15:00:22 #startmeeting neutron_northbound_2015_03_13 15:00:22 Meeting started Fri Mar 13 15:00:22 2015 UTC. The chair is regXboi. Information about MeetBot at http://ci.openstack.org/meetbot.html. 15:00:22 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 15:00:22 The meeting name has been set to 'neutron_northbound_2015_03_13' 15:00:34 #chair flaviof, edwarnicke 15:00:34 Current chairs: edwarnicke flaviof regXboi 15:00:42 #topic agenda bashing 15:00:51 #link https://wiki.opendaylight.org/view/NeutronNorthbound:Meetings agenda 15:01:21 folks, take a look at the agenda - is there anything else we need to cover? 15:01:56 * dfarrell07|sick is here 15:02:23 dfarrell07|sick thanks 15:02:33 regXboi: LGTM 15:02:35 #info flaviof 15:02:52 mestery: thanks - will give others a few minutes 15:03:16 yes, i'm here 15:03:45 if there are no changes, we will move to action items from previous meeting... 15:04:01 #topic cover action items 15:04:19 flaviof: do you have the links for the emails on how to migrate? 15:04:33 * regXboi knows you sent them, but was wondering if we can point to links 15:04:35 #info flaviof checked all uses of neutron and sent email 15:05:28 regXboi: I made all the changes needed instead of showing how to do it. So I have gerrits, listed in dev-neutron 15:05:56 ok, we'll see about linking that on the project page 15:06:02 can I ask you to do that? 15:06:26 regXboi: link to the email I sent? 15:06:40 yes - add that to the Lithium project page (somewhere) 15:06:48 dfarrell07|sick: weren't you instructed to get well 15:06:49 regXboi: np 15:06:55 flaviof: thanks 15:07:20 #action flaviof to add link in Lithium project page on the info for migrating projects to the new neutron repo 15:07:24 edwarnicke: lol, working on it ;) 15:07:26 you beat me to it 15:07:30 ok next item 15:07:40 #link https://wiki.opendaylight.org/view/NeutronNorthbound:Lithium_Release_Plan#User_Facing_Karaf_Features_List <- list of karaf features 15:07:47 that look correct edwarnicke? 15:08:01 regXboi: yep :) 15:08:20 ok... next 15:08:31 #link https://git.opendaylight.org/gerrit/#/c/16340/ <- initial document commit 15:08:44 IIRC I listed that on the project page 15:09:10 and 15:09:33 #link https://git.opendaylight.org/gerrit/#/c/16155/ <- dummy provider code 15:09:40 I think we can try writing test cases to that 15:10:31 #link https://git.opendaylight.org/gerrit/#/c/15989/ <- yang model 15:10:37 we'll talk about this more later 15:10:38 regXboi: Looks good to me :) Mind if I merge 15:10:53 edwarnicke, we should change the comment first? 15:11:23 regXboi: I was referencing the test dummy 15:11:32 edwarnicke: I know 15:11:41 the comment on that patch is "DO NOT MERGE" 15:11:57 Ryan Moats proposed a change to neutron: Dummy provider for neutron testing https://git.opendaylight.org/gerrit/16155 15:12:05 and that change should do the trick 15:12:36 so once that verifies - sure we can +2 and merge it 15:13:02 back to the yang model (action item #5) I consider that closed 15:13:08 as we are now breaking the model apart 15:13:12 regXboi: ? 15:13:20 regXboi: The yang model still needs a bit of workj 15:13:30 yes, and we have a topic for it 15:13:40 Doh 15:13:43 the action item was "to add all the extensions into a single model" 15:13:43 read to fast 15:13:55 I consider that action item closed because we are breaking it apart :) 15:14:06 Is kkoushik_ around ? 15:14:09 yes, i'm here 15:14:23 kkoushik_: Cool, do you see the two comments to your updates? 