Super User is a question and answer site for computer enthusiasts and power users. Join them; it only takes a minute:

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

I have a RAID 5 array using mdadm and ext4 as its filesystem. We got a strong power surge and one of the drive faulted, while another one got a bit flakky, aka "click of death".

I replaced the disks one by one, first changing the broken one, letting resync, and then the buggy one with a clean disk and another resync. I now have a full RAID 5 (3 drives + 1 parity drive) running fine.

My issue is that during those resyncs, some sectors were probably failed on the buggy drive and I now have some faulty inodes. Here is an exemple of what I get in the /var/log/messages (gentoo):

Aug 16 14:16:13 shiva kernel: EXT4-fs error (device md127): ext4_lookup: deleted inode referenced: 270795020

Most of the files works juste fine, there are only about ~12 such inodes failed like that and the files it correspond to seems to be mostly very small ones next to each other (the array is ~4TB large).

My issue now is that I want to cleanup my disk but I'm not sure how, the bootup tells me to run a fsck but I don't know what it does exactly and plenty of things I read on the web scare me a little :)

Considering that:

  1. the hardware issue is 100% resolved
  2. most of the filesystem works just fine, only a dozen inodes are screwed
  3. I want to "fix" the filesystem to get it back to normal
  4. I'm okay with losing the few failed files (although I would like to know which one it is if possible)
  5. again, the system boot/mount fine, almost every file works fine, the one that fails are always the few same

Should I be running fsck on the drive to do that ? Will I risk losing everything ?

I know the best way would be to backup and then redo, but I can't afford 3TB of backups right now after already just buying 2 new drives ... I'm trying to clean up the mess.

share|improve this question
up vote 2 down vote accepted

Yes, running fsck will do exactly what you want. Of course, anything can happen, including a meteorite landing on your computer and destroying all your disks; but if anything running fsck is safer than continuing to operate with a filesystems with errors.

I recommend booting from a rescue CD (or USB pen drive) and running fsck from there. If you're feeling uneasy, you can start by running fsck -n, which will not write to the filesystem, only tell you what errors it's found.

share|improve this answer
Thanks, that worked – azera Aug 16 '10 at 17:33

You must log in to answer this question.

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