14:01:05 #startmeeting networking 14:01:06 Meeting started Tue Nov 8 14:01:05 2016 UTC and is due to finish in 60 minutes. The chair is ihrachys. Information about MeetBot at http://wiki.debian.org/MeetBot. 14:01:07 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 14:01:08 Hi 14:01:09 The meeting name has been set to 'networking' 14:01:11 o/ 14:01:15 Hi 14:01:20 hola 14:01:37 * ihrachys gives a minute for everyone to wake up :) 14:01:44 o/ :) 14:01:49 Hi 14:01:49 hi 14:01:53 hello 14:02:00 o/ 14:02:27 hi 14:02:31 hi 14:02:42 ok let's run it thru 14:02:47 it will be packed today... 14:02:51 #link https://wiki.openstack.org/wiki/Network/Meetings 14:03:14 an attentive reader may notice that the layout of the wiki page, as well as topics covered is changed from the last time we had the meeting; that's all armando's fault :) 14:03:22 some topics went away (keystonev3) 14:03:35 some transformed (neutron-lib now includes announcements about breaking changes in review) 14:03:54 o/ 14:04:02 and some were added for greater visibility (there is now an explicit no-api-downtime-upgrades subsection in blueprints section) 14:04:26 Armando believes the layout will better reflect the priorities of the team for Ocata and will help us focus on what matters 14:04:47 #topic Announcements 14:05:04 1. neutron-lib 1.0.0 is released yay! 14:05:07 #link http://lists.openstack.org/pipermail/openstack-dev/2016-November/106725.html 14:05:27 the version number (1.x instead of 0.x) does not mean the effort is complete in any way 14:05:30 \o/ 14:05:32 we have a lot of work to do 14:05:50 we have Ocata-1 in one or two weeks 14:06:04 that will, traditionally, not include anything major :) 14:06:40 also, in case you haven't noticed, we had a summit a week or two ago 14:06:59 design session chairs were asked to send recap emails to openstack-dev@ [neutron] 14:07:08 nova/neutron session: http://lists.openstack.org/pipermail/openstack-dev/2016-November/106478.html 14:07:08 neutron-server recap: http://lists.openstack.org/pipermail/openstack-dev/2016-November/106684.html 14:07:08 neutron-lib recap: http://lists.openstack.org/pipermail/openstack-dev/2016-November/106719.html 14:07:08 ops/end user session: http://lists.openstack.org/pipermail/openstack-dev/2016-November/106903.html 14:07:08 neutron-lbaas: http://lists.openstack.org/pipermail/openstack-dev/2016-November/106911.html 14:07:24 * HenryG sneaks in late 14:07:31 please review them to stay in sync with reality 14:07:36 there are probably more recaps the queue, I hope other session chairs will update the ML on their topics in due course 14:07:37 I have to send a client one... I had a nasty cold last week. I will send it tomorrow. 14:07:47 amotoki: get well! and thanks. 14:08:04 finally, the last organization announcement 14:08:21 some of you may probably know that I will change my timezone in a month to the one not too comfortable to chair the meeting 14:08:29 to anticipate that, armax decided to move the responsibility for the Tuesday meeting to jlibosva starting from the next week 14:08:43 (who hasn't showed up today; what a shame!!!) 14:08:55 he's taking a day off due to working when he wasn't supposed to 14:09:17 I trust in our fearless leader, Jakub :) 14:09:21 * ihrachys hoped the number of !!! reflected it was a joke! 14:09:24 amuller: i believe there is no such thing like "working when wasn't supposed to" in OpenSource ;) 14:09:32 :D 14:09:35 I will work with Jakub to make the transition smooth 14:09:39 there is a patch for the official change at https://review.openstack.org/#/c/394716/ in case you care to vote 14:09:56 dasm: Unfortunately code monkies are actually humans 14:09:58 congrats for everyone tired of me mumbling on the meeting :) 14:10:10 I hear humans need rest 14:10:23 amuller: rumors :P 14:10:27 ok, let's finally switch to more meaningful stuff! 14:10:36 thanks ihrachys 14:10:38 #topic Blueprints 14:10:54 #link https://launchpad.net/neutron/+milestone/ocata-1 14:11:18 I checked the list an hour ago, and I believe the list of High importance RFEs neatly reflects reality of Ocata 14:11:48 there are some Low importance RFEs in the list that may well slip over into a future release 14:12:08 there is not much to say except... 