Moving the thread from netconf to netmod. Will the authors pull 6020bis back into the WG to reach the rough consensus?
> On Aug 17, 2016, at 2:13 AM, Martin Bjorklund <[email protected]> wrote: > > Hi, > > I have read this long ML thread twice now, and I agree with Andy that: > > 1) We should not / cannot make design changes in an errata or late in > AUTH48; in order to do this we need to pull the document back to > the WG and reach (rough) consensus on the behavior (note btw that > this thread is currently in NETCONF, it really should be NETMOD). > > 2) Since servers MAY delete NP-containers in some cases, clients can > easily handle NP-containers by using "merge" on them. > > > I also agree with Jason that ideally the server should never fail on > any kind of operation on an NP-container, regardless of current state > and requested operation. (But again, this is not a simple > clarification of the current text.) > > > And to answer the original question, I think the server that first got > a request to create the empty NP-containers and then a request w/ > operation "none" is not correct when it fails with a "data-missing" > error. There is no text in 6241 or 6020 that supports this behavior. > > > /martin > > _______________________________________________ > Netconf mailing list > [email protected] > https://www.ietf.org/mailman/listinfo/netconf Mahesh Jethanandani [email protected] _______________________________________________ netmod mailing list [email protected] https://www.ietf.org/mailman/listinfo/netmod
