21:00:08 <mlavalle> #startmeeting networking 21:00:09 <openstack> Meeting started Mon Jul 22 21:00:08 2019 UTC and is due to finish in 60 minutes. The chair is mlavalle. Information about MeetBot at http://wiki.debian.org/MeetBot. 21:00:10 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 21:00:13 <openstack> The meeting name has been set to 'networking' 21:00:13 <njohnston> o/ 21:00:58 <igordc> hi 21:01:42 <mlavalle> slow mid-summer meeting? 21:01:57 <slaweq> hi 21:01:58 <mlavalle> I know haleyb|away is off on PTO 21:02:00 <njohnston> It would seem so! 21:02:08 <njohnston> bcafarel is also on PTO 21:02:44 <mlavalle> yeah, Spaniards, French and Germans close the entire country this time of the year and all go to the beach 21:03:15 <mlavalle> That's what I call an advanced civiliazation :-) 21:03:54 <mlavalle> ok, enough of goofing around 21:04:02 <mlavalle> #topic Announcements 21:04:50 <mlavalle> This is the week of the Train-2, per: 21:04:52 <mlavalle> https://releases.openstack.org/train/schedule.html 21:05:48 * njohnston blows his Train Whistle 21:06:16 <mlavalle> also, we have a lanning etherpad for the Shanghai PTG: 21:06:28 <mlavalle> https://etherpad.openstack.org/p/Shanghai-Neutron-Planning 21:06:43 <mlavalle> we can keep adding topics to the pad until November. 21:07:27 <njohnston> I am not planning on being there but I am going to timeshift myself and try to participate as much as possible remotely 21:07:33 <mlavalle> But I ask folks planning to be there to please add your names and irc nicks in the "Attendees" section, so we can help the Foundation plan for meeting rooms 21:08:26 <mlavalle> it's good you mention it. That means that we need to plan to have some sort of video feed, like we've done in the past 21:08:34 <njohnston> +1 21:08:38 <tidwellr> +1 21:09:10 <slaweq> mlavalle: I hope it will be possible this time 21:09:29 <slaweq> mlavalle: I never was in China but I heard that there may be some issues with access to some services 21:09:37 <mlavalle> It is more than just having a camera attached to a laptop or a google hangout 21:10:12 <mlavalle> because we need to take the Chinese Great Firewall into consideration 21:10:44 <mlavalle> I'll talk liuyulong to start testing an approach 21:11:17 <mlavalle> any other announcements? 21:12:29 <mlavalle> ok, let's move on 21:12:39 <mlavalle> #topic Blueprints 21:13:03 <mlavalle> This our dashboard for the T-2 milestone: https://launchpad.net/neutron/+milestone/train-2 21:13:21 <mlavalle> any blueprints updates? 21:13:56 <tidwellr> I just updated https://review.opendev.org/#/c/667511/ with some additional tests to cover prefix delegation cases, I think it's ready to go 21:14:38 <mlavalle> I was about to ask. Thanks for the update 21:15:50 <mlavalle> I can see that estraroutes atomic has been busy: https://review.opendev.org/#/q/topic:extraroute+(status:open+OR+status:merged) 21:15:59 <mlavalle> although it seems a little green to me 21:16:07 <mlavalle> I mean for T-2 21:17:20 <mlavalle> and neutron-classifier seems to be the same 21:17:31 <mlavalle> https://review.opendev.org/#/q/topic:qos-classifier+(status:open+OR+status:merged) 21:18:00 <mlavalle> any other updates on blueprints? 21:18:47 <mlavalle> btw, njohnston I am going to create a blueprint for the non ip ethertypes and target it for T-3. are ok with that? 21:19:38 <njohnston> mlavalle: Sure thing 21:19:43 <mlavalle> cool 21:19:54 <mlavalle> ok, moving on 21:20:02 <mlavalle> #topic Community goals 21:20:17 <mlavalle> any updates on community goals 21:20:18 <mlavalle> ? 21:21:14 <slaweq> gmann recently started pushing some patches related to IPv6 testing goal 21:21:16 <slaweq> https://review.opendev.org/#/c/671899/ 21:21:33 <slaweq> I will take care of review of them 21:22:14 <mlavalle> added to my pile 21:22:58 <njohnston> note that there are both neutron and neutron-fwaas changes. fwaas change is: https://review.opendev.org/671907 21:23:46 <mlavalle> in my pile as well 21:25:57 <mlavalle> #topic Bugs 21:26:44 <mlavalle> This is last weeks deputy report for last week: http://lists.openstack.org/pipermail/openstack-discuss/2019-July/007952.html 21:26:52 <mlavalle> Thanks hongbin 21:27:33 <mlavalle> it seems it was more or less a quite week 21:27:55 <mlavalle> only 1 high piority bug: https://bugs.launchpad.net/neutron/+bug/1836642 21:27:56 <openstack> Launchpad bug 1836642 in neutron "Metadata responses are very slow