15:14:36 before we get to that - we have one more action item :) 15:14:39 i've addressed most of the comments from patch6 in patch7 15:14:46 ok, i'll wait 15:14:47 #link https://wiki.opendaylight.org/view/NeutronNorthbound:Meetings <- meeting wiki page 15:14:53 #topic Yang Model 15:15:05 kkoushik_ -> ok *now* :) 15:15:19 kkoushik_: There are two new ones to patch 7 :) 15:15:38 kkoushik_: I commented asking why we don't just use the standard yang uuid (example of its use here: https://git.opendaylight.org/gerrit/gitweb?p=ovsdb.git;a=blob;f=southbound/southbound-api/src/main/yang/ovsdb.yang;h=ac27f72709aec7bbb15c4e75f90f5807c2415f8a;hb=HEAD ) 15:15:41 ok - all the comments except breaking up the model into more pieces + new comments on patch7 have been addressed 15:15:56 I tried that - but it won't compile in the neutron ws 15:16:19 not sure where/how to bring the latest model in which has the uuid: definition 15:16:36 edwarnicke: can you help out after we are done? 15:16:37 kkoushik_: And Wojiech commented about binding host_id not being a uuid 15:16:47 kkoushik_: That's why I pointed you to the example :) 15:16:48 well 15:16:51 kkoushik_: https://git.opendaylight.org/gerrit/gitweb?p=ovsdb.git;a=blob;f=southbound/southbound-api/src/main/yang/ovsdb.yang;h=ac27f72709aec7bbb15c4e75f90f5807c2415f8a;hb=HEAD does it 15:16:52 i'll address that 15:16:57 well 15:17:05 the binding items all need to go into an extension 15:17:21 as does the router:external in networks (second extension) 15:17:22 ed: i'll fix that - thx 15:17:38 regXboi: agreed on the new extension models 15:17:45 kkoushik_: thanks 15:17:53 regXboi: will get to them by monday 15:18:02 note: the router:external extension also adds the router and floatingip models 15:18:10 because that is all part of the L3 extension 15:18:26 so I'd suggest calling that one neutron-L3.yang 15:18:38 regXboi: ok - that makes sense 15:18:38 or something like that 15:19:11 i'm having trouble replying to comments on gerrit webpage(they get marked as "Draft") 15:19:29 not sure how to publish them for public visibility 15:19:31 IIRC, that's the case until you "post" a review 15:19:31 kkoushik_: You have to submit your 'Reply' for them to go in 15:19:47 so on the yang model, here's what I see that we still need: 15:20:05 neutron-security-groups (security groups and rules) 15:20:26 neutron-metering (metering labels and rules) 15:20:40 neutron-lbaas (LBaaS v2.0) 15:20:51 mestery: am I missing anything? 15:21:25 * regXboi wondering about vpnaas 15:23:03 vpn 15:23:05 lets do that too 15:23:10 pointer? 15:23:17 to what? The API doc? 15:23:33 will that be part of Kilo? 15:23:36 mestery: Think we could get Prem to do VPN ;) 15:23:47 if so, then pointers to the API and the datamodel 15:23:48 It's been a part of Havana, Icehouse and Juno, non-experimental, so yes. 15:24:08 interesting - it's not documented on the Networking API page 15:24:35 * edwarnicke would gets how hard documentation is 15:24:52 s/would// 15:24:56 ok, so we will need links to the API and the data model 15:25:08 but yes, we need to add neutorn-vpnaas (VPNaaS) as well 15:25:35 regXboi: Should we loop in the VPN Services guys to work on that? 15:25:37 * regXboi admits that he still needs to do attribute level checking 15:25:43 regXboi: I bet Prem would be all over it :) 15:26:25 edwarnicke: if they are willing to commit both yang model and data model/Aware interface/Dummy Provider code, sure 15:26:42 regXboi: Question on that point 15:26:51 but they don't get to cherry pick the yang model 15:26:57 edwarnicke: ask away 15:27:02 regXboi: Since we do not currently have the Aware for VPN Services, do we necessarily need it added? 