[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: Separation of configuration and control - good or bad?
[Thank you for responding directly to me. I am still not receiving
netconf mail.]
Larry Menten wrote:
As the draft currently stands, the deletion of an interface requires
providing the complete, desired
configuration subtree to be contained in the parent element and doing a
"replace" operation.
Do you agree that this is the implication of the design, or am I missing
something?
I agree with the implication of the design, and I believe you're missing
something. The deletion of an interface requires the execution of an
edit-config operation which in turn will modify whatever portion of the
configuration you have specified. How that is structured will depend
very much on the individual device schema.
You have missed the point. The spec forces a mechanism that makes many
simple configuration operations fragile and dangerous. Nothing to do
with the
schema. All to do with the model adopted in the draft.
This is just hyperbole. Adding even a non-granular lock fixes your
complaint.
It wasn't meant to be hyperbole. A design that supports
add/delete/replace/merge
and explicitly names the target of the operation does not require as
much locking
to assure robust management as one that expresses transactions as
overlay/replace/merge
and leaves it to the agent to determine what target was meant.
I agree with you to a point. If you can do granular locking then life
is better. However, that requires two things: (a) a standard and (b)
that each network element implement that standard. You're asking for a
very high bar on implementations, particularly low end ones. On those
low end implementations there truly is no need for granular locking.
Eliot
--
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/>