I literally am jaw-dropped by this.

 

I have two unRAID servers which are mirrors of one another (thank god) and have a sync job which uses a commercial product GoodSync to sync between them, I just received an alert from it indicating more than 30% of the sync has changed and when I went and looked my media folder on my unRAID 6.4 box is 3.22 TB smaller than my 6.3.5 box. In looking at the share, the entire folder structure is there sans the files. The Main tab also reports the disk space is all in use. Dropping to the command line the files are not there in /mnt/disk or /mnt/user.

 

This is scary, the 6.4 box is btrfs, the 6.3.5 box is xfs. I am just thankfull I have a fully mirrored server. What information can I provide to troubleshoot this? I am currently running a btrfs scrub and all sync jobs have been suspended.

quazar-diagnostics-20180126-1303.zip

16 minutes ago, neurocis said:

The Main tab also reports the disk space is all in use.

Do you mean the used space is reported as if the files are still there?

 

I don't see nothing wrong on the syslog, can you post the output of
 

btrfs fi show /mnt/disk#

for all disks?

Exactly, the used space looks accurate, it is not depleted by the 3.22TB ...

 

Linux 4.14.13-unRAID.
root@QUAZAR:~# btrfs fi show /mnt/disk#
ERROR: not a valid btrfs filesystem: /mnt/disk#
root@QUAZAR:~# btrfs fi show /mnt/disk1
Label: none  uuid: b6eace12-2351-4f56-9781-39433327b319
        Total devices 1 FS bytes used 2.36TiB
        devid    1 size 2.73TiB used 2.48TiB path /dev/md1

root@QUAZAR:~# btrfs fi show /mnt/disk2
Label: none  uuid: ebc5fd60-c87a-4f2d-9de6-9f173947b15d
        Total devices 1 FS bytes used 998.28GiB
        devid    1 size 1.36TiB used 1.14TiB path /dev/md2

root@QUAZAR:~# btrfs fi show /mnt/disk3
Label: none  uuid: 0e6c8b99-bf31-4e93-b9d8-bd2d9d816e8c
        Total devices 1 FS bytes used 116.98GiB
        devid    1 size 465.76GiB used 122.05GiB path /dev/md3

root@QUAZAR:~# btrfs fi show /mnt/disk4
Label: none  uuid: 3c96d353-bb8d-4f92-855e-807b5dd3005c
        Total devices 1 FS bytes used 102.33GiB
        devid    1 size 465.76GiB used 105.05GiB path /dev/md4

root@QUAZAR:~# btrfs fi show /mnt/disk5
Label: none  uuid: 206d2efc-bc27-4298-a057-19ee51c1da46
        Total devices 1 FS bytes used 969.00GiB
        devid    1 size 1.82TiB used 1.42TiB path /dev/md5

root@QUAZAR:~# btrfs fi show /mnt/disk6
Label: none  uuid: 31ebe411-8bff-4a50-95fb-dc51be382113
        Total devices 1 FS bytes used 1.08TiB
        devid    1 size 1.82TiB used 1.27TiB path /dev/md6

root@QUAZAR:~#

 

Everything looks normal, did you check disk by disk, i.e., check a missing file know location on every disk, the same folder could exist but be empty on different disks.

Yes, one of the first things I did. files are not under the shfs mount /mnt/user /mnt/user0 or under any of the 6 diskx mounts. The issue also persists after a server reboot, and whether or not folder caching is enabled.

So what is the result of running "du -sh /mnt/disk<id>" on the different disks? Does the sum of all files for each individual disk reasonably match the btrfs usage size for that disk?

No idea then, the space is being used by something, if it was filesystem corruption there should be something in the log and since the 3.2TB can't all be missing from the same disk, the chances of multiple disks having filesystem corruption at the same time are very low, never seen a similar issue before,

A possibility that comes to mind, maybe you moved one or more folders by accident like with drag and drop, make sure you search for a know missing file in the complete share.

Think I found it ... GoodSync seems to have the files in its _saved_ folder. Following up with them, but this no longer looks like an unRAID / shfs or btrfs issue..

 

Thank you for the pointers, once the space was confirmed I just did a search for one of the missing files in /mnt/user and it popped up.

 

?