Re: [netmod] Schema Mount with Inline Type

Rohit R Ranade <rohitrranade@huawei.com> writes:

> Hi,
>
> I donot have a specific scenario as of now. But the scenario used in the draft is that a module which is mounted needs to refer to its parent-schema. I donot see how that is related to "mount point instances having shared-schema or different schema". 
>
> If we look at the LNE draft, I think it avoids the "parent-reference"
> by having a " bind-lne-name" which binds the interface to the LNE and
> also creates a "system" configuration for that interface inside the
> LNE instance. So the interface references inside the mount jail get
> resolved.

If I understand it correctly (maybe not), the "bind-lne-name" is used at
the level of Network Device to assign an interface to a configured
LNE. This information IMO is not meaningful inside the LNE itself. If it
is so, there is in fact no parent reference.

>
> So in future, if "inlined" schema needs to use parent-schema, it needs
> to use a "bind" mechanism to add entries from the module in
> parent-schema to the same module under mount-point ?

In my view, the inline case makes only sense in the split management
scenario where each mount jail is managed via a separate NC/RC
server. In this case it makes no sense to refer to data outside the
mount jail.

My standard complaint is that schema mount mixes up two rather different
concepts: the shared-schema case is a data modelling concept whereas the
inline case is more about combining instance data. This causes a lot of
complexity and confusion.

Lada

>
>
> -----Original Message-----
> From: Ladislav Lhotka [mailto:lhotka@nic.cz] 
> Sent: 02 January 2019 19:29
> To: Rohit R Ranade <rohitrranade@huawei.com>; netmod@ietf.org
> Subject: Re: [netmod] Schema Mount with Inline Type
>
> Hi Rohit,
>
> Rohit R Ranade <rohitrranade@huawei.com> writes:
>
>> Hi All,
>>
>>    module: ietf-yang-schema-mount
>>      +--ro schema-mounts
>>         +--ro namespace* [prefix]
>>         |  +--ro prefix    yang:yang-identifier
>>         |  +--ro uri?      inet:uri
>>         +--ro mount-point* [module label]
>>            +--ro module                 yang:yang-identifier
>>            +--ro label                  yang:yang-identifier
>>            +--ro config?                boolean
>>            +--ro (schema-ref)
>>               +--:(inline)
>>               |  +--ro inline!
>>               +--:(shared-schema)
>>                  +--ro shared-schema!
>>                     +--ro parent-reference*   yang:xpath1.0
>>
>> Any reason for not adding "parent-reference" for "inline" type ? What 
>> is the solution for the modules defined under such mount points to 
>> refer to parent schema ?
>
> The inline case was intentionally designed with an impenetrable "mount jail". Do you see any use case where parent references are needed and the "shared-schema" strategy cannot be used?
>
> Lada
>
>>
>> With Regards,
>> Rohit
>> _______________________________________________
>> netmod mailing list
>> netmod@ietf.org
>> https://www.ietf.org/mailman/listinfo/netmod
>
> --
> Ladislav Lhotka
> Head, CZ.NIC Labs
> PGP Key ID: 0xB8F92B08A9F76C67

-- 
Ladislav Lhotka
Head, CZ.NIC Labs
PGP Key ID: 0xB8F92B08A9F76C67