup. – Criggie Oct 23 '16 at 6:17
  • I put it into a secondary slot, zero-filled it (which seems to have eliminated the unstable sectors) and pointed my torrents and etc (read: stuff that I couldn't care losing in case anything) to it. That would also lift the load from my primary drive as well. – Vladislav Korotnev Oct 23 '16 at 6:28
  • I strongly recommend you throw the failing drive. Even if its "just stuff I don't care about" AND set up an autonomous automatic backup. You've had a warning from your gear, and that's uncommon. Learn from this warning, and set up a backup. – Criggie Oct 23 '16 at 6:32

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.