14:12:10 #topic OVO/no API downtime 14:12:41 tl;dr for those lazy to read my neutron-server recap email, we target the feature(s) to Ocata 14:12:50 at this point, we are looking at a significant number of transitioning patches that lack reviews 14:12:53 #link https://review.openstack.org/#/q/status:open+project:openstack/neutron+branch:master+topic:bp/adopt-oslo-versioned-objects-for-db 14:13:06 the link should capture most of relevant patches 14:13:06 lazy or not quick enough ;) 14:13:27 it would be wise to give those patches some attention from core and not-yet-core reviewers in next months so that we can make progress on the switch to versioned objects for database interaction 14:14:08 ajo: aren't we all superhumans here? laziness is the only viable explanation. 14:14:10 wow, that's a lot of patches 14:14:15 I understand that people are still recovering from the summit, but since the cycle is rather short, it's better not to hang on the patches for too long 14:14:26 in my case assume lazyness :) 14:14:30 ajo: right that's a lot, and we need to squash all of them and more! 14:14:49 * ajo takes good note of the URL to do more proactive reviews on that area 14:15:07 sadly some of us were and some are still distracted by other stuff, but we should get back to work, taking each and shaving it until it lands 14:15:35 on a relevant note, there is a spec up for review that captures the current thinking around how we plan to deliver no API dowtime upgrades in Ocata 14:15:41 #link https://review.openstack.org/#/q/status:open+branch:master+topic:bp/online-upgrades 14:16:04 there are no more patches for the topic except the spec atm, but except more actual code changes in near future 14:16:34 I don't have anything specific of OVO transitioning patches to point out here, it's just a matter of working on landing those patches in queue at this point 14:17:12 in next meetings, we may run thru specific patches that may be blocking progress; for now, moving on 14:17:18 #topic Bugs and Gate failures 14:17:44 before we bug our bug deputy (pun!) and look at the gate, I'd like to mention something 14:17:47 #link http://lists.openstack.org/pipermail/openstack-dev/2016-November/106906.html 14:17:56 tl;dr ^ infra wants all projects to switch to Xenial for both newton and ocata branches in a less than a month 14:18:11 which is rather tough I admit 14:18:23 de we try with our jobs?, any issues with xenial? 14:18:38 we run some jobs with xenial already 14:18:49 I think tempest based, and unit tests, docs, pep8 are all xenial 14:18:58 some jobs are not, like grenade (?), or functional/fullstack 14:19:08 and I talk about master at this point 14:19:15 and what's the usual way?, replicating a job with xenial in -nv, and then switching when all is finel? 14:19:19 I think jlibosva was tracking the switch for functional/fullstack 14:19:31 finel=fine 14:19:33 ajo: or experimental path 14:19:41 well, experimental right 14:19:56 for functional/fullstack, I think we have some experimental jobs that everyone can trigger and debug 14:20:21 who handled the other changes?, are we tracking this somewhere? etherpad/something ? 14:20:33 even if we switch ocata to all xenial, there is a non obvious work to identify the needed backports for newton 14:20:40 I'd be glad to help a bit with that matter, I don't have lots of time, but I could try to team up and take a job 14:20:47 ajo: yeah, at this point, I believe we need a tracker 14:20:53 and etherpad may be a good solution 14:21:06 and probably not just a tracker, but a squad team to make it to completion 14:21:12 because the time is tight 14:21:36 https://etherpad.openstack.org/p/moving-neutron-jobs-to-xenial sounds good? 14:21:40 I guess I will ask armax to identify what our next steps are 14:21:46 ajo: let's use it, yes! 14:22:08 ajo: let's fill it in after the meeting 14:22:31 ok, so that's one high impact thing 14:22:51 as for day to day stuff, the gate looks quiet for the most part: http://grafana.openstack.org/dashboard/db/neutron-failure-rate 14:23:43 the only suspicious thing is some tempest jobs - linuxbridge and dvr - are a bit higher in the trend, 10% and 15% correspondingly 14:24:04 we may want to have someone to take a look, but overall it does not seem critical (just yet) 14:24:36 obviously, functional is at 15%, but unless folks step in to fix the job, it won't become better 14:24:53 and that's about it for the gate 14:25:19 as for bugs, we had several bug deputies to cover during the summit time and around it 14:25:32 I believe it was dasanind, dasm, and dougwig (all d's!) who handled bugs in last three weeks for us. 14:25:52 folks, would you like to give some brief status for weeks you covered? 14:26:01 anything critical to mention? 14:26:15 dasanind_: you first 14:26:22 ihrachys: there were 31 bugs and 4 RFE's 14:26:56 most of the bugs were medium or low and some were marked as invalid 14:27:38 dasanind_: anything that you could not triage/tag yourself? 14:28:28 there were couple of bugs for fwaas, lbaas and cisco which I could not triage 14:29:10 cisco is not our scope, we just move it to networking-cisco and move on 14:29:16 I can take a pass on the lbaas bugs 14:29:23 dasanind_: as long as you tag the bugs with fwaas/lbaas, that should work fine 14:29:34 blogan helped with the lbass bugs 14:29:34 Yep 14:29:42 lbass=lbaas 14:29:58 njohnston helped with the fwaas one 14:29:58 I don't think we necessarily need to triage to completeness all of them; the most important thing is to pull attention of matter experts via tags. 14:30:11 ok cool. thanks dasanind_ for covering for us! 14:30:17 dasm: your turn 14:30:27 summit week was very calm. just 3 or 4 bugs. 14:30:53 i was expecting to have *plenty* of time, but was buried under pile of other responsibilities, so currently trying to triage all outstanding bugs. 14:30:53 lucky you are, but not surprising :) 14:31:23 dasm: thanks. dougwig may have swiped through some of them already. but I suspect he won't update just now :) 14:31:45 ack 14:31:52 dasm: let's at least make sure nothing slips thru the cracks, and all bugs are tagged properly. 14:31:57 all new bugs I mean 14:32:16 dasm: thanks for working on it 14:32:31 we also need new volunteers to serve as bug deputies for next two weeks. anyone? 14:32:43 I can take one 14:32:56 which doesn't matter 14:33:15 I can be after andreas_s 14:33:32 sold! 14:33:52 andreas_s: electrocucaracha: please update the wiki page after the meeting to document you are the deputies for next weeks 14:34:07 thanks volunteers for stepping in, that was quick this time, and I appreciate that! 14:34:18 #topic Docs 14:34:51 anyone to update us on docs? 14:35:16 ihrachys: Just a note that the bi-weekly netowkring guide meeting has been removed 14:35:35 Networking guide discussion will now take place in the regular documentation team meeting 14:35:56 john-davidge: does it indicate it's too good or it's too bad? :) 14:36:19 ihrachys: It was so good everyone was too intimidated to ever join :) 14:36:27 haha 14:36:35 but now everything will change!! 14:36:39 /s 14:36:54 thanks john-davidge for the notice 14:36:58 I think the main purpose of the dedicated meeting is to discuss the reorganization of the guide, but it looks over. 14:37:33 amotoki: makes sense. so it seems like it's 'good' in the end. 14:37:41 * ihrachys breathes with relief 14:37:43 on a relevant note, we have doc bugs to squash, as usual 14:37:46 #link https://bugs.launchpad.net/neutron/+bugs?field.tag=doc 14:38:14 swipe through the list and take what you can handle (of course if it was your DocImpact patch, everyone expects you to close the loop) 14:38:37 I just did assign some to myself 14:38:52 #topic Transition to OSC 14:39:01 rtheis: are you still with us? :) 14:39:42 ok, anyone else to update on the OSC progress? 14:39:43 around and wrapping up most of my work 14:39:47 amotoki will take over 14:39:50 after talking with rtheis, i took over his role of OSC transition. 14:40:16 rtheis: thanks a lot for all your work on the topic, we are definitely in a better place now 14:40:22 yw 14:40:23 to the point where we deprecated neutronclient CLI :) 14:40:24 I was in bad health last week and I am now catching up with his role. 14:40:57 no solid report this week. will report next week 14:41:06 #link https://review.openstack.org/#/c/393903/ neutronclient CLI deprecation patch 14:41:19 amotoki: thanks for stepping in 14:41:29 #topic Neutron-lib and planned neutron refactoring 14:41:58 as I said, 1.0.0 released lately 14:42:03 amotoki: rtheis ihrachys : Is there a timeline/deadline for the OSC work? 14:42:27 the main work will be done by ocata feature freeze 14:42:27 sindhu: I guess it's always yesterday; but no new features or even some bug fixes are now allowed in neutronclient 14:42:38 #link http://lists.openstack.org/pipermail/openstack-dev/2016-November/106725.html the 1.0.0 release announcement 14:42:57 ihrachys: Ok, thanks! 