26

I first noticed that fsck_hfs was running, taking up 50-75% of a CPU, yesterday. It continues to run today.

ps shows that it is doing /sbin/fsck_hfs -f -n -x -E /dev/disk3. Only problem: I don’t think I have a /dev/disk3.

  1. Why is it running?
  2. Will it ever finish?
    • Can I kill it?
  3. What is /dev/disk3? Could it be my Time Machine volume, which is not mounted at the moment?

System Info: MacBook Pro (2008). It has two disks installed—the internal disk (/dev/disk1) and a PC Card SSD (/dev/disk0, surprisingly). It connects to a remote Time Machine volume attached to an Airport Extreme base station.

  • 1
    another datapoint; my 10.13.5 laptop just spent two hours "verifying" a backup on a NAS drive, apparently stuck at 84%. Majority of the CPU time was going to fsck_hfs (similar command line to the OP) with basically zero network traffic (~5 packets per second) and I think disk IO Ops/s were low as well but I wasn't checking that as much. No apparent change in progress was reported in the GUI (that I noticed) and then it just completed successfully. – Sam Mason Jul 3 '18 at 16:47
19
  1. It was probably running from when your Time Machine volume was mounted.
  2. If the volume isn't present anymore, I doubt it.
    • I'm sure you can sudo killall fsck_hfs; it wouldn't hurt anything. (Have you tried restarting?)
  3. It probably is.
  • Thanks for the help! You were right. Moments after killing it, Time Machine popped up saying it was unable to verify my backup. It probably became confused last night when I closed my MacBook and took it home. Crazy that it will try to verify a ~750 GB disk image over Wi-Fi! I would have to leave my computer turned on and at the office for several days for it to complete… – Nate Jun 17 '10 at 19:30
  • 1
    You probably don't want to force quit it because it forces you to rebackup your entire computer and you then loose the existing back up history – Sirens Sep 1 '15 at 2:19
  • 1
    @Nate "… it will try to verify a ~750 GB disk image over Wi-Fi! …" – that's not what happens; only a fraction of the image is transmitted. Keyword: CopyHFSMeta, please see for example forums.freenas.org/index.php?threads/… – Graham Perrin Jan 30 '16 at 12:04
26

Clicking on the Time Machine icon in the menu bar and choose "Skip verification" caused the fsck_hfs process to stop itself. Maybe a bit nicer than kill...

  • Just let it run, just wait for it finish. – Zhang Buzz Jun 29 at 2:00
3

You can also click on the Time Machine icon, select "Open Time Machine Preferences...", and you should see a progress bar for how far along the "Backing Up" process is.

You can then click the "x" next to the progress bar to stop the verification process.

1

I would recommend against killing it.

Last time I did, it left my Time Machine backup in a state where OSX believed it to be corrupted. This resulted in OSX refusing to continue using this backup, i.e. OSX queried me to create a new backup (which would effectively mean I would loose any history in my backup, e.g. deleted files, previous versions of modified files, etc.)

I managed to bring the Time Machine backup into a good state, allowing OSX to continue using this backup, but it was quite a hassle.

  • Care to explain your process getting the backup back into a good state? I have an empty sparsebundle on my Time Machine now. – Matt M. Apr 28 '17 at 4:37
  • @MattM. If your sparsebundle is truly empty, I don't think you can save it. But if it's not empty, just marked as corrupt as mine was, it should be salvageable. I can't remember exactly what I did, found the solution through a google search. But the sparsebundle contains some metadata file that has a property telling that the backup is being checked/corrupt. So the process was something like unmount the time machine volume, find the metadata file, adjust those properties, and possibly remount the time machine volume (maybe I manually ran a verification step afterwards) – Pete Apr 28 '17 at 10:24
  • I used Data Rescue 4 for Mac, and it worked like a champ for $99. I didn't want to risk any irreversible loss by editing the metadata. It restored the data directly onto my local drive, reading the sparsebundle from Time Capsule connected over ethernet. – Matt M. May 2 '17 at 8:33

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.