16:02:24 <yamahata> #startmeeting neutron_northbound 16:02:24 <odl_meetbot> Meeting started Mon Nov 21 16:02:24 2016 UTC. The chair is yamahata. Information about MeetBot at http://ci.openstack.org/meetbot.html. 16:02:24 <odl_meetbot> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 16:02:24 <odl_meetbot> The meeting name has been set to 'neutron_northbound' 16:02:30 <yamahata> #chair mkolesni 16:02:30 <odl_meetbot> Current chairs: mkolesni yamahata 16:02:36 <yamahata> #topic agenda bashing and roll call 16:02:49 <mkolesni> #info mkolesni 16:02:55 <yamahata> #info yamahata 16:03:12 <yamahata> from me, I'd like to summarise neutron stadium effort and Ocata plan 16:03:21 <yamahata> In addition to usual topics 16:03:28 <mkolesni> sure 16:03:32 <yamahata> hello john_a_joyce 16:04:28 <john_a_joyce> hello 16:04:41 <mkolesni> hi john_a_joyce 16:04:53 <yamahata> we've just started. Now agenda bashing call. 16:05:02 <yamahata> do you have any topic today? 16:05:14 <john_a_joyce> hi mkolesni 16:05:24 <john_a_joyce> nothing specific for me to bring up 16:06:19 <yamahata> okay usual topics, neutron stadium, and ocata planning. 16:06:26 <yamahata> move on 16:06:32 <yamahata> #topic Announcements 16:06:47 <yamahata> Ocata 1 was Nov 14 and ocata-1 assessment for networking-odl was done. 16:06:57 <yamahata> The final summary is beind reviewed. 16:07:11 <yamahata> #link https://review.openstack.org/#/c/389397/ Ocata: Assessment summary 16:07:20 <yamahata> It's not finalised yet. 16:07:37 <mkolesni> i think in terms of docs were half way there 16:07:52 <mkolesni> better than nothing, but its very bare bones at the moment 16:08:14 <yamahata> mkolesni: agreed. 16:08:47 <yamahata> the assessment of networking-odl is merged 16:08:56 <yamahata> #link https://review.openstack.org/#/c/383910/ Ocata: Assessment for networking-odl 16:09:13 <yamahata> The sentense of "networking-odl will be removed from neutron stadium" was removed 16:09:41 <yamahata> So I suppose we're out of "emergency" status. 16:09:57 <mkolesni> seems so 16:09:59 <yamahata> anyway we need to continue to fill those gaps. 16:10:16 <mkolesni> agreed 16:10:28 <mkolesni> im currently working on functional tests as you know 16:10:38 <yamahata> So far I have intentionally merged patches very aggressively to make progress without other's reviews. 16:11:11 <yamahata> #topic neutorn stadium 16:11:29 <yamahata> But the pace could be slowed down to wait for reviews. 16:11:52 <mkolesni> i think so 16:12:05 <mkolesni> im doing very basic functional tests 16:12:18 <mkolesni> do you know if anyone would like to add some more complex ones? 16:12:30 <yamahata> So far I'm not aware of anyone. 16:13:18 <mkolesni> also in that regard 16:13:26 <yamahata> We have 4 or 5 items to fill gaps. 16:13:28 <mkolesni> do you know whats happening with this patch? 16:13:30 <mkolesni> https://review.openstack.org/#/c/394476/ 16:13:43 <mkolesni> to add functional job to the CI 16:14:01 <mkolesni> #link https://review.openstack.org/#/c/394476/ add non-voting functional tests job for networking-odl 16:14:10 <yamahata> He seems like not so active now. 16:14:20 <yamahata> we should take it over. 16:14:25 <yamahata> mkolesni: do you want to take it? 16:14:31 <mkolesni> hes from Intel, i thought maybe you know 16:14:49 <mkolesni> i dont mind to push it forward, we need the tests to start running 16:15:09 <yamahata> thanks. 16:15:38 <yamahata> he's not dedicatedly assigned to networking-odl. but generally on CI. 16:16:21 <mkolesni> ok dont want him to get offended if i take over that patch 16:17:50 <yamahata> Okay. the followings are items for neutron stadium. 16:18:09 <yamahata> 1. documentation: not critical now. but we need more 16:18:33 <yamahata> 2. stable/branch maintenance. there are several patches under review. those need to get merged 16:18:49 <yamahata> 3. functioal tests: to have CI (and more tests) 16:19:08 <yamahata> 4. fullstack tests: to have CI (and more tests) 16:19:16 <yamahata> 5. mutlinode tests: to have CI 16:19:25 <yamahata> 6. gerenade test: to have CI. 16:19:29 <mkolesni> im not sure what we need fullstack for 16:19:36 <mkolesni> do you have an idea? 16:20:05 <mkolesni> i think functional should be enough 16:20:34 <yamahata> Ideally, we can have test cases that check network connectivity with only neutron, ODL without creating readl VM, but with netns. 