Re: [netmod] [netconf] netconf-config-change and NMDA

Hi,

Michal Vaško <mvasko@cesnet.cz> wrote:
> Hi,
> I have encountered a problem while validating
> ietf-netconf-notifications netconf-config-change notification while
> following NMDA. The RFC [1] says that the "target" instance-identifier
> in this notification should be validated against operational
> datastore. In my use-case there was a running datastore modification
> and the "target" was not found in the operational datastore so the
> notification failed to be validated. I still believe the notification
> should be valid and this seems to be the result of a flaw in the
> specification somewhere. Thanks for any input.

I think that the problem existed even before NMDA.  The
netconf-config-change says that the target refers to a node in the
affected datastore, and if the affected datastore was startup the
instance-identifier wouldn't validate with the given XPath context.

Also, even in the case that the datastore was 'running', if the
operation was 'delete', the target node could presumably point to the
deleted node, which wouldn't validate.

Not sure what the right fix is; maybe we shouldn't use an
instance-identifier at all in this notification?  Or maybe it is ok
since the description clearly says that the target refers to a node in
the affected datastore.

[OTOH, in general it is very hard to tell if validation against the
operational state works or not, since it requires a stable snapshot of
the operational state.]

/martin

> 
> Regards,
> Michal
> 
> [1] https://tools.ietf.org/html/rfc8342#section-6.1
> 
> _______________________________________________
> netconf mailing list
> netconf@ietf.org
> https://www.ietf.org/mailman/listinfo/netconf
> 