Re: [netmod] draft item 104 meeting 1/2 minutes

[joel jaeggli <[e-mail]>]

I posted a lightly edited version of this as the meeting minutes in the
tracker.

regards
joel

On 4/13/19 15:35, Joel Jaeggli wrote:
> IETF 104 netmod meeting minutes
> 
> Monday Session 1 - Second half of netconf slot, 2019/03/25 1000
> https://www.youtube.com/watch?v=aw0fRh1pvec
> 
> Monday Session 2 - 2019/03/25 1350
> https://www.youtube.com/watch?v=GVSYk_PLPNY
> 
> * Monday Session 1 *
> 
>   Rob Wilton and Company (55 min)
>   YANG Versioning Design Team Update
> 
> Requirements Draft update – Joe Clarke (5 mins)
> draft-verdt-netmod-yang-versioning-reqs-02
> • Design Team update & Solution overview – Rob Wilton (10 mins)
> • Semantic versioning for modules – Balazs Lengyel (20 mins)
> draft-verdt-netmod-yang-semver-00
> • Semantic versioning for YANG schema – Rob Wilton (10 mins)
> draft-rwilton-netmod-yang-packages-01
> • Schema version selection – Reshad Rahman (10 mins)
> draft-wilton-netmod-yang-ver-selection-00
> 
> Joe presenting on requirements - 
> https://datatracker.ietf.org/meeting/104/materials/slides-104-netmod-sessa-yang-versioning-design-team-update-00.pdf
> 
> slide 10: who feels rfc7950 section 11 is sufficient and does not require changes?
> 
> Martin: Is the question really about NBC
> 
> Joe: That is later
> 
> Lada: Section 11 doesn't belong in yang definition
> 
> Joe: Interesting observation, that was not discussed as part of the solution discussion. 
> 
> Lou: Would be more comfortable with question if it didn't include section 11 in the question
> 
> Joe: The question is a bit rough, but wanted to scope the question, e.g., exclude yang next discussion
> 
> On questions 2: "Are NBC changes required?"
> 
> Juergen: question is under specified.
> 
> joe: are nbc changes required (allowed) to an existing node in a given yang module?
> 
> Martin: (Andy's solution from the list) They happen, might be better to describe them as sometime more like a deviation, i.e. a way to express them, but don't indicate that they are good.
> 
> Acee: It would be useful to describe the workflow for clients and servers.
> 
> Joe: We are trying to describe this as guidelines in the solutions document.
> 
> lou: I think we all agree that we need a better solution about non-backward compatible changes.
> 
> joe: do you object to what the requirements documents?
> 
> lou: not but it could be improved to not specify the solution.
> 
> statement I don't really care if it's published an RFC I think it's very important that we have
> a working group document that establishes consensus on or defines the consensus of what it is we're trying to solve
> 
> martin b: agreed, but want it understand if it will be published or not?
> 
> Kent: (Polling)
>   Should requirements doc be published as an RFC - very few
>   Sould requirements doc NOT be published as an RFC -  the same
>   Should we defer this decision - most (~2x more, but still few in the room)
>   
> slide 14
> 
> joe:  is branching required, and if so how much branching is needed and for clarity?
> 
> Christian Hopps:(For branching) Should narrow down the scope? Focus on vendor specific modules. don't see any  need for branching in standards models. vednor models have release trains and other requirements.
> 
> joe: just because something is supported doesn't mean one has to use it.
> 
> I don't know if it's still valid but this was heard on the design team a few months ago so maybe there is a need maybe not here in ops area that we've heard but in routing area that was something that came up on one of our calls
> 
> lada: agree that ietf shouldn't need branches
> 
> joe: from requirements standpoint yes branching should be supported.
> 
> lou: replace required with allowed
> 
> rob: l2sm is an example. the just published a new revision to get around lots of bugs.
> 
> andy: multiple release trains only for vendors and other sdos. 
> 
> joe: sounds like this should be allowed.
> 
> tim c: we do have this model in place in the bbf
> 
> runs out of time for additonal presentations on the agenda.
> 
> scheduling of informal open YANG versioning Design Team meeting 03/28 at ietf.
> 
> meeting ends.
> 
> * Monday Session 2 *
> 
>    Introduction
> 
>      Chairs (10 minutes)
>      Session Intro & WG Status
> 
> Chairs: the following will be LCed after the meeting 
> –draft-ietf-netmod-intf-ext-yang-07
> –draft-ietf-netmod-sub-intf-vlan-model-05 
> 
> Robert Sparks: IETF is managing YANG Catalog 
> 
>    WG documents items:
> 
>      Balazs Lengyel (10 min)
>      YANG Instance Data File Format
>      https://tools.ietf.org/html/draft-ietf-netmod-yang-instance-file-format-02
> 
> Balazs Presenting.
> 
> kent: could I get a sense for the room how many people have read this draft raise your hand please this version of the draft yes so it's a few oh good 
> 
> 	after we see some more review comments we can gauge where we are on last call so it's really
> 
> 
>      Martin or Andy (10 min)
>      YANG Data Structure Extensions
>      https://tools.ietf.org/html/draft-ietf-netmod-yang-data-ext-02
> 
> martin presenting
