[10:43:29] lhotka joins the room [10:50:49] mark joins the room [11:00:17] andy.bierman joins the room [11:01:22] DavidAlanReid joins the room [11:01:43] DavidAlanReid leaves the room [11:01:57] David Reid joins the room [11:02:14] dlpartain joins the room [11:02:38] dlpartain has set the subject to: NETCONF Session at IETF75 (Stockholm) [11:03:10] hi all [11:03:27] if you're on audio (or think you should be), please tell me if it's OK [11:03:44] fine for me [11:03:56] also, after a rather chaotic ISMS jabber session, i have a favor to ask [11:04:09] if you have something that you want read, please prefix it with READ: [11:04:19] that way i know it's not just an internal jabber thing [11:04:56] also be aware that the delay is on the order of 5-15 seconds in the audio. so, it's helpful if you tell me what you're commenting on [11:05:04] thanks, mark [11:05:20] i'll assume silence means it's fine for D. Reid and Andy [11:05:23] mehmet talking [11:05:33] agenda slides [11:05:45] "Note Well" [11:05:49] slide 4 [11:05:50] Audio is fine for me [11:05:50] Juergen Schoenwaelder joins the room [11:06:15] Bert Wijnen joins the room [11:07:02] http://www3.ietf.org/proceedings/75/slides/netconf-0.ppt [11:07:09] slide 5 [11:07:44] weshardaker joins the room [11:08:02] W500 joins the room [11:08:21] http://www3.ietf.org/proceedings/75/slides/netconf-1.ppt [11:08:36] martin starts _now_ (to test latency) [11:09:04] yes [11:09:19] slide 2 [11:09:39] W500 leaves the room [11:09:53] Mark Ellison joins the room [11:11:22] juergen [11:12:28] asking about references to YANG documents [11:13:44] READ: no, the XSD was created standalone, and has it's own IP addr definition [11:14:21] Atarashi Yoshifumi joins the room [11:14:27] READ: no, don't see the benefit in generating from YANG [11:15:58] dan [11:18:17] how can partial-locking monitoring advance before partial-locking? are both drafts going to IESG together? [11:18:45] clns joins the room [11:18:59] do you want me to read that? [11:19:23] yes: READ (sorry) [11:19:24] washam.fan joins the room [11:19:33] he is doing it now [11:20:23] ray joins the room [11:20:30] OK -- do not see reason to hold back monitoring draft [11:20:47] q: do we make the yang stuff normative? [11:21:19] READ: are we effectively stating that no further models be published until YANG completes? [11:26:15] opinions here? [11:26:35] publish sooner than later [11:26:50] i prefer make yang normative [11:27:42] done with that presentation [11:27:53] http://www3.ietf.org/proceedings/75/slides/netconf-2.pdf [11:27:55] Greetings to all at the IETF meeting :) [11:28:03] thanks :-) [11:28:45] some confusion here about martin's slides... [11:28:47] hang on [11:29:41] slide 2 [11:29:55] Summary of changes [11:30:48] 3 [11:31:28] 4 [11:33:22] coments? [11:33:30] bob cole [11:33:31] Linus Nordberg joins the room [11:33:38] READ: edit-config with test-then-set can generate both kinds of error-path in same response [11:34:41] slide 5 [11:34:57] 4 will sent to the mailing list as a proposed solution [11:34:59] 5 [11:35:18] Bert Wijnen leaves the room: Replaced by new connection [11:35:19] Bert Wijnen joins the room [11:36:10] comments on 5? [11:38:09] wes [11:39:11] keep & fix? or not? [11:39:15] Bert Wijnen leaves the room [11:39:15] Bert Wijnen joins the room [11:39:15] keep it [11:39:25] READ: will introduce a new error-tag break old managers? [11:39:27] hhhmmmm... [11:39:50] that was a hmmm for keeping [11:40:13] room / jabber consensus for "keep & fix" [11:40:21] seems to me, anyway [11:40:29] andy's question needs to be discussed, clearly [11:40:35] slide 6 [11:41:49] comments on 6? [11:41:54] phil [11:42:56] READ: isn't this data-model specific? [11:43:27] will read after phil [11:44:41] READ: All NETCONF content is XML. whether you break big chucks of text into subtrees is up to your data model [11:46:38] READ: but we never define that format and how to convert arbitrary XML content to 'text' [11:48:43] agree or disagree on the proposal to remove this text? [11:48:57] agree to remove this text [11:49:26] hmmms are for removing [11:49:38] slide 7 [11:49:39] hmmm in favor of removing [11:49:43] thanks mark [11:51:30] READ: IMO, we MUST bump the NETCONF base capability version to add new error-tags, or old managers will break [11:51:40] assuming you're looking at the slides, please send your comments as soon as you can [11:52:24] lhotka leaves the room [11:53:21] READ: what about a capability-changed notification? [11:55:28] READ: the manager can also poll the ietf-netconf-state data model for the capabilities [11:55:43] do not want this error-code [11:56:08] comments? [11:56:12] washam.fan leaves the room [11:56:23] hums for? [11:56:25] against? [11:56:39] hmmm, in favor [11:56:40] READ: I had the impression that the set of caps available for a session were the intersection fo manager and agent caps...adding caps would not be in that intersection [11:57:53] the proposal will be sent to the list [11:58:14] slide 8 [11:58:19] utf-8 / 16 [11:58:31] opinions here? [11:59:03] I think utf-8 only is fine [12:01:41] hum for adding text that says netconf servers must support utf-8 [12:01:48] for/against? [12:02:18] for UTF-8 only [12:02:25] balazs.lengyel joins the room [12:02:37] andy, proposal is MUST UTF-8, but more is up to implementations [12:03:16] sorry... on a slide that hasn't been posted. i'll type [12:03:24] Title: 013: confirmed-comit (new) [12:04:02] * If the session dies, a pending confirmed commit