Re: Full support of a capability

[Andy Bierman <ietf@andybierman.com>]

Vincent Cridlig wrote:

> Hi,
>
> A few questions on the Netconf draft:
>
> If an agent advertises the #writable-running capability, does it mean 
> that the full data model has to support it ?
> Differently: if a part of the agent data model doesn't support 
> #writable-running capability, is the agent allowed to advertise this 
> capability ?


The capability is for the whole device and is independent of the data model
being modified.

> Or can this be managed with "operation-not-supported" or 
> "operation-failed" error when needed ?


You mean the error-tag for the <copy-config> from <running> to <startup>
(i.e., explicit NV-save for #writable-running) or some other operation?

Let me see if I understand the details here...

Depending on the data model content that is altered (or simply installed?)
during a particular NETCONF session,  the agent sometimes updates
the non-volatile storage as soon as the <running> config is altered,
and sometimes it waits for an explicit <copy-config> operation to
update the <startup> config.    What a mess.   How is an application
supposed to know when it needs to use <copy-config> or not?
Also, <startup> only exists if #writable-running is true, and it is
supposed to be a conceptual mirror or <running>, not a subset.

Perhaps you mean that you have some data models that do not
get saved in NV-storage at all.  In NETCONF terms, this is
state data, not config data.



> Vincent Crdlig


Andy

> --
> to unsubscribe send a message to netconf-request@ops.ietf.org with
> the word 'unsubscribe' in a single line as the message text body.
> archive: <http://ops.ietf.org/lists/netconf/>


--
to unsubscribe send a message to netconf-request@ops.ietf.org with
the word 'unsubscribe' in a single line as the message text body.
archive: <http://ops.ietf.org/lists/netconf/>