> 
> 
> lada - don't see benfit of using restconf yang data model
> 
> joe: yang catalog would be of benifit from instance data
> 
> erick: would use it for yang push
> 
> kent: is it important to maintain the -ext structure?
> 
> rob: why does the instance document need to use yang data at all? (lada's comment)
> 
> 	other approach is to just use an annotation.
> 
> joe: this extension allows lists to not need a key element
> 
> lou: please add to you list how this impacts tree diagrams?
> 
> lou: how many have read this version ( a few)
> 	how many have read any version? (one more)
> 	for both documents we'll have to talk among chair about last call before montreal, may use a as forcing function for review.
> 
>      Kent Watsen (10 min)
>      Handling Long Lines in Inclusions in Internet-Drafts and RFCs
>      https://tools.ietf.org/html/draft-ietf-netmod-artwork-folding-01
> 
> christian - is the martins suggestion work for all cases?
> 
> kent - seems like it has a dangerous chance of collision.
> 
> martin: cases where the pretty thing doesn't work
> 
> kent: for automated folding you would need a smart folder
> 
> martin prefers one
> 
> christian: should have one way even if it's a little bit harder
> 
> rob: don't think a single backslash can cover all cases.
> 
> kent:  poll on approaches - double backslash only  (no suppport)
> 	both methods (some support)
> 	only single (also some support)
> 
> kent question footer
> 	a (leave draft as is)
> 	b add a footer 
> 
> lou: clear majority for a over b
> 
>    Non-Chartered items:
> 
>      Kent Watsen (15 min)
>      YANG Next Analysis
>      <no associated draft>
> 
> kent: 70 issues over 3 years created in yang next issue tracker.
> 
> Kent: Where should we focus
> 
> juergen: two other dimensions, do we know how to solve? do we have consensus?
>     
>      Qin Wu (10 min)
>      Factory default Setting
>      https://tools.ietf.org/html/draft-wu-netmod-factory-default-02
> 
> Kent (polling)
> - How many have read draft: a reasonable number
> - How many interested in problem addressed in draft: a reasonable number
> - Now many think we should not work on this topic: none
> - How many think draft should be adopted as starting point: a reasonable number
> - How many think it should not be adopted: none
> 
> Adoption poll will be taken to the list
> 
>      Qin Wu (10 min)
>      NMDA Base Notification for Applied Intended Configuration
>      https://tools.ietf.org/html/draft-wu-netmod-base-notification-nmda-01
> 
> Kent (polling):
>     -  How many have read draft: very few
>      - how man think intersting:  also few
>     
> lou: how many people think this is an area we should be spending time on? (a little more)
> 
> Lou: Need more feedback from group, please send a summary to the list of objectives  to list to try to generate interest
> 
>      Christian Hopps (10 min)
>      YANG Geographic Location
>      https://tools.ietf.org/html/draft-chopps-netmod-geo-location-01
> 
> Kent (polling)
>     - How many have read draft: a few
>     - how man think intersting:  more
>     - how many would like to see document used as basis: a good number
>     - How many think it should not be adopted: none
> 
> Adoption poll will be taken to the list
> 
>      Juergen Schoenwaelder (15 min)
>      Update of Common YANG Data Types (RFC 6991)
>      https://tools.ietf.org/html/draft-schoenw-netmod-rfc6991-bis-00
> 
> Kent (polling)
> - How many have read draft: a reasonable number
> - How many interested in problem addressed in draft: a reasonable number
> - Now many think we should not work on this topic: none
> - How many think draft should be adopted as starting point: a reasonable number
> - How many think it should not be adopted: none
> 
> Adoption poll will be taken to the list
> 
> Update of Common YANG Data Types
> Jürgen Schönwälder
> 
> rob: type defs are cheap so define both nanoseconds and minutes (time resolution)
> 
> lada: host type should be restricted narrowly
> 
> chris: does the canonical format go down to seconds?
> 
> kent: hundreths of seconds
> 
> 	possibly more types (days weeks hours months)
> 
> Lou (polling)
> - How many have read draft: a reasonable number
> - How many interested in problem addressed in draft: a reasonable number
> - Now many think we should not work on this topic: none
> 
> chris: we're adding more common types
> 
> lou:  does it make sense to rev this as a bis. lets use the model of keep reving this
> 
> rob: if we delay by and extra year we'll have more to add.
> 
> lou: lets see how fast we can do it.
> 
>      Michale Wang / Chongfeng Xie (10 min)
>      A YANG Data model for Policy based Event Management
>      https://tools.ietf.org/html/draft-wwx-netmod-event-yang-01
> 
> 
> lou: how many have read this (a few)
> 	maybe it's early but how many this is a starting point (some)
> 	adopting now (less)
> 	wait a bit (few)
> 
> 	this is coming out of yang push dt in netconf
> 
> meeting concludes
> 
> 
> 
> _______________________________________________
> netmod mailing list
> [e-mail]
> https://www.ietf.org/mailman/listinfo/netmod
>