I'm running 6.1.6, but have been encountering this since I upgraded to 6. 

 

My unraid server is strictly a NAS. No docker containers running on it.    I'm exposing my shares via smb for all my windows machines, and NFS for mounting on my download server.  My download machine is running XenServer (v.XS65ESP1016) with a CoreOS (v.835.9.0) VM running on it.  Inside of the CoreOS VM, I'm running the ususal suspects of docker containers (nzbget, sickrage, etc...) After a (seemingly random) period of time, the UI of these apps starts acting wonky and looking in the logs I see the infamous Stale File Handle error.  It happens multiple times a day on my two Couchpotato containers (I run one for kids movies and one for non-kids), but have seen the Stale File Handle in other containers as well, just not nearly as frequently.  A restart of the docker container temporarily fixes the issue. 

 

my cloud-config has entries like this to acomplish the mounting:

- name: mnt-jabba-movies.mount
      command: start
      content: |
        [Mount]
        What=192.168.0.16:/mnt/user/Movies
        Where=/mnt/jabba/movies
        Type=nfs

 

I've done quite a bit of google searching but haven't seen a good recent solution.  any ideas what I can do to eliminate the Stale File Handle errors when using NFS?  If any more info would be helpful, ask away.  I'd like to stay using NFS so unless it's really the only option, I'm hoping not to hear "switch to smb"

 

Thank in advance

I believe the issue is that the shares go to sleep.  They are mounted to the host not the container,  so the container sees it different.  Without getting too technical,  I had a similar issue,  sonarr container,  kept telling me that it was removing media because it didn't see it.  I would just restart the container and everything was fine.

 

So,  is the stale file handle in the container or the host? What happens if you try accessing the same share from the host?

 

My noted situation is the only reason I moved my containers to unRAID,  as the shares were directly passed to the containers

I don't know if my problem is caused by the same thing or not, but just to add another datapoint, I upgraded to 6.1.6 yesterday or maybe the day before, and last night, my NFS connected VMs (ESXi 6.0) began to fail, and not be able to be restarted or moved.  Looking closer, it says ESXi can't write to the filesystem.  I SSH'd into the UnRaid, and tried to remove a .lck file from one of the datastores, and Unraid from the SSH console says it's a read-only filesystem?!? 

 

I rebooted the UnRaid, and it began to work again, but then within a few hours, it starts happening again.  Is anyone else having problems with 6.1.6, either with the filesystem itself, and/or with the NFS aspect??  This has been working on previous versions for me for years..

-Steve