Re: [Netconf] I-D Action: draft-wu-netconf-restconf-factory-restore-00.txt

[Robert Wilton <[e-mail]>]

On 02/07/2018 17:08, Andy Bierman wrote:
>
>
> On Mon, Jul 2, 2018 at 7:01 AM, Juergen Schoenwaelder 
> <[e-mail] 
> <mailto:[e-mail]>> wrote:
>
>     I suggest to try to make the proposal simpler, not more complex.
>
>
>
> +1
>
> IMO the only thing needed is 1 simple identity named "factory".
Yes, I basically agree.

In addition, when a device boots then <startup> is initialized to 
<factory> if it doesn't exist.  Or alternatively, <running> is 
initialized to <factory> if <startup> doesn't exist.
A client can use an RPC to copy <factory> to <startup> or to <running> 
if they want, but then can never write to <factory> (although factory 
could change via a software update).

> Existing protocol operations can be used (e.g, copy-config from 
> factory to running).
Now RESTCONF is datastore aware, it looks like we need to add a 
"copy-config" RPC (or should it just be "copy"?) to RESTCONF to allow 
the contents of one datastore to be copied to another datastore.  Such 
an RPC should be entirely generic and not tied to the factory datastore 
in anyway.  Possibly, related to this, it might be worth considering if 
there are any other operations in NETCONF that should be supported in 
RESTCONF and to do that as a single update to the protocol rather than 
lots of piecemeal extensions.

Thanks,
Rob


>
>
>     /js
>
>
> Andy
>
>
>     On Mon, Jul 02, 2018 at 01:56:54PM +0000, Qin Wu wrote:
>     >
>     >
>     >
>     > 发件人： Juergen Schoenwaelder
>     > 收件人： Qin Wu<[e-mail]
>     <mailto:[e-mail]><mailto:[e-mail]
>     <mailto:[e-mail]>>>
>     > 抄送： Kent Watsen<[e-mail]
>     <mailto:[e-mail]><mailto:[e-mail]
>     <mailto:[e-mail]>>>;Ladislav Lhotka<[e-mail]
>     <mailto:[e-mail]><mailto:[e-mail]
>     <mailto:[e-mail]>>>;netconf<[e-mail]
>     <mailto:[e-mail]><mailto:[e-mail]
>     <mailto:[e-mail]>>>
>     > 主题： Re: [Netconf] I-D Action:
>     draft-wu-netconf-restconf-factory-restore-00.txt
>     > 时间： 2018-07-02 20:23:06
>     >
>     > On Mon, Jul 02, 2018 at 12:16:27PM +0000, Qin Wu wrote:
>     > > Good point and suggestion. Here is my thought:
>     > > In case :startup capability is supported, the factory
>     datastore can be copied into <startup>, restart is not needed
>     since we have loaded content of factory datastore into startup.
>     Startup will be updated with running each time the running is
>     altered. In case of system fatal error, we will consider copy
>     factgory datatore into <startup> again for restore.
>     >
>     > I doubt this will work. If you copy <factory> to startup> and
>     > subsequently <running> to <startup>, there is a copy of
>     <running> left
>     > in <startup>, i.e., the copy of <factory> to startup> had no effect.
>     > [Qin] To address this issue, we can introduce multiple target
>     data sores in the new operation factory restore operation, copy
>     factory datastore to startup and running in one operation, the
>     source will be set to the same factory datastore.
>     >
>     > > In case the restart is needed or device power on is needed,
>     the factory datastore as source may not be set, instead, URL is
>     set as source, the content of source identified by URL can be
>     loaded into target datastore during restart or device repower on.
>     In this case, the proposed factory datastore
>     > > and new operation can work together with zero touch
>     bootstrapping procedure proposed in draft-ietf-netconf-zerotouch.
>     >
>     > URLs are an optional capability so far and I do not know what
>     "URL is
>     > set as source" means to me. What is target datastore here? I am not
>     > sure how data flows.
>     > [Qin] see device power on procedure defined in zero touch
>     netconf WG draft, factory restore scheme, in my opinion can be
>     integrated into it. The target datastore(s) can be set to any
>     datasore(s) you want to return factory default.
>     >
>     > > In case :writable-running capability is supported, the factory
>     datastore can be directly copied into <running>, in case of
>     multiple conceptual flows, we can consider to copy one factory
>     datastore into multiple <running> targets.
>     > > In case :candidate capability is supported, the factory
>     datastore can be first copied into <candiate> and then the
>     <candidate> is committed into <running>, in this case, startup is
>     not touched.
>     >
>     > What are multiple <running> targets? What are "multiple
>     conceptual flows"?
>     > I am confused.
>     > [Qin] see above, in the above cases,multiple datasores can be
>     set to <startup> and <running> And all other datastore that need
>     to set to factory default. That means in the new operation,
>     multiple target list can be included. If multiple factory defaults
>     are allowed,multiple sources can be included as well.
>     >
>     > Not sure multiple target running case exists,if we can copy one
>     source to multiple instances distributed in multiple logical
>     network elements, that will be great.
>     > /js
>     > 
>     > --
>     > Juergen Schoenwaelder Jacobs University Bremen gGmbH
>     > Phone: [phone] Campus Ring 1 | 28759 Bremen | Germany
>     > Fax: [phone] <https://www.jacobs-university.de/
>     <https://www.jacobs-university.de/>>
>
>     -- 
>     Juergen Schoenwaelder           Jacobs University Bremen gGmbH
>     Phone: [phone]         Campus Ring 1 | 28759 Bremen | Germany
>     Fax:   [phone]         <https://www.jacobs-university.de/
>     <https://www.jacobs-university.de/>>
>
>     _______________________________________________
>     Netconf mailing list
>     [e-mail] <mailto:[e-mail]>
>     https://www.ietf.org/mailman/listinfo/netconf
>     <https://www.ietf.org/mailman/listinfo/netconf>
>
>
>
>
> _______________________________________________
> Netconf mailing list
> [e-mail]
> https://www.ietf.org/mailman/listinfo/netconf