Re: [yang-doctors] Dealing with BFD RFC 9127 client-cfg-parms for PIM, OSPF, ISIS and other BFD clients on some platforms

[Jeffrey Haas <[e-mail]>]

Martin,


> On Nov 10, 2021, at 10:26 AM, Martin Björklund <[e-mail]> wrote:
> 
> Jeffrey Haas <[e-mail]> wrote:
>> On Wed, Nov 10, 2021 at 03:24:19PM +0100, Martin Björklund wrote:
>>> I'm not ok with this violation in general, but I am ok with it in this
>>> particular case.  Hence my questions.
>> 
>> Understood.
>> 
>>> If it turns out that in the end
>>> you modify the AUTH48-docs and wait for the bis anyway, then I don't
>>> think this is the right way to go.
>> 
>> So, your preference is "ship AUTH48 docs unchanged, even though it'd have a
>> potentially redundant 'feature bfd'"?
> 
> Of course my preference is to not violate the upgrade rules, if
> possible.
> 
> 1.  publish the auth48-docs now with redundant "bfd" feature
> 2.  publish the auth48-docs now and remove the redundant "bfd" feature
> 3.  remove the redundant "bfd" feature and wait for -bis
> 
> I am not sure I understand what the proposal is at this point (as I
> understood it, your original proposal was (1), but also w/o "bfd" in
> -bis).

Correct.  The critical detail the yang doctors was solicited for was if we can proceed with a quick 9127-bis to deal with the missing if-feature.  Adding that if-feature in the -bis avoids having to maintain the contents of a grouping in multiple documents that were otherwise inheriting it.

> 
> With (2), the published docs won't be very useful until the -bis is
> published, right?

It creates a new dependency, yes.

> 
> In the case of (3), I think it is better to make a proper fix to -bis.
> 
> And in the case of (1), you could as well introduce redundant
> "client-cfg-param" features (as I suggested earlier), and avoid the
> proposed -bis YANG upgrade rule violation.
> 
> Anyway, if you still decide to go with (1) or (2), I am ok with the
> proposed -bis.  (I prefer (2) over (1)).

Thanks for your opinion.  We'll hopefully close this with Rob and the routing-ADs shortly.

-- Jeff

> 
> 
> 
> /martin