16:21:15 <yamahata> Given that ODL netvirt team has its own CSIT, its priority is not so high. 16:21:26 <yamahata> Oh I forgot 16:21:40 <yamahata> 7. make tempest test(scenario tests) run and pass 16:21:57 <mkolesni> i dont know since ODL already has CSIT what value do we get from fullstack 16:22:08 <mkolesni> i think its more trouble than its worth, for now 16:22:16 <yamahata> Right now we have test with old netvirt, but we should migrate to new netvirt(vpnservice) 16:22:18 <mkolesni> but we can always decide we need some in the future 16:22:35 <yamahata> mkolesni: yeah. 16:22:59 <mkolesni> also what does multi node mean? 16:23:39 <mkolesni> i.e. we have Neutron on one node and ODL on another? 16:23:53 <mkolesni> 911651 16:24:03 <yamahata> we would have control node and compute only node 16:24:18 <yamahata> Ooops, all-in-one node and compute node. 16:24:29 <yamahata> #link https://review.openstack.org/#/c/397093/ add odl-server service in devstack for networking-odl multinode job 16:25:06 <yamahata> It can test e.g. live migration. 16:26:14 <yamahata> any other questions? 16:26:24 <mkolesni> so basically its tempest tests for aio + compute node? 16:27:17 <yamahata> aio? 16:27:33 <mkolesni> all in one 16:27:37 <yamahata> Yes. 16:27:43 <mkolesni> ok makes sense 16:27:50 <yamahata> Some scenario tests requires multiple nodes. 16:28:43 <yamahata> As we have added more scenario tests, I noticed that tempest CI hits timeout(120min) 16:29:14 <yamahata> gate-tempest-dsvm-* takes about 1h30-50min 16:29:35 <yamahata> So we need to consider a way to run tempest in pararell mode. 16:29:52 <mkolesni> how is it done in other projects? 16:30:12 <yamahata> Normally it's run in parellel mode. 16:30:18 <yamahata> e.g. neutron, ovn, midonet 16:30:34 <yamahata> I tested it for networking-odl with https://review.openstack.org/#/c/399986/ 16:30:39 <yamahata> It seems unstable. 16:30:51 <yamahata> Probably we should attack it after migrating to new netvirt. 16:31:09 <yamahata> execution time reduces to around 40mins 16:31:31 <mkolesni> seems reasonable 16:32:23 <yamahata> any items regarding to neutron stadium? 16:33:28 <mkolesni> none from me 16:33:31 <yamahata> next topic 16:33:36 <yamahata> #topic ocata planning 16:33:49 <yamahata> at first, let me summarise the status of newton 16:34:00 <yamahata> The 3.0.0 was released, but stable/newton branch isn't cut yet. 16:34:12 <yamahata> #link https://review.openstack.org/#/c/398624/ 16:34:12 <yamahata> networking-odl: Release networking-odl newton/3.1.0 16:34:23 <yamahata> I'm asking infra team to cut it. 16:34:35 <yamahata> Hmm, it get -1. I'll respin it. 16:34:40 <mkolesni> i think that a bit problematic some code already is there that is for ocata only 16:34:45 <yamahata> #action respin https://review.openstack.org/#/c/398624/ 16:35:21 <yamahata> so do you have any specific git changeset/hash we should use? 16:35:41 <mkolesni> so they either should cut branch at earlier hash or we will need to revert some changes 16:36:22 <yamahata> For me either way is okay. mkolesni you have an opinion. Can you please drive it? 16:37:19 <mkolesni> will release team accept it? 16:37:49 <yamahata> For the patch of 398624, I'll respin it based on hash you'll give. 16:38:10 <yamahata> After that, we would need adjustment patches to stable/newton. 16:38:30 <mkolesni> i dont mind which has you use for 3.1 16:38:41 <yamahata> Okay, then I'll update the patch. 16:38:43 <mkolesni> just pointing out the fact that some reverts will be necessary 16:40:22 <yamahata> Gotcha. 16:41:11 <yamahata> For Ocata planning, I have the following items in mind. 16:41:27 <yamahata> 1. neutron stadium effort continued 16:41:56 <yamahata> 2. CI migration to new features(new netvirt, v2 driver, pseudo agent port binding) 16:42:09 <yamahata> 3. more on v2 driver 16:42:17 <yamahata> any other items? 