On Tue, Feb 07, 2017 at 11:01:08AM -0800, James Bottomley wrote: > I was talking about inode stability in the filesystem underlying the > export. I believe you're talking about inode number stability > guarantees of the nfs client code itself, which are unrelated to the > inode number guarantees of the exported filesystem? They are 1:1 correlated for a Linux server at least.