sometimes" [High,Confirmed] - Assigned to Slawek Kaplonski (slaweq) 21:28:31 <slaweq> but this one is IMHO more an nova issue rather than neutron bug 21:28:54 <njohnston> That is a very interesting bug. I had no idea that before Nova could respond to a metadata request it would query Neutron, more than once. 21:30:03 <mlavalle> in the work I am doing with the dvr multinode job, I can see this happening 21:30:10 <slaweq> yes, and I think that we will need some help from nova's side with that one 21:30:39 <mlavalle> a lot of VMs, even if eventually succesful, have to query metadata several times 21:30:54 <mlavalle> there is always 2 or 3 retries 21:31:19 <slaweq> mlavalle: problem in CI is that script which gets metadata used on cirros image don't retry if e.g. call for public-keys fails 21:31:40 <slaweq> it will only retry in case of /instance-id/ request will fail (the first one) 21:31:53 <mlavalle> yes, that's what I meant 21:32:49 <mlavalle> but sean-k-mooney seems to be paying attention, right? 21:32:59 <slaweq> and I really don't understand why nova needs to ask neutron for ports and security groups when instance is asking for /public-keys/ 21:33:38 <slaweq> sean did some initial look at it but I don't have feeling that he really things that it's something what can/should be improved on nova's side :/ 21:34:15 <slaweq> he rather said that it's some slowdown sometimes on ci nodes and we can't do too much about it... 21:34:22 <slaweq> that's my impression at least 21:35:09 <mlavalle> if by end of this week we don't get attention, I can ping efried 21:35:22 <mlavalle> would you please remind me? 21:35:31 <slaweq> mlavalle: sure :) 21:36:07 <mlavalle> any other bugs we should discuss today? 21:37:00 <mlavalle> ok, moving on 21:37:07 <mlavalle> #topic neutron-lib 21:37:21 <mlavalle> boden is not in attendance 21:38:23 <mlavalle> he only wanted me to mention that neutron-classifier is joining the list of neutron-lib-current projects 21:39:07 <mlavalle> anything else on neutron-lib? 21:39:45 <mlavalle> btw, boden is the bugs deputy this week and he mentioned he is on it 21:39:55 <mlavalle> #topic On demand agenda 21:40:09 <mlavalle> Anything else we should discuss today? 21:40:22 <slaweq> one quick thing 21:40:32 <slaweq> I proposed new release of neutron-tempest-plugin https://review.opendev.org/#/c/671994/ 21:40:44 <slaweq> mlavalle: and amotoki please take a look at it if You will have some time 21:40:48 <slaweq> thx :) 21:40:59 <clarkb> I wanted to mention that neutron not running the common tempest-full(-py3) job potentially opens the gate up to breakage (basically neutron makes changes that work in its gate jobs but not on the nova/cinder/glance/etc job) 21:41:14 <mlavalle> Done 21:41:15 <clarkb> I don't necessarily think this is a problem until something goes wrong, but wanted to make suer you all were aware of that being a possibility 21:41:35 <slaweq> clarkb: but this was discussed with QA team in Denver 21:41:56 <clarkb> sure I'm not saying it is wrong. I'm saying it is a risk and I wasn't sure if people were aware of the risk 21:42:10 <slaweq> and we are now switched to run tempest-integrated-networking job instead of tempest-full 21:42:21 <mlavalle> we are aware, but thanks for the heads up anyways 21:43:34 <mlavalle> slaweq: I +1ed the neutron-tempest-plugin release patch 21:43:42 <slaweq> mlavalle: thx a lot 21:43:53 <mlavalle> anything else we should discuss today? 21:44:01 <tidwellr> not sure if this is the right forum, perhaps the drivers meeting is better. I wanted to see if we can do something about floating IP's for routed networks, the solution is look rather simple but some agreement on an approach would be needed. For reference - https://review.opendev.org/669395 and https://review.opendev.org/486450 21:44:26 <mlavalle> This is the correct forum 21:44:37 <mlavalle> I'll take a look over the next couple of days 21:44:39 <tidwellr> this would really round out the routed networks functionality in train 21:45:13 <mlavalle> in my pile 21:45:16 <slaweq> tidwellr: I will also take a look 21:45:51 <tidwellr> thanks, seeing as we're closing in on feature freeze the sooner we get some consensus the better chance we have at landing something 21:46:31 <mlavalle> ok, in keeping with the summer vacation spirit that seems to prevail, , we return 15 minutes to your agendas 21:46:37 <mlavalle> have a great week! 21:46:40 <mlavalle> #endmeeting