15:27:14 for consistency's sake - yes 15:27:23 regXboi: My understanding is we were maintaining the Aware stuff to ease migration 15:27:30 which I wholeheartedly support 15:27:36 But in this case, there is no migration 15:27:56 It feels strange to introduce an API we already intend to deprecate in Be 15:27:59 as we are not requiring folks to switch to yang in this release, I can't ask a project to support both I*Aware and Yang 15:28:10 regXboi: Fair 15:28:13 so, for example 15:28:31 I'm thinking of folks like OVSDB and VTN here 15:28:43 if they want VPNaaS, we'd leave them a bit out in the cold 15:28:58 but yes, it *does* look odd providing a new interface that we are going to deprecate in Be 15:29:13 regXboi: Might it be worth asking if they intend to do anything with that in Li ? 15:29:26 regXboi: I think the only folks looking to do stuff in Li with VPN are VPN Services (though could be mistaken) 15:29:32 edwarnicke: you've got the action item :) 15:29:44 * edwarnicke should learn to keep his mouth shut ;) 15:30:03 #action edwarnicke to poll users of neutron to see if any (other than VPN services) are looking to use Neutron VPNaaS in the Li timeframe 15:30:16 hey, if you get a no answer, then we just go Yang on that one 15:30:33 regXboi: Fair :) 15:30:50 edwarnicke: can I also ask you to approach Prem about providing the initial neutron-vpnaas yang model? 15:30:56 regXboi: Shall I send email to neutron-dev, vpn-dev, and disucss ? 15:30:56 you know him better than I 15:31:03 regXboi: Happy to 15:31:10 regXboi: And my guess is he will be excited to do so :0 15:31:11 :0 15:31:13 :) 15:31:16 * edwarnicke can't type today 15:31:32 I'd be fair - I'd send the email to the projects that flaviof submitted patches to :) 15:32:16 #link https://lists.opendaylight.org/pipermail/neutron-dev/2015-March/000048.html list of neutron dependencies 15:32:24 so more than just neutron, vpn and discuss 15:32:30 regXboi: Yep 15:32:48 anything else on Yang for now or can we move on to testing? 15:33:02 obviously, yang stays on the agenda for next week :) 15:33:11 doh 15:33:30 regXboi: Sounds good... next topic :) 15:33:36 wait - need infos :) 15:33:51 #info we still need neutron-security-groups (security groups and rules) 15:34:02 #info we still need neutron-metering (metering labels and rules) 15:34:10 #info we still need neutron-lbaas (LBaaS v2.0) 15:34:14 ok 15:34:18 #topic testing 15:34:38 first item here - we need to finish the M3 testing item, which is ETA today 15:34:51 "Simple system test on a karaf distribution with the project's recommended features installed " is what it literally says 15:34:57 do we have a patch on intergration for that? 15:35:11 or a patch anywhere? 15:35:37 regXboi: I don't think so yet 15:35:52 do we think we can get one by COB today? 15:35:57 or do we slip this to red :/ 15:36:05 regXboi: I think for us that would consist of something which loads the dummy provider and simply checks to see if we are unmarshalling all the REST correct, right? 15:36:08 regXboi: by "project's recommended features installed" do you mean feature:install odl-ovsdb-openstack 15:36:10 ? 15:36:14 regXboi: I don't think it can be done today 15:36:22 flaviof: no 15:36:40 regXboi: you mean dummy test? 15:36:51 I would think it is odl-neutron-service 15:37:01 regXboi: make sense. 15:37:04 and then the dummy provider 15:37:11 and edwarnicke is correct 15:37:20 it is a simple unmarshalling of all the REST at this point 15:37:29 flaviof: neutron has to be tested standalone 15:37:33 understood. 15:37:50 #link https://wiki.opendaylight.org/view/NeutronNorthbound:JUnitTestStrategy <- wiki page on JUnit testing strategy for more complex cases 15:37:52 flaviof: neutron is provider agnostic :) 15:38:18 we are bleeding into the second sub-topic here, which is the JUnit stuff 15:38:19 edwarnicke: indeed. 