Best practice setting up datasets for apps: SMB or GENERIC?

[pixelwave]

I recently run into some (rights-related) issues with mit datasets which are connected to apps (hostpath).

I usually set them up as SMB types for easier access control and manipulate files.

Some error messages pointed towards a problem potentially using NFSv4 permission management versus the generic one?

Should datasets which are connected to apps / containers usually be generic or smb?

 

[anodos]

I recently run into some (rights-related) issues with mit datasets which are connected to apps (hostpath).

I usually set them up as SMB types for easier access control and manipulate files.

View attachment 54179

Some error messages pointed towards a problem potentially using NFSv4 permission management versus the generic one?

Should datasets which are connected to apps / containers usually be generic or smb?

SMB should be fine, but some apps may want to be able to chmod(2) and so you'll need to change the aclmode to passthrough in this case. You'll also probably want to make sure you grant EVERYONE@ the TRAVERSE rights set at a minimum for the dataset mountpoint (and make sure your apps user has rights to the dataset in question).

 

[anodos]

One other nuance to SMB preset is that the filesystem is case-insensitive, which may not be appropriate for some apps. If you're just trying make plex data writable over SMB then case sensitivity doesn't really matter, but you'd need to figure out what's appropriate on a case-by-case basis.