[6.7.0] Disk unmountable [SOLVED]

Rebuilding a disk won't help with filesystem corruption, you need to check filesystem on disk17:

 

https://wiki.unraid.net/Check_Disk_Filesystems#Checking_and_fixing_drives_in_the_webGui

or

https://wiki.unraid.net/Check_Disk_Filesystems#Drives_formatted_with_XFS

 

Seems to be relatively common some xfs disks going unmountable after a kernel upgrade, it happened before and it's happening to some users now, likely newer kernel is detecting some previous undetected corruption.

Thanks, Johnny, that worked.

 

A xfs_repair -Ln was needed and the drive is ok again. However, during the repair it mentioned that it is moving inodes to lost+found but I was not able to find any folder with that name on the drive looking at it using the disk share.

 

It is sad to note that a file system corruption is not protected for by parity and can lead to data loss on a single drive failure.

If there's no lost+found folder then nothing was moved there, message always appears regardless.

 

Drive failure and filesystem corruption are two very different things, parity can't help with the latter, same as protecting against accidental deletions or ransomware, that's what backups are for.

 

Checking all other drives now. Noticed that the message: "- found root inode chunk" appears always. Probably normal though "chunk" sounds negative. The scan, regardless of -n or -v is added only ends with "done" but doesn't state clearly that errors were found or not. I guess when nothing in-between is mentioned that says "error" it is ok. But wondering if there is a better way of telling if everything is ok nor not.

1 hour ago, tazman said:

The scan, regardless of -n or -v is added only ends with "done" but doesn't state clearly that errors were found or not.

With xfs_repair -n, when there's nothing obvious in the output, the only way to know if errors were detected is to check the exit status, 0 means no errors detected, 1 means errors detected

 

 

Edited May 21, 2019 by johnnie.black