15:38:26 regXboi: I tend to prefer system tests for REST stuff to unit tests 15:38:48 edwarnicke: if you look at that page - I leave the marshall/unmarshall to system test 15:38:54 regXboi: We literally hit two issues this week on the RESTCONF migration where the underlying classes had 100% unit test code coverage, and we still had bugs because of jersey interactions 15:39:02 regXboi: Cool, read to fast ;) 15:39:03 those methods I call out are actual logic that we can test in JUnit 15:40:00 edwarnicke: refresh - I clarified 15:40:25 by stating that the majority can be tested via system test, so this page concentrates on the other (bold) 15:40:36 edwarnicke: the chatty issues on jetty.... is that something that will get fixed soon? 15:40:53 flaviof: chatty issues on jetty ? 15:41:03 * flaviof not sure if that is what edwarnicke meant by REST issues 15:41:43 edwarnicke: project's recommended features installed 15:41:54 undo that.... :) 15:42:00 edwarnicke: java.lang.IllegalStateException: No LoginService for org.eclipse.jetty.security.authentication.BasicAuthenticator@678309d4 in org.eclipse.jetty.security.ConstraintSecurityHandler@4bfa7263 java.lang.IllegalStateException: No LoginService for org.eclipse.je 15:42:28 can we take that up after we are done? 15:42:45 * regXboi doesn't want to go off and troubleshoot jetty problems here 15:42:51 regXboi: np. thought that is what we were talking about. sorry 15:43:56 not precisely, I'm talking about have a patch to integration that does a simple unmarshalling test with the dummy providers against odl-neutron-service 15:44:06 so that we can claim we are done with M3 15:44:52 edwarnicke: although now that I've typed the above, I'm thinking that we may need a separate dummy provider feature for integration to be happy :/ 15:45:43 so, who can I #action for that item :) 15:45:53 * regXboi waits for the crickets ..... 15:46:07 that item == the integration patch 15:47:24 * regXboi listens to the crickets 15:48:44 flaviof: as test contact, can I look to you to do set that up? 15:50:07 did everybody just run away? 15:50:18 * dfarrell07|sick is still here 15:50:55 * regXboi watches the time for documentation run away 15:51:09 ping flaviof 15:52:17 regXboi: sorry, here 15:52:17 well...lol 15:52:25 ah cool :) 15:52:38 regXboi: dfarrell07|sick: I'm busy with mdsal work on ovsdb; would prefer to defer that action to dfarrell07|sick if possible 15:53:07 dfarrell07|sick are you up to doing that ? 15:53:48 I think it's pretty out of scope for my current TODOs, would be very non-trivial for me to get spun on that much NNB project detail 15:54:06 dfarrell07|sick - that is why I asked 15:54:30 dfarrell07|sick - do you know how integration sets up the robot framework for testing? 15:54:59 i.e. are you willing to help someone else (who has the NNB project expertiese) get the intergration patch in? 15:55:02 regXboi: to be honest, I know the basics but most of my Integration work hasn't been in that domain (integration with outside projects/tools) 15:55:16 regXboi: yes, I can help muster Integration team help 15:55:25 dfarrell07|sick: ok that's fair 15:56:39 #action regXboi to put together integration patch (with resource help from dfarrell07|sick) for REST unmarshalling testing of odl-neutron-service with dummy provider. 15:56:51 * regXboi sighs 15:57:00 ok, next topic 15:57:15 #topic documentation 15:57:51 #link https://git.opendaylight.org/gerrit/#/c/16340/ <- intiial documentation link 15:58:05 edwarnicke, flaviof: I'm going to action you both to review 15:58:24 and add comments about fleshing out the two parts 15:58:53 #action edwarnicke, flaviof to review initial documentation patch and provide comments/structure for fleshing out the two sides 15:59:06 anything else? 15:59:15 otherwise... 15:59:24 * dfarrell07|sick has nothing else 15:59:27 going once? 15:59:28 cookies? 15:59:44 edwarnicke, flaviof: going twice? 16:00:06 sold... 16:00:09 #topic cookies 16:00:12 #endmeeting