2

I came into work this morning and one of my BSD servers is unable to boot properly. I can get into single user mode but when running fsck I get the error "unexpected soft update inconsistency" Does anyone know either how to fix this or how I should proceed next? Thanks!

enter image description here enter image description here

The specific problem is that I can't access my system properly. I am unsure on what the error stated above means and how to fix it. I uploaded the images of the fsck -y per the request below. Any help would be very much appreciated.

UPDATE The system came back up after a few hours of running fsck -y!! I am able to login in multiuser mode and all is good. I am going to put in a recommendation to my boss that everything be backed up. Thanks for all of your help and input everyone.

7
  • 2
    What about fsck -y? Please paste the full output (or at least, some error message).
    – Nathan C
    Jul 10, 2013 at 17:41
  • I added the images of fsck -y so far.
    – Fr0ntSight
    Jul 10, 2013 at 19:03
  • did you lose power to the server?
    – Doon
    Jul 10, 2013 at 19:17
  • Yes, there was a power outage here on Sunday.
    – Fr0ntSight
    Jul 10, 2013 at 19:17
  • 1
    @Fr0ntSight That is sadly an all too common situation in small offices. I suggest a UPS that has a supported companion tool in ports so you can shut down the server automatically if the UPS is in danger of running its battery down (There's an old FreeBSD Diary article about doing this with APC hardware, and there are a bunch of other options if you google around or search the ports collection)
    – voretaq7
    Jul 10, 2013 at 19:33

2 Answers 2

4

As the next three words of the error message imply, the right thing to do is run fsck manually (which you're doing).

fsck (usually run with -y so you don't have to manually say yes to all the prompts) will resolve the soft update inconsistencies and usually leave you with a working system again.
(Exceptions exist where whatever caused the inconsistency has trashed something important, like the kernel, or where the underlying cause is a dying disk.)

If fsck can't do the job you're in a position where you probably want to restore your system from known good backups. (The brave among us, and the few with skills in the arcana of fsdb may elect to attempt a manual resurrection, but restoring from backups is almost always less painful.)

2
  • Thank you voretaq. I really appreciate your response. I will continue to let fsck run and do its thing. At first I thought it was hanging but after reading your answer, I will let it be and hope it fixes the disk. That was the first time I saw that specific error message and couldn't pin down what it meant. Thanks again.
    – Fr0ntSight
    Jul 10, 2013 at 19:20
  • depending upon the size/type of disk this can take a bit.
    – Doon
    Jul 10, 2013 at 19:40
3

fsck -y will force fsck to "fix" the errors. Considering you have superblock, bit map, and summary errors the disk is likely fried or something truly horrible happened. In any case, backup whatever you can off the disk and consider reinstalling/reimaging.

Looks like you're running a SATA disk. If you can get the system to boot you should check the SMART status with the 'smartmontools' port/package. It's pretty likely to have some additional information.

2
  • I've seen nasty like that on a heavily-trafficed DB server that took a power hit -- the disk might be OK (though if there's any reason to suspect disk problems definitely back up and replace the hardware).
    – voretaq7
    Jul 10, 2013 at 19:16
  • I've not, at least not in the last few years - Anytime I get multiple metadata errors the file system is cooked and if it wasn't something stupid I did then the disk was usually making weird noises.
    – Chris S
    Jul 10, 2013 at 19:21

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .