16:01:03 #startmeeting Octavia 16:01:04 Meeting started Wed Dec 4 16:01:03 2019 UTC and is due to finish in 60 minutes. The chair is johnsom. Information about MeetBot at http://wiki.debian.org/MeetBot. 16:01:05 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 16:01:07 The meeting name has been set to 'octavia' 16:01:09 rm_work Ping 16:01:12 hi 16:01:14 Hi 16:02:14 hi 16:02:23 #link https://wiki.openstack.org/wiki/Octavia/Weekly_Meeting_Agenda#Meeting_2019-12-04 16:02:41 I did put together a quick agenda today 16:02:44 hi 16:03:10 #topic Announcements 16:03:17 I guess by default I will run today's meeting 16:03:44 The one announcement I am aware of is about some students that may be helping us out. 16:04:18 The OpenStack foundation has had students from NDSU work on OpenStack projects before. 16:04:39 sweet! 16:04:58 rm_work had a conversation about it with Kendall at the Summit/PTG and we may be getting some students to help us out. 16:05:21 As the TLS protocols and ciphers stories are pretty complete it was proposed that they could work on those tasks. 16:05:38 It sounds like it's about four students in their senior year. 16:06:11 Any questions / concerns about this? 16:06:46 how can we help them onboarding? 16:06:52 Otherwise I hope we welcome them and give them a good OpenStack experience. 16:07:29 I think Kendall will do the initial onboarding (setup gerrit and OSF accounts, etc. I will volunteer to give the Octavia intros. 16:07:43 thank you 16:08:44 That was all I had for announcements today. Pretty sure I'm forgetting something. Does anyone else have anything? 16:09:19 There are some proposed infra team changes: 16:09:21 #link http://lists.openstack.org/pipermail/openstack-infra/2019-December/006537.html 16:09:36 But that is mostly an FYI and should not impact day-to-day 16:10:32 #topic Brief progress reports / bugs needing review 16:11:08 I was out for a few days due to the US holiday. Otherwise it has been working on reviews, bug fixes, and making progress on the failover work. 16:11:30 Big thatnks to rm_work and johnsom for review of refactor jobboard changes! 16:11:57 We are starting to see a number of issues with the transition between python2 and python3 in the master gates. We have been trying to work through those as we see them. 16:12:07 I fixed a DIB issue last night for example. 16:13:00 But if we spoke of ciphers I will highlight the one that we have on review already 16:13:06 #link https://review.opendev.org/#/c/685337/ 16:13:34 could be starting point for students as well 16:13:47 Yep, it doesn't look like that is on the priority review list. I will add it. 16:14:41 Yes, having a default is a head start on that one for sure. 16:15:47 yeah, now with py2 jobs removed we started seeing issues also in centos 7 amphorae. kuryr team reported the amp no longer works. 16:16:54 I posted a patch to build centos 8 amps. it seems to be working well on a couple of rechecks and maybe even faster than centos7 16:16:58 Is someone looking into that? There should be no problem continuing to build py2 centos 7 images for stable 16:17:11 Faster would be super good 16:17:32 the py2 libs were removed so... 16:17:51 haleyb started working on centos7-py3 support 16:18:04 while we don't have centos8 controllers 16:18:10 Ok 16:18:11 cgoncalves: the building of the py2 libs? that was a setup.cfg change we can maybe put back 16:18:18 it would be nice if we could have a centos7 controller + centos8 amp job for now 16:18:56 the centos7 py3 job is not happy, https://review.opendev.org/#/c/697198/ - seems python3 isn't installed by default 16:19:11 last 3 centos8 amp scenario jobs completed in 1h49, 1h40, 2h07 16:19:27 so i either start digging into devstack or we get centos-8 working 16:20:02 centos 8 controller isn't available yet. I know of some folks working on that, there are WIP patches in Gerrit 16:20:17 cgoncalves: getting back to the py2 libs being removed - that was octavia or octavia-lib ? 16:20:25 #link https://review.opendev.org/#/c/688614/ 16:20:44 haleyb, octavia at least. 16:21:21 I don't think py2 removal has merged in octavia-lib yet 16:21:29 kuryr folks reported the amphora-agent started crashing on centos 7. I'd have to scroll up to find the exact reason 16:21:51 we could have been too aggressive - if there was a py2 consumer we shouldn't have broken them. i learned the upstream goal was to remove the testing this cycle, more than that could have caused this 16:22:35 This is the master branch, so I think some transition issues are expected. 16:22:56 +1, as long as it happens early in the cycle which is the case here ;) 16:23:04 +1 to that! 16:23:40 #link https://review.opendev.org/#/c/697128/ 16:23:47 ^ centos 8 amphora patch 16:23:59 FYI, I kept compatibility with centos 7 amphorae 16:24:33 so is there anyone/anything consuming octavia as a library? if so starting python2 removal (like six library) should wait until next cycle 16:24:37 Frankly, some py2/py3 combinations running DIB will break until this merges: 16:24:39 #link https://review.opendev.org/697211 16:24:49 haleyb No 16:25:11 Only octavia-lib, Octavia should never be imported 16:25:28 burn the boats i say then :) 16:25:46 http://codesearch.openstack.org/?q=octavia%3E%3D&i=nope&files=&repos= 16:26:16 "should never" != "is not" 16:26:18 I think they were/are fixing that 16:26:30 haleyb Sadly true 16:27:19 I don't think that is a blocker for us though, we should still move forward 16:27:34 +1 16:28:15 ack, i'll try and split it into many patches to make it reviewable (and revertable) 16:28:59 Any other updates today? 16:30:17 So this next topic I wish our PTL was present.... 16:30:29 #topic OVN driver under Octavia project governance 16:30:54 There were a few patches proposed this week to move the OVN provider under the Octavia project. 16:31:41 I don't think that was proposed or discussed with the Octavia team prior to those patches. 16:32:24 I don't remember that being proposed or discussed either 16:32:39 There were a number of -1 votes on those. I think there are a number of concerns about that and it should be discussed, but without the PTL here I'm not sure we should jump into that. 16:33:02 I can apologize for that, think it was just a decision based on the progress of the networking-ovn code movement 16:34:03 haleyb Was there any push back from the neutron team in retaining ownership of that code? 16:34:13 it was basically an orphan, and we can hold any merge until we're able to discuss more, think it's on the neutron-drivers menu this week as well 16:34:33 If there is, maybe we should add this to next week's agenda and make sure that rm_work can join. 16:36:33 i know there are no in-tree drivers, seemed odd to consider this a third-party driver since it is associated with neutron and not a third-party company 16:37:27 in the end i don't think the neutron team will not accept it 16:37:36 I propose we table this topic until rm_work (PTL) can be present. Is that ok or is this time sensitive? 16:38:27 we can wait until rm_work is around, i think having an answer in a week is good enough 16:38:58 Ok, thanks. I'm just concerned that we don't have quorum at the moment. 16:39:08 #topic Open Discussion 16:39:33 Other topics this week? 16:41:51 I think the two patch chains folks are focused on reviewing this week are the AZ patches and the jobboard chain. Thanks for all of the help reviewing those! Great progress on both fronts. 16:42:24 It will be nice to have jobboard in early in Ussuri as it was one of our priorities for Ussuri from the PTG. 16:43:07 ataraday_, thanks a lot for all the great work and sticking around despite low review attention in past months 16:43:48 +100 16:44:06 cgoncalves, johnsom Thanks! Though it is not finished yet :) 16:44:18 If there are not any more topics this week I will close out the meeting. 16:44:41 #endmeeting