> On Wed, Oct 14, 2015 at 10:49:02AM -0400, David Reid wrote:
> > > >I'm reviewing 6020bis since it is in working group last call. 
> > > >I see a new requirement that a server MUST NOT implement
> > > >more than one revision of a module. I understand that supporting
> > > >more than one revision can cause problems, but I expect that
> > > >it will happen in practice. I know it happens sometimes with
> > > >MIBs in SNMP. I think MUST NOT is too strong.
> > > 
> > > I've encountered the same phenomenon in the SNMP universe,
> > > so if I expected Netconf to used as a replacement for SNMP
> > > I'd have the same concern.
> > > 
> > > Randy
> > 
> > Here is the situation I face. We put a netconf server in our SNMP Master
> > agent. Using the MIB to YANG conversion rules from RFC 6643 we can provide
> > read-only access (or read-write in a non-standard way) via netconf and yang
> > to all of the MIB data. We have existing customers using different revisions
> > of the same MIB module in different subagents. If they convert those
> > MIB modules to YANG modules and access the information via netconf, it
> > would violate the proposed rule.
> 
> The RFC 6643 translation generates no mandatory statements in YANG and
> the RFC 6643 translation is read-only. So for read-only access, using
> the latest MIB module as a basis should work since SMIv2 has strict
> backwards compatibility rules, no?

For read-only objects, yes. But Andy argued that for read-write objects
I should advertise the oldest revision. This is not a problem in YANG yet
because we don't have multiple versions of YANG modules in the field yet.

Why does yang 1.1 add the new requirement that a server MUST NOT implement
more than 1 revision? If there is an e-mail thread, you can point me at that
rather than re-hash the arguements here.

-David Reid

_______________________________________________
netmod mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/netmod

Reply via email to