Last month I upgraded the Unraid OS from 6.2.4 to 6.6.5, then to 6.6.6 about a week later.  All seemed fine until I noticed some of my docker appdata was missing.  I didn't notice a problem with the cache pool until now mostly because all disk statuses were green.  I have an unmountable cache disk.

 

I've managed to mount the fs and copy the data I need.  Considering that, is there any harm in attempting a repair with this command?  Or am I better off to simply reformat?  I've never used it before and all refs say use as last resort.

btrfs check --repair <device>

  

The file system checker in the GUI reported no errors.  I'm kind of confused.  Anyone know of possible reasons why the disk got to this state?

 

 

 

If you already copied the data best bet is to reformat and restore, btrfs fsck isn't reliable yet, and even if it worked for your current problem, and without seeing the diags can't say either way, it would be more likely to get corrupt again in the future, so best to start fresh.

I tried the repair for kicks and although it reported the cache disk returned to normal operation; it wasn't.

19 hours ago, johnnie.black said:

best bet is to reformat and restore

I had a heck of a time trying to do this.  This is more or less the steps I took.   Is there an easier way? 

1. Stop array
2. Unassign both
3. Start array
4. Stop array
5. Assign #1 cache
6. Start array
7. Format #1 cache
8. Stop array
9. Assign #2 cache
10. Start array (data on #2 is wiped upon array start)

You could use wipefs on them, that would be the faster way.