Stale file handles after updating to 6.9.1.

[phasik]

Hello!

 

Here is my setup:

Unraid - 6 disks + 1 parity. I have several user shares.

Proxmox - Various VMs including Docker, Plex, etc. I have 2 NFS shares for ISOs and VM backups.

Docker - Running on an Ubuntu VM. Various containers, some of which use the same NFS shares (Transmission, Radarr, Sonarr, etc)

 

A few days ago, I updated to 6.9.1 from version 6.8.3. That's when I started to get stale file handles on my various Linux boxes (NFS shares). I updated to 6.9.2 in hopes of a fix, but the issue still persists. I have searched the forums and tried some of the recommendations like changing fuze_remember to 0 or -1, hard links on or off, NFS 4.0 only, etc but nothing helped. I switched to using CIFS shares on the Linux boxes/Docker containers but still have the issue. The only computer that I haven't experienced the issue (knock on wood) is my main desktop, running Windows 10. It has the main user share as a mounted drive.

 

The only fix is to reboot the container/VM or unmount and remount the share in question. This occurs at random times too. My only guess, after reading up on NFS and stale file handles, is that when one of my VMs or containers access/change data on the share, it will screw up the others boxes that are using the share. For instance, Radarr will signal Transmission to download a file. Once Transmission has completed the download, Radarr will move the file into an appropriate folder. Because 2 different hosts are changing/moving files around, this is where the stale handles come from? I believe thats what is going on.

 

Anyone have any thoughts on this or have a way to fix this? Its odd that I started running into this issue only after I updated Unraid. I feel like I am running around in circles here.

[phasik]

Update: I switched everything over to SMB shares and everything seems to be working as intended. I guess NFS was the problem all along..

[dlandon]

Please see this post: