Re: [mpls] [netmod] YANG - Intended-Config & Applied-Config & Derived-State & Operational-state...grrrr... !!

[Gert Grammel <[e-mail]>]

This case is another reason why the intended config should not be touched
by the server. Juergen is right that there are lots of different
implementations. It is the implementer¹s choice where to put configuration
state. Typically if a server supports rollback, the state is stored in a
database, otherwise it is often stored directly on HW. Both versions
survive a reboot. If you remove the LC (=Line Card?) the result is either
a) intended = config != active in case of a central config repository or
b) intended != config = active in case of config stored on LC

a) is a case that is often tolerated or even desired, as it allows to
pre-configure HW before it becomes physically available. B) is an error
condition.

Gert

On 2016-06-04 03:28, "netmod on behalf of Juergen Schoenwaelder"
<[e-mail] on behalf of
[e-mail]> wrote:

>On Wed, Apr 06, 2016 at 02:28:03AM +0000, Rajiv Asati (rajiva) wrote:
>> Hi Juergen,
>> 
>> Thanks for the clarity. I somewhat agree. Different lifetime shouldn't
>>really cause any impact.
>> 
>> One nit,
>> 
>> When I pull out a configured interface, the operational state of this
>> interface is gone but not the config - if I put the interface back, it
>> should come up configured as before.
>> 
>> Depends. If LC itself is removed, then all the related interfaces would
>>be removed from the running-config. And reinserting the LC may not bring
>>their individual conf back (ignoring the notifications).
>
>Assuming LC means line card if not discard the following (and explain
>what LC means). I believe the answer here is that systems treat this
>differently and there is no universal answer. Some systems seem to
>store line card config on the line card itself, other systems
>associate config with the physical location of a port or interface,
>yet others associate config with a name of an interface (which may not
>indicate the physical location of an interface). Since implementations
>really differ here, we support multiple implementation options in the
>interfaces data model.
>
>/js
>
>-- 
>Juergen Schoenwaelder           Jacobs University Bremen gGmbH
>Phone: [phone]         Campus Ring 1 | 28759 Bremen | Germany
>Fax:   [phone]         <http://www.jacobs-university.de/>
>
>_______________________________________________
>netmod mailing list
>[e-mail]
>https://www.ietf.org/mailman/listinfo/netmod