On Sun, Apr 19, 2026 at 01:36:48AM -0400, Mouse wrote: > > This is a matter of where the valuable programming time goes, to make > > it easier to use the old protocol versions, or to be able to use NFS > > with a sound security model. > > If programmer time were fungible, that would be true (or at least > vaguely close; the amount of effort, and skill required, may not be > equal for the two tasks).
True. Time spared on support for the NFSv3 functionality would not bring us NFSv4 by itself. > what gets done is what people > feel like working on. Whether the results are important to anyone is > relevant only insofar as it influences what people choose to work on. Exactly. This is why it helps to mention which kínd of effort might be most useful. > > As soon as you need security, the performance of v3 becomes > > irrelevant. > > Unless your threat model is such that you can get that security through > infrastructure instead of NFS version choice/config, in which case it > becomes a choice. (For example, I once ran `insecure' NFS over the > open Internet...with security provided by what amounted to a VPN.) Yes, strictly speaking, not totally irrelevant. OTOH NFSv3 itself and the security workarounds come with a cost (not least the inevitable constraints on the system's management and evolution/adjustment). Relying on some mainstream OS with support for NFSv4 does not bring similar disadvantages. As a result, without NFSv4 it is hard to expect that NetBSD would be considered for new NFS installations. :-( kind regards, od2uvb
