Merging 2 NFS mounts each of 100TB capacity

[gpradeepjain]

Hi,

I am having 2 NFS mounts each of 100TB capacity i.e 100+100 = 200 TB. I am planning to mount these 2 NFS on Kubernetes POD and use mergerfs to show them as a single filesystem. Then put Nginx in front of the new union folder as a file explorer. Just want to understand do you see any problems or pitfalls in this design, as the size of the mount is large i.e. 100TB each. Will mergerfs have any problem if the capacity of mount is large as in my case?

[trapexit]

No. Why would mergerfs have issues with the capacity? 200TB is not that much storage and well within the 64bit limits of modern POSIX filesystem size and offset values. The biggest possible issue is using NFS as it's not fully POSIX compliant and has some known complications when interacting with FUSE but that depends entirely on your usage patterns.

[trapexit]

Outside normal concerns for NFS, yes, it's fine.

[gpradeepjain]

I am using via this command config mergerfs -o allow_other,use_ino any other config to add for NFS?

[trapexit]

There really isn't anything not in the docs regarding NFS. Your not exporting mergerfs through NFS so all your options are preference.

[gpradeepjain]

Do i need to add any kind of monitoring on the fuse itself? Do you see any scenario where it may hang or break.

[trapexit]

I don't know what you mean. FUSE is a kernel driver. What exactly is there to monitor?