Take the 2-minute tour ×
Ask Ubuntu is a question and answer site for Ubuntu users and developers. It's 100% free, no registration required.

There are other questions like this bumbling around the site (that I'm working through) but I'm in the strange position where I think I know what I'm doing but can't explain what's going on.

  • I had 10x ~10GB files. None of them were open. Previews in Dolphin are turned off for files that big. Confirmed with lsof.
  • I selected the files and pressed Shift+Delete. This should skip the trash and nuke the files.
  • No progress was recorded. Usually Dolphin/KDE make a big fuss about deleting things.
  • The files are "gone" (not showing in Dolphin or the command line)
  • I'm still 100GB heavier than I should be. Most importantly, I'm at 100% disk use, which is why this is an urgent issue.
  • I've checked .trash on that drive and they weren't in there.
  • I've checked lsof again. They're not open.

So it seems like the inodes were deleted but the space wasn't freed. I would assume I have 100GB of orphan files. Hooray.

Is there any way I can purge these orphan files from an active filesystem (I realise a fsck would probably shift them off an offline filesystem)? I can't take the filesystem offline because important things (system files, home items, etc) are bind-mounted out of here.

Other details: the filesystem is EXT3 (I know, I should probably upgrade but changing FS is scaring when you're dealing with 4TB of data). The device is a mdadm RAID5 array.

share|improve this question
1  
Did u try baobab to find the files? –  blade19899 Nov 26 '13 at 12:50
    
I didn't think of that. Will report back. As I say, there's 4TB of data in here so it's going to take a little while to scan. –  Oli Nov 26 '13 at 12:53
    
Annoyingly, as I've just moved another 10GB of stuff off the disk and there's now 60GB free. I think it's possible it's just being really slow. –  Oli Nov 26 '13 at 12:55

1 Answer 1

up vote 3 down vote accepted

Ahh, deleting large files on ext3 takes a while since it has to zero out a lot of indirect blocks ( though I could have sworn the rm command blocks until it is done ). Yes, you certainly should upgrade to ext4. It shouldn't be scary when you have a backup, and you do have a backup right? ;)

Don't forget; raid is not for preventing data loss, it is for preventing downtime due to mechanical failure.

share|improve this answer
    
Yeah most of it is backed up but just reinitialising a RAID5 array of that size takes a day. Restoring that much data from slow off-site backup locations is another week. I've just decided to ignore it until I replace the array (which won't be long if I keep filling it up). –  Oli Nov 26 '13 at 14:44
1  
@Oli, if you are using lvm and can get the free space back up to 20% or so, you could shrink the volume, create a new ext4 one, migrate some files over, shrink some more, enlarge the new one, etc. –  psusi Nov 26 '13 at 14:47
    
@Oli, oh, and you can use tune2fs to enable extents without reformatting, which will give at least some of the benefits of ext4, including making deletes ( of newly created files after enabling the feature ) go faster. You can also use chattr to convert existing files to use extents. –  psusi Nov 27 '13 at 23:06

Your Answer

 
discard

By posting your answer, you agree to the privacy policy and terms of service.

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