Fuse mounts on host passed to container via bind mount/NFS not working


(David Gutowsky) #1

I run a datahoarder/homelab situation at home with cloud storage. I am trying to use rclone and plexdrive on the host to mount my cloud storage. I can mount them on the host and see the files fine, but when I try to share them either NFS or bind mount, the files within those mounts of my g-drive or AWS are empty. Like they do not even exist. The folders are there with the right permissions, but nothing in the folders.
The reason I am trying to do this is because I would like to have a central location of my cloud storage for my containers instead of individually mounting them in each and every container. I’d like to be able to run those fuse mounts on the host and pass their directories via bind mount or something to the containers as a centralized sharing of folders,
I’ve been banging my head against the wall for a day and getting no where. Is what I am trying to do even possible?
Thanks in advance.

Edit: My container(s) are privileged.


#2

I think you are stumbling on this,

That is, in order to do what you want, you need to use NFSv4 instead, with a special parameter.


(David Gutowsky) #3

I’ll take a look at it. I’ll report back once I get access to the host. Thank you.


(David Gutowsky) #4

I have to say, first, thank you for your help.

Secondly, let me say that it works, but NFS crashes in some way after some heavy usage, or what appears to be heavy usage, especially around the RClone Fuse mount portion of the NFS mount. On the client end, it looks like the mount is down. NFS Server is reporting that it’s down. If I try to remount it from the client side, it doesn’t work. Nothing works until I restart the NFS server. It’s frustrating and it is not clear, to me, what the solution is.

So, in all, it’s not exactly a LXC issue, but I wanted to complete the thread. I might just go ahead and mount the rclone mount directly in the LXC, but I wanted to avoid this, but I am not sure I can.


#5

Thanks for reporting back.
If the software crashes, then it’s a bug report to be written.

You mention heavy usage. Is that over days or it can happen within a day of use?
Does dmesg show something crashing?
It is likely to be some memory corruption.


(David Gutowsky) #6

I think it might be ENCFS crashing on the host computer, which caused NFS to crash and cause issues; I was exporting the de-encrypted version. Going to try running ENCFS in the LXC container and see if that stabilizes things.