On Wed, Nov 28, 2018 at 03:43:01PM +0000, Robert Wilton wrote: > > > > > > I do not understand why a datastore leaf is not sufficient and why we > > need yet something new. If ever needed, NMDA does allow us to define > > new datastores. > > Because a distinction between "candidate" vs "running" vs "intended" won't > necessarily be that useful. Although knowing "running" vs "intended" would > allow the client to know whether it is pre/post template expansion and that > might be useful.
I think it is. I would also assume that the need to save <candidate> instance data in a file is rare. So its likely <running> or <intended> and for some systems the difference is small, for others it may not be small. My point is we have a way to distinguish things and we should use that instead of creating something new that lumps things together again. > The second reason is that I don't know whether things like capabilities and > diagnostics will be new datastores or just part of operational. I don't > think that either of these two areas have really been properly worked out > yet. I presume that they will come over time, once more network management > becomes more automated. By default this is all <operational>. The schema (aka YANG library) exists in <operational> already today. I do not know what your definition of 'diagnostics' is but until we have something else, I assume this is in <operational>. And yes, if systems have other creative datastores, we have a way to deal with that as well. Its all there and I am afraid that creating yet another way to classify data just adds pain and brings back ambiguity. /js -- Juergen Schoenwaelder Jacobs University Bremen gGmbH Phone: +49 421 200 3587 Campus Ring 1 | 28759 Bremen | Germany Fax: +49 421 200 3103 <https://www.jacobs-university.de/> _______________________________________________ netmod mailing list netmod@ietf.org https://www.ietf.org/mailman/listinfo/netmod