Re: [netconf] Clarification on NETCONF edit-config default-operation replace

On Mon, Jan 11, 2021 at 7:33 AM Rob Wilton (rwilton) <rwilton@cisco.com>
wrote:

> Hi Andy,
>
>
>
> Perhaps we are crossing wires here.
>
>
>
> I’m not trying to change the behaviour, just trying to ensure that the
> spec describes the behaviour unambiguously.  My proposed text was intended
> to clarify the behaviour to be consistent with Jason’s statement below and
> the conclusion of this thread:
> https://mailarchive.ietf.org/arch/msg/netconf/DTPlkJf4enKNHXkBWn49np2e-8c/
>
>
>
> Are you disagreeing with Jason’s statement “An edit-config can't be used
> as a replacement for a full "replace at root" of the entire config like a
> copy-config” or the prior conclusion to the linked thread above?
>
>
>

No.

I guess the wording in both cases is not correct.
There is a description of the operation attribute in sec 7.2 (not very
good, but normative).
The default-operation text should just say that "merge" or "replace" is the
assigned value of
the operation when the attribute is not present in the representation of
the data node.
The value "none" indicates that no operation is assigned if the attribute
is not present.

> Regards,
>
> Rob
>

Andy

>
>
>
>
> *From:* Andy Bierman <andy@yumaworks.com>
> *Sent:* 11 January 2021 15:18
> *To:* Rob Wilton (rwilton) <rwilton@cisco.com>
> *Cc:* Sterne, Jason (Nokia - CA/Ottawa) <jason.sterne@nokia.com>;
> netconf@ietf.org
> *Subject:* Re: [netconf] Clarification on NETCONF edit-config
> default-operation replace
>
>
>
>
>
>
>
> On Mon, Jan 11, 2021 at 6:48 AM Rob Wilton (rwilton) <rwilton=
> 40cisco.com@dmarc.ietf.org> wrote:
>
> Sorry, digging up an old thread here.
>
>
>
> I agree with the outcome of the referenced thread, but I wonder whether it
> wouldn’t be helpful to have an errata to clarify the NETCONF RFC, at least
> so that we fix this text in future?
>
>
>
> One of my colleagues read the NETCONF RFC and took “default-operation
> replace” to be equivalent to a copy-config operation.  I also note that
> some of the references on the web seem to the describe the wrong behaviour.
>
>
>
> Specifically, I wonder whether we should change:
>
>
>
>          replace:  The configuration data in the <config> parameter
>
>             completely replaces the configuration in the target
>
>             datastore.  This is useful for loading previously saved
>
>             configuration data.
>
>
>
> to:
>
>
>
>          replace:  The configuration data in the <config> parameter
>
>             replaces the related configuration in the target
>
>             datastore.
>
>
>
>
>
>
>
> No this should not be changed.
>
> It has been in use for  about a decade.
>
> You propose to remove the only way to replace the entire config and make a
> huge NBC-change
>
> to provide a redundant mechanism instead.
>
>
>
> Regards,
> Rob
>
>
>
> Andy
>
>
>
>
>
>
>
> *From:* netconf <netconf-bounces@ietf.org> *On Behalf Of *Sterne, Jason
> (Nokia - CA/Ottawa)
> *Sent:* 09 March 2020 13:20
> *To:* Ivory, William <william.ivory@intl.att.com>; netconf@ietf.org
> *Subject:* Re: [netconf] Clarification on NETCONF edit-config
> default-operation replace
>
>
>
> Hello William,
>
>
>
> There was a discussion about this on the list a few years ago.  See the
> thread with subject:
>
>
>
> Clarification request for NETCONF edit-config default-operation replace
>
>
>
> An edit-config can't be used as a replacement for a full "replace at root"
> of the entire config like a copy-config.
>
>
>
> Jason
>
>
>
> *From:* netconf <netconf-bounces@ietf.org> *On Behalf Of *Ivory, William
> *Sent:* Thursday, March 5, 2020 7:35 AM
> *To:* netconf@ietf.org
> *Subject:* [netconf] Clarification on NETCONF edit-config
> default-operation replace
>
>
>
> Hi,
>
>
>
> I'd appreciate clarification on the following NETCONF operations relating
> to copy-config and edit-config:
>
>
>
> My understanding is as follows:
>
>
>
> - <copy-config> completely replaces any existing configuration
>
>
>
> - <edit-config> with operation:replace attribute will replace any existing
> configuration with new configuration for nodes specified inside the
> <config> operation. The attribute may be at any level in the request, and
> applies only to the node specified and nodes under that node.
>
>
>
> Where I'm not clear is the <edit-config> operation with the
> default-operation parameter set to replace. RFC 6241 section 7.2 states:
>
>
>
>          replace:  The configuration data in the <config> parameter
>
>             completely replaces the configuration in the target
>
>             datastore.  This is useful for loading previously saved
>
>             configuration data.
>
>
>
> Does this mean that this is the equivalent of the <copy-config> operation,
> ie ALL existing configuration should be removed, even if there is no
> explicit replacement in the new <config> section?
>
>
>
> Thanks,
>
>
>
> William
>
>
>
> _______________________________________________
> netconf mailing list
> netconf@ietf.org
> https://www.ietf.org/mailman/listinfo/netconf
>
>