Hi Juergen, your particular example is from peer mount (i.e. remote mount). Yes, you are mounting everything below /a/b/c, including augmentations etc. You mount indeed instance information (not just a schema to instantiate just under that mountpoint). So, I think our understanding is consistent. However, the mountpoint itself (foo) is part of a model. You cannot place foo dynamically just anywhere, attaching data/mountpoints into random points in the tree "after the fact", but you need a model in which you define the mountpoint, using the YANG mountpoint extension - this is what I meant with regards to defining a model with mountpoints declared. (And, while each mountpoint is restricted to a single path, there is nothing that would prevent you from defining multiple mountpoints (each referring to a different path).)
--- Alex -----Original Message----- From: Juergen Schoenwaelder [mailto:j.schoenwael...@jacobs-university.de] Sent: Wednesday, February 24, 2016 3:17 PM To: Alexander Clemm (alex) <a...@cisco.com> Cc: Martin Bjorklund <m...@tail-f.com>; netmod@ietf.org Subject: Re: [netmod] explicit mount On Wed, Feb 24, 2016 at 09:04:00PM +0000, Alexander Clemm (alex) wrote: > Juergen, I think you are correct. Also alias-mount and peer-mount (not just > schema-mount) specify mountpoints in the schema. They are not about mounting > arbitrary data in arbitrary places, but defining a model with mountpoints > declared. My reading of draft-clemm-netmod-mount-03 is that you proposed: ac:mountpoint foo { ac:target "pointer-to-place-where-remote-system-information-can-be-found"; ac:subtree "/a/b/c"; } I guess you intend to mount everything below the path /a/b/c, independent of where the path ends and how many augmentations etc. are there - you mount everything in the subtree. Lada and Martin I think do not propose to mount arbitrary paths but only complete YANG modules and everything (in particular instance identifier) is interpreted 'inside' like if you did a chroot (at least this is how I understood things so far). It is not clear yet to me how 'inside' and 'outside' translate to NETCONF and RESTCONF but that is for the protocol people to figure out I guess. My understanding is that Martin can mount several YANG modules on a single mb:mount-point, your mountpoint is restricted to a single specific path. Lada does not need mount points, the stuff goes to any place the implementations likes to put it. /js -- Juergen Schoenwaelder Jacobs University Bremen gGmbH Phone: +49 421 200 3587 Campus Ring 1 | 28759 Bremen | Germany Fax: +49 421 200 3103 <http://www.jacobs-university.de/> _______________________________________________ netmod mailing list netmod@ietf.org https://www.ietf.org/mailman/listinfo/netmod
