On Wed, 2019-08-28 at 10:40 -0400, J. Bruce Fields wrote: > On Wed, Aug 28, 2019 at 10:16:09AM -0400, Jeff Layton wrote: > > For the most part, these sorts of errors tend to be rare. If it turns > > out to be a problem we could consider moving the verifier into > > svc_export or something? > > As Trond says, this isn't really a server implementation issue, it's a > protocol issue. If a client detects when to resend writes by storing a > single verifier per server, then returning different verifiers from > writes to different exports will have it resending every time it writes > to one export then another. > Huh. I've always considered the verifier to be a per-inode quantity that we just happened to fill with a global value, but the spec _is_ a bit vague in this regard. I think modern Linux clients track this on a per-inode basis, but I'm not sure about really old Linux or other clients. It'd be good to know about BSD and Solaris, in particular. -- Jeff Layton <jlayton@xxxxxxxxxx>
