Re: [netmod] Opsdir telechat review of draft-ietf-netmod-rfc8022bis-07

[Joe Clarke <[e-mail]>]

On 1/6/18 18:27, Acee Lindem (acee) wrote:
>>The only substantive comment I have is in the comments ahead of the
>>now-obsolete state branches.  Currently, these comments just state
>>"Obsolete
>>State Data".  I wonder if it would make sense to add a bit more text here
>>to
>>reference why these branches are now obsolete.  Perhaps a reference to
>>the NMDA
>>document would be beneficial.
> 
> How about something like:
> 
>   The subsequent data nodes are obviated and obsoleted by the “Network
>   Management Architecture” as described in
> draft-ietf-netmod-revised-datastores.”

Works for me.

>>Another minor nit I noticed (and this is likely an issue with pyang) is
>>that
>>when using a grouping, the YANG tree lists nodes like routing-state ->
>>router-id with a '+' instead of a 'o' (i.e., indicating obsolete).  Not a
>>big
>>deal since the parent container is obsolete.
> 
> Good catch. Due to some subsetting and formatting, these were not
> regenerated. I
> will fix.

Cool.  I wasn't in a place to validate with pyang yesterday.  I had
assumed an issue there, but if it just takes a regen to fix, that would
be great and clearer.

>>
>>One comment I have is that the imports clauses here definitely point out
>>a need
>>to be able to import by some kind of version that will allow to set a
>>minimum
>>requirement (e.g., import by semantic version).  Having comments such as
>>are in
>>the modules now are not machine-consumable, and will likely cause
>>operational
>>challenges for those that do not pay attention.
> 
> We discussed this on the NETMOD list and it is also undesirable to hard
> code a
> version. It would be good to have “greater than or equal to” semantics.

Yes, I have seen and been part of the semver discussion.  This was more
to get keep beating the drum as this will pose an operational issue for
those not as familiar with NMDA.

Thanks, Acee.

Joe