Re: [netmod] NMDA <operational> output with "when' condition

[Robert Wilton <[e-mail]>]

Hi Rohit,

The rule is that the operational state datastore always reports the 
value in use.  So, if f1="foo" and f2="bar" then that is what should be 
returned.  The fact that it violates a when statement is allowed, but in 
this scenario, not ideal.

However, if leaf f2 can hold values even when f1 is not set to "hello" 
then I don't think the model below is as clear as it could be.

Perhaps it would be better as:

list list1{

   key f1;

   leaf f1{

     type string;

   }

   leaf f2{

     config false;

     type string;

   }

   leaf f2-cfg-override {

     when ".../f1='hello'";

     type string;

     description "Force f2 override"

   }

Or, alternatively, remove the when statements from f2, perhaps add a 
default value "bar", and state in the description that it is only 
configurable when "f1 is set to hello".

Thanks,
Rob


On 11/07/2018 15:23, Rohit R Ranade wrote:
>
> Hi All,
>
> Consider the model below:
>
> ----------------------------------
>
> list list1{
>
>   key f1;
>
>   leaf f1{
>
>     type string;
>
>   }
>
>   leaf f2{
>
>     when ".../f1='hello'";
>
>          type string;
>
>   }
>
> }
>
> The intention of this model is to allow configuration of leaf f2 by 
> user, only when value of f1 is 'hello'.. For all other values of f1, 
> say 'foo', the system will internally choose some value say 'bar'.
>
> When query all fields on <operational>, using condition f1='foo', if 
> need to output its corresponding f2 value (i.e ‘bar’) the when 
> condition will get violated .  Is there some way to keep the 
> restriction on <running> , but relaxing it when showing the 
> <operational> output ?
>
> RFC 8342 has the below section , but I think we cannot use the 
> statements below as there is no “abnormal” value here.
>
>    <operational> SHOULD conform to any constraints specified in the data
>
>    model, but given the principal aim of returning "in use" values, it
>
>    is possible that constraints MAY be violated under some circumstances
>
>    (e.g., an *abnormal value* is "in use", the structure of a list is
>
>    being modified, or remnant configuration (see Section 5.3.1) still
>
> exists).
>
> Any help is appreciated.
>
> With Regards,
>
> Rohit R Ranade
>
>
>
> _______________________________________________
> netmod mailing list
> [e-mail]
> https://www.ietf.org/mailman/listinfo/netmod