CIFS smb2 prevents node from starting

[CashewCaliphate]

Here are my package versions for my current install of proxmox. I added a CIFS share through the GUI used for backups and container data, but when I try to restart my node, the containers that have the share mounted prevent the node from shutting down, giving the classic systemctl error exit code=1 message. After having no other choice but a hard reboot, the node then loads the GUI, but all options are inaccessible, marked with question marks. Console has only this as an output:

proxmox cifs vfs: iotcl error in smb2_get_dfs_refer rc=-5

If proxmox is running on linux kernel 4.15, why is it trying to use SMB2? SMB3 became the default dialect in kernel 4.13. On my windows side, it's using only SMB3 (I disabled all over versions).

I found this bug in kernel 4.15, which should be fixed in 4.17. Does this apply to my circumstance? Does Proxmox have a timeline to upgrading to 4.17?

I'm assuming SMB2 is the issue here, but if there's something else I should also be checking, let me know.

 

[Alwin]

I added a CIFS share through the GUI used for backups and container data, but when I try to restart my node, the containers that have the share mounted prevent the node from shutting down, giving the classic systemctl error exit code=1 message.

If they CT/VMs don't shutdown in time (120sec), they are killed. This will OFC delay the shutdown.

The default is SMB3 on Proxmox VE. According to the bug entry the DFS site support seems to be activated on the NAS.
https://pve.proxmox.com/pve-docs/chapter-pvesm.html#storage_cifs

As Proxmox VE 6.x is out you could upgrade to the new PVE or take the 5.0 kernel only.