[6.12.8] Mover throwing errors: "Operation not supported" followed by "file exists"

[Gymnae]

Hi there,

 

I'm a pretty new user. But ever since installing unraid, I have an issue with the mover. My cache drive (It's a pool of one SSD) fills up, and the mover cannot move data off it. At least, it appears to fail at deleting data after moving them off the cache to the array

 

For example, when moving freshly created TimeMachine Backups, the following error is thrown:

move_object: <file/folder> Operation not supported

I assume that could be a permission issue, or some attribute that is set and causes problems. I ran the "New Permissions" function a couple of times for no change. 

 

When initiating Mover again, another error is thrown, for the same file/folder:

move_object: <folder/file> File exists

So far, I deleted the specified file/folder on /mnt/cache with mc after such a log entry. But this is not a sustainable practice. 

What could cause this? My cache drive is btrfs with compression, while the array is zfs. 

 

 

datengrab-diagnostics-20240218-1923.zip

[Gymnae]

I tried setting up the cache pool after emptying the SSD - the problem persists. I always have to manually delete files via mc off the cache drive.

[JorgeB]

If possible please post non anonymized diags, or it's difficult to see the mover log.

[Gymnae]

1 hour ago, JorgeB said:

If possible please post non anonymized diags, or it's difficult to see the mover log.

I could, but I rather not in this public forum. Is there another place where I could share it, which is accepted, maybe where I can password protect the file?

[JorgeB]

In this case you can PM them to me.

[JorgeB]

On 2/18/2024 at 6:31 PM, Gymnae said:

So far, I deleted the specified file/folder on /mnt/cache with mc after such a log entry. But this is not a sustainable practice.

Not sure what's causing the error, but just to confirm:

 

-first time you run the mover you get "Operation not supported"

-second time the you get "File exits" for those same files

 

Is that correct, and after deleting the duplicates, does the mover succeed or does it enter the same loop?

 

 

[Gymnae]

40 minutes ago, JorgeB said:

Not sure what's causing the error, but just to confirm:

 

-first time you run the mover you get "Operation not supported"

-second time the you get "File exits" for those same files

 

Is that correct, and after deleting the duplicates, does the mover succeed or does it enter the same loop?

 

 

Yes ,that is what is happening

 

To your second question: I clean the cache drive and have a new backup pour files onto the cache drive, or copy files to it, then the loop starts anew.

[JorgeB]

After consulting with someone who uses a Mac, since I don't, I've been told that for TM shares it would be best to avoid the mover, since having parts of the backup in different devices may cause issues, and ideally you'd want to use an exclusive share, but for that would need a large enough pool.

[Gymnae]

2 hours ago, JorgeB said:

After consulting with someone who uses a Mac, since I don't, I've been told that for TM shares it would be best to avoid the mover, since having parts of the backup in different devices may cause issues, and ideally you'd want to use an exclusive share, but for that would need a large enough pool.

 

It doesn't appear to be a problem with TimeMachine.

I copied via file manager to a normal SMB share and then initiated the mover. The issue happens no matte if I copy the file via MacOS or Linux:

 

Feb 20 21:24:24 Datengrab move: mover: started
Feb 20 21:24:24 Datengrab move: file: /mnt/cache_pool/media/proxmox-ve_8.1-1.iso
Feb 20 21:24:36 Datengrab move: move_object: /mnt/cache_pool/media/proxmox-ve_8.1-1.iso Operation not supported
Feb 20 21:24:36 Datengrab move: mover: finished
Feb 20 21:24:51 Datengrab emhttpd: shcmd (271): /usr/local/sbin/mover |& logger -t move &
Feb 20 21:24:51 Datengrab move: mover: started
Feb 20 21:24:51 Datengrab move: file: /mnt/cache_pool/media/proxmox-ve_8.1-1.iso
Feb 20 21:24:51 Datengrab move: move_object: /mnt/cache_pool/media/proxmox-ve_8.1-1.iso File exists
Feb 20 21:24:51 Datengrab move: mover: finished

 

Edited February 20 by Gymnae
clarifying that the issue is not limited to TimeMachine

[JorgeB]

Post the output of:

ls -la /mnt/cache_pool/media/

and

ls -la /mnt/user/media/

 

 

You can just include a couple of files, likes those ISOs

[Gymnae]

Gladly, this is the result:

root@Datengrab:~# ls -la /mnt/cache_pool/media/
total 1262328
drwxrwxrwx 1 nobody users         40 Feb 20 21:21 ./
drwxrwxrwx 1 nobody users         32 Feb 20 21:16 ../
-rwxrw-rw- 1 nobody users 1292605440 Nov 23 20:21 proxmox-ve_8.1-1.iso*
root@Datengrab:~# ls -la /mnt/user/media/
total 1262358
drwxrwxrwx 1 nobody users         40 Feb 20 21:24 ./
drwxrwxrwx 1 nobody users          7 Feb 20 21:16 ../
drwxr-xr-x 1 nobody users         16 Feb 17 05:13 games/
drwxrwxrwx 1 nobody users          5 Dec 11 16:51 iMovie\ Library.imovielibrary/
drwxrwxrwx 1 nobody users         32 Feb 14 23:08 music/
drwxrwxrwx 1 nobody users         23 Feb 17 05:51 photos/
-rwxrw-rw- 1 nobody users 1292605440 Nov 23 20:21 proxmox-ve_8.1-1.iso*
drwxrwxrwx 1 nobody users         10 Feb 15 21:54 video/

 

[JorgeB]

I'm afraid I cannot see where the problem is, everything looks normal to me, and I've never seen this issue before, if you can suggest reformatting the pool, try a different filesystem mostly to be different, if you can also re-format the array disk with xfs for example, then create a new test share and retry.

[Gymnae]

2 minutes ago, JorgeB said:

I'm afraid I cannot see where the problem is, everything looks normal to me, and I've never seen this issue before, if you can suggest reformatting the pool, try a different filesystem mostly to be different, if you can also re-format the array disk with xfs for example, then create a new test share and retry.

I can recreate the pool no problem, but the array is tougher. I have stored data on it which I have, currently, no other place to put. No Backup. Yes. Also, I'd want to benefit from ZFS and the RAM I have, so I'd like to stick with ZFS.

Is there something else I could try? Might compression be an issue?

[JorgeB]

17 minutes ago, Gymnae said:

Might compression be an issue?

Doubt it, one thing I didn't remember to ask before, please post output of

 

zfs get all disk1/media

 

[Gymnae]

Re-formatting the cache SSD seemed to have done the trick. Time Machine backups as well as other shares are moved without a hitch.

I think it might have been my tweaking of the BTRFS partition with the same flags I used on my system drive. Now I kept it vanilla and will keep it this way