Re: [Netconf] Comments on draft-wu-netconf-restconf-factory-restore-01

Qin Wu <bill.wu@huawei.com> Tue, 17 July 2018 03:24 UTC

From: Qin Wu <bill.wu@huawei.com>
To: Balazs Lengyel <balazs.lengyel@ericsson.com>, "netconf@ietf.org" <netconf@ietf.org>
Thread-Topic: [Netconf] Comments on draft-wu-netconf-restconf-factory-restore-01
Thread-Index: AdQdcQJ6YAv/CNcZQFOK/EKsmnd7sg==
Date: Tue, 17 Jul 2018 03:24:06 +0000
Message-ID: <B8F9A780D330094D99AF023C5877DABA9AF4FF3D@nkgeml513-mbx.china.huawei.com>
Accept-Language: zh-CN, en-US
Content-Language: zh-CN
Content-Type: multipart/alternative; boundary="_000_B8F9A780D330094D99AF023C5877DABA9AF4FF3Dnkgeml513mbxchi_"
MIME-Version: 1.0
Archived-At: <https://mailarchive.ietf.org/arch/msg/netconf/ob06urGuNUGnTJKrudd1U71hpnY>
Subject: Re: [Netconf] Comments on draft-wu-netconf-restconf-factory-restore-01
Precedence: list

Hi, Balazs:
发件人: Netconf [mailto:netconf-bounces@ietf.org] 代表 Balazs Lengyel
发送时间: 2018年7月17日 4:36
收件人: netconf@ietf.org
主题: [Netconf] Comments on draft-wu-netconf-restconf-factory-restore-01

Hello Qin,
Factory default Setting Capability for RESTCONF
https://tools.ietf.org/html/draft-wu-netconf-restconf-factory-restore-00

From the ietf-system YANG module:
"system-restart : Request that the entire system be restarted immediately. " So we already have a restart command,

and I still not understand the difference between your definition of reboot and restart.
[Qin]:system-restart means the entire system (not just NETCONF server, but also include DNS server，etc.) needs to be restarted while reboot means the NETCONF/RESTCONF server needs to be restarted. I see this slight difference.
From section 3.6 of RFC7317:
"The 'system-restart' operation can be used to restart the
entire system (not just the NETCONF server). "

If we implement this we need it both for netconf and restconf. Can't we do it as a YANG model defining an rpc instead of a protocol operation?
Why not make this one more rpc in ietf-system. We already have there the system-restart.
[Qin]: Yes, that is the option. The problem is reboot is not system level rpc, when we talk about reboot, I believe it means device reboot or restart NETCONF/RESTCONF server rather than restart the whole system.
In addition, do you think we should define transport independent operations to support factory default setting. Such as copy-datastore?
There are network devices that you can reset to factory default, but where you can not read the factory-default configuration.
[Qin]: That’s why we introduce <factory> datastore.
If this is implemented such devices must also be supported. Actually this i also how Netconf works. You can not read what will be the result of a delete-config beforehand.
[Qin]: Exactly.

regards Balazs

Balazs Lengyel Ericsson Hungary Ltd.

Senior Specialist

Mobile: +36-70-330-7909 email: Balazs.Lengyel@ericsson.com<mailto:Balazs.Lengyel@ericsson.com>

[Netconf] Comments on draft-wu-netconf-restconf-f… Balazs Lengyel
Re: [Netconf] Comments on draft-wu-netconf-restco… Qin Wu