16:42:42 <yamahata> Ocata cycle is short than usual. 16:42:44 <mkolesni> specifically id like to try move the dependency calculations 16:42:55 <yamahata> #link https://releases.openstack.org/ocata/schedule.html Ocata Release Schedule 16:43:01 <yamahata> mkolesni: Great. 16:43:54 <mkolesni> so basically it gives us 2 months till feature freeze? 16:44:00 <yamahata> Yes. 16:44:07 <yamahata> So far release type of networking-odl is independnet. 16:44:15 <mkolesni> we should try to keep up with Neutron schedule this release 16:44:44 <mkolesni> its a bit poroblematic that we miss the openstack deadlines each time 16:44:45 <yamahata> it means networking-odl team has flexibility. (with effort to follow neutorn schedule) 16:45:14 <mkolesni> sure just that i think if we follow the schedule more tightly it will help with ODL adoption 16:45:17 <yamahata> Another choice is to follow openstack schdule. 16:45:46 <yamahata> It means networking-odl team needs to follow neutron release more strictly. 16:45:47 <mkolesni> right now its causing uncertainty since we dont have a release schedule 16:46:17 <yamahata> Do we want to change networking-odl release type from independent to neutron release? 16:46:44 <mkolesni> i think we should strive to keep up with the Neutron schedule as much as possible 16:47:03 <mkolesni> perhaps a 1-2 week slip is OK but we should aim to be as close as possible 16:47:11 <mkolesni> that's my opinion 16:48:28 <yamahata> I meant more formal one. 16:48:39 <yamahata> Now networking-odl is released at the page of https://releases.openstack.org/independent.html 16:49:01 <yamahata> If we follow openstack schedule, we can move to https://releases.openstack.org/newton/index.html 16:49:09 <mkolesni> lets keep it independent this cycle and see if a formal change is needed at the beginning of next one? 16:49:24 <yamahata> It mean networking-odl team really has to follow its release schedule. 16:49:37 <yamahata> mkolesni: sounds good idea. 16:49:40 <mkolesni> were all grown ups here i think we can manage :) 16:50:36 <yamahata> Yeah, we should experiment if we can manage or not. 16:50:53 <yamahata> any other topic to discuss regarding to ocata cycle? 16:51:18 <mkolesni> nothing from me 16:51:45 <yamahata> #topic action items from last meeting. 16:51:49 <yamahata> we don't have any 16:51:53 <yamahata> #topic patches/bugs 16:52:04 <yamahata> do we have any specific one to discuss? 16:52:16 <yamahata> Recently patch proposal/review is quite active. 16:52:22 <mkolesni> none from me 16:52:27 <mkolesni> seems velocity is good now 16:52:39 <yamahata> #topic open mike 16:52:53 <yamahata> this time slot seems working. 16:53:06 <mkolesni> works for me :) 16:53:22 <john_a_joyce> this slot works fine for me - although I am not that active in the project at the moment 16:53:37 <yamahata> cool 16:53:55 <mkolesni> john_a_joyce: feel free to jump in if you have time 16:54:04 <yamahata> thanksgiving, december and new year is approaching. We would like to make our schedule clear. 16:54:14 <yamahata> when to have or skip. 16:54:47 <yamahata> at least I can attend Nov 28, dec 5. 16:55:06 <mkolesni> we dont have anything special here so i dont mind 16:55:08 <yamahata> Dec 26 would be skipped. 16:55:20 <yamahata> so would Jan 2. 16:55:40 <john_a_joyce> yeah Dec 26 and Jan 2 I would not be able to attend 16:55:48 <mkolesni> Dec 12/19? 16:56:13 <yamahata> I can attend Dec 12. Not sure Dec 19 at this point. 16:56:55 <mkolesni> OK so lets raise this again on Dec 5 and well know more clearly? 16:57:04 <yamahata> mkolesni: sounds reasonable. 16:57:12 <mkolesni> cool 16:57:20 <yamahata> any other issue? 16:57:25 <mkolesni> none from me 16:57:29 <yamahata> or can we close this meeting? 16:57:35 <mkolesni> +1 16:57:42 <yamahata> thanks everyone 16:57:50 <mkolesni> thank you Isaku 16:57:56 <mkolesni> see you around 16:58:03 <yamahata> #topic cookies 16:58:07 <john_a_joyce> thanks everyone 16:58:09 <yamahata> #endmeeting