is reverted This, If a config change is done so that the session dies, the config change will immediately be reverted. [12:04:28] slide 10: against namespace wildcarding because has a namespace assigned. This proposal violates XML. [12:04:31] * Proposal: add a confirmed-commit:1.1 capability which adds a parameter to the . [12:05:04] where is slide 12 and 13 online? [12:05:41] * It is unclear (existing implementations behave differently) if the confirming commit can be sent from another session or not. Seems insecure to allow other session to do it. But in the persist example above, the confirming commit MUST be sent from another session (since the first died). [12:05:55] * Proposal: Allow other session to explicitly confirm. [12:05:57] (slide done) [12:06:17] Andy, we're on an 8 with the above title. [12:07:27] (whole slide) Title: 013: confirmed-commit (new) * If the session dies, a pending confirmed commit is reverted This, If a config change is done so that the session dies, the config change will immediately be reverted. * Proposal: add a confirmed-commit:1.1 capability which adds a parameter to the . * It is unclear (existing implementations behave differently) if the confirming commit can be sent from another session or not. Seems insecure to allow other session to do it. But in the persist example above, the confirming commit MUST be sent from another session (since the first died). * Proposal: Allow other session to explicitly confirm. [12:07:42] washam.fan joins the room [12:08:05] READ: one reason is separate is to not be forced to keep old baggage [12:09:05] Juergen is correct about lock; the candidate can be changed again without affecting confirmed commit [12:09:40] balazs.lengyel leaves the room [12:11:35] the discussion will continue on the mailing list [12:12:16] moving to "with-defaults", maybe back here after [12:12:32] http://www3.ietf.org/proceedings/75/slides/netconf-3.ppt [12:13:14] Balazs [12:13:29] slide 2 [12:14:44] slide 3 [12:15:48] slide 4 [12:16:21] slide 5 [12:16:35] comments from jabber? [12:16:40] phil [12:16:49] READ: we need to describe what happens for copy-config to the candidate (set to trim or explicit); proposal: with-defaults ignored for target= candidate [12:19:30] martin [12:22:01] wg last call? for/against? [12:22:09] READ: need to be able to control defaults sent to with copy-config [12:22:49] balazs.lengyel joins the room [12:23:01] back to martin [12:23:04] slides updated! [12:23:12] http://www3.ietf.org/proceedings/75/slides/netconf-2.pdf [12:23:16] slide 10 [12:24:14] sorry, it's slide 9 on the ietf site still [12:24:29] comments? [12:24:47] NO! already has a namespace. this breaks XML [12:25:43] please explain... [12:25:49] in request or reply? [12:26:12] in the PDU is in the default namespace. it has to be declared [12:26:53] the PDU shown is invalid XML; libxml2 will not accept it [12:28:52] The xmlns="" would need to be present if the NULL namespace is used [12:30:05] READ: deal with this in your NMS code; users don't need to enter XML cruft [12:30:21] lada [12:32:22] the proposal will go to the mailing list [12:32:43] slide "009:partial-operation error (new)" [12:34:08] agree with proposal to deprecate this error [12:35:20] slide 012: format of copy-config (new) [12:41:53] http://www3.ietf.org/proceedings/75/slides/netconf-4.pdf [12:41:58] Bob Cole [12:42:09] slide 2 of 6 [12:42:35] arneos joins the room [12:42:51] Objectives and Benefits [12:43:08] (still hearing ok?) [12:44:06] sound is OK [12:44:20] Changes from 00 to 01 [12:45:06] :verified-commit Capability [12:45:52] Example ping.yang [12:46:53] Mailing List Discussion [12:49:47] comments? [12:50:23] Bert Wijnen leaves the room: Replaced by new connection [12:50:23] Bert Wijnen joins the room [12:54:30] Bert Wijnen leaves the room [12:54:31] Bert Wijnen joins the room [12:55:16] yes -- RPCs would be used. the question is whether any standard RPCs would help [12:57:03] andy, i am in line [12:58:32] I am planning to implement; if only as proprietary improvement to the flawed confirmed-commit [12:59:25] agree verify tests should be runnable not during commit; need baseline test numbers [13:00:46] Atarashi Yoshifumi leaves the room [13:00:55] slide 7 on agenda [13:01:16] questions will go to the mailing list [13:01:36] balazs.lengyel leaves the room [13:01:45] balazs.lengyel joins the room [13:01:52] balazs.lengyel leaves the room [13:02:00] wrt: database arch ;; need partial locking to work for all databases, not just 1 [13:02:22] i think they want the comments on this list now... [13:02:25] we're done [13:02:33] bye; thanks David! [13:02:37] thanks! [13:02:45] andy.bierman leaves the room: Logged out [13:02:47] y'all be careful out there [13:02:50] ray leaves the room [13:02:56] dlpartain leaves the room [13:03:02] David Reid leaves the room [13:03:08] arneos leaves the room [13:03:22] Mark Ellison leaves the room [13:03:26] thanks david. [13:05:25] mark leaves the room [13:06:14] Juergen Schoenwaelder leaves the room: Disconnected [13:07:02] washam.fan leaves the room [13:07:38] Bert Wijnen leaves the room [13:18:15] weshardaker leaves the room: Disconnected. [13:19:12] Linus Nordberg leaves the room [13:47:09] Bert Wijnen joins the room [13:47:50] Bert Wijnen leaves the room [13:48:55] Bert Wijnen joins the room [14:40:00] clns leaves the room: offline [14:46:33] Juergen Schoenwaelder joins the room [14:49:37] Juergen Schoenwaelder leaves the room [15:57:29] Bert Wijnen leaves the room