14:42:58 since neutron-lib now adopted reno, we have a neat and easy way to advertise new developments in the library 14:43:08 for details, see in the announcement emails 14:43:17 to remind, there was a decision made on neutron-lib design session that we don't want to linger removal of the code moved into neutron-lib for too long in neutron repo, so we change the procedure for maintaining the code. 14:43:29 And I hope everyone is aware of the faster adoption plan 14:43:31 instead of deprecating such code for cycles in neutron repo until removal, we do the following: 14:43:43 1. move the code into neutron-lib 14:43:50 2. release a new version of the library 14:43:59 3. announce changes on the team meeting (in this exact section) 14:44:08 4. in several weeks the least, remove the code from neutron tree 14:44:13 in this way, subprojects are given a chance to switch to the new code without leaving their gates broken 14:44:21 and neutron team can proceed with other work instead of maintaining deprecated code 14:44:28 questions? critics? concerns? 14:45:03 ihrachys: does that apply to hacking checks as well? 14:45:05 subprojects are expected to track new lib announcements, AND read meeting logs to be informed about things that may affect their gates 14:45:40 perhaps thats a topic for the neutron-lib meeting 14:45:51 boden: what do you mean? it applies only to neutron in-tree code that is not supposed to be shared with subprojects. hacking checks from neutron tree are afaik not used by anyone anymore 14:46:06 if they are, they do it wrong and should switch to neutron-lib factory/checks asap 14:46:22 and honestly I wouldn't cry if they are hit by a gate failure because of that at this point 14:46:39 * ihrachys thinks how an 'evil' smiley would look like 14:46:43 ihrachys: the general question is if we can roll out new hacking checks that quickly as well.. and also there are subprojects using neutron hacking check factory 14:46:54 we can take this offline; don’t want to derail 14:47:10 With the new "faster" plan we must be more aggressive about the hacking checks too 14:47:14 boden: the new change in the process is not about how we roll in new things in neutron-lib, but about how quickly we can remove deprecated code from neutron tree 14:47:34 yeah, we can discuss it in neutron-lib venue in more details 14:47:39 thanks for bringing it up thoug! 14:48:11 so one thing that is currently in review and that projects should be aware of is the removal of neutron manager code from neutron tree 14:48:13 #link https://review.openstack.org/#/c/386845/ 14:48:49 I believe the new neutron-lib now includes the manager, and so all subprojects are advised to adopt it asap 14:48:53 or else ;) 14:49:24 everyone is silent. stunned by the evil? 14:49:24 I like this new process.. “or else” ;) 14:49:49 ok, report your complaints to your magnificent PTL 14:49:57 I get the impression that people actually like the new plan 14:50:02 ihrachys: is it an announcement? 14:50:11 HenryG: haha, until they are hit by the bus :P 14:50:46 amotoki: it's in a way an announcement, yes; we don't know for how long the deprecation warning for the manager will stay with us, even though the current patch set leaves some deprecation path. 14:51:00 amotoki: or you mean more generally the announcement of the new process? 14:51:19 no. i think it is enough. 14:51:24 one suggestion is to use #info. 14:51:47 ihrachys: btw "hitting by bus", is it already decided, which projects are not any longer in stadium? https://review.openstack.org/#/c/389397/ 14:51:50 ok cool. obviously, we are not completely evil, and so we will consider impact of our actions on case by case basis when ripping off code from neutron tree. 14:52:17 ihrachys: just to be precise the point 4, how many weeks are we talking about? 14:52:41 ihrachys: Make an #info blurb about the new process, then I have one other lib item 14:52:53 electrocucaracha: I don't think we set any specific number, but we are all adults here. I guess 2-3 would be a good compromise. 14:53:37 #info new neutron-lib obsolete code process in place, subprojects are to track new library releases and read meeting minutes and adopt accordingly 14:53:42 HenryG: your stage 14:54:05 We need more core reviewer attention in neutron-lib repo patches 14:54:31 Please. :) 14:54:39 I second the gentleman 14:54:47 * ihrachys feels guilty himself 14:55:06 HenryG: im sure you're looking for just more reviews, but are you asking for more +2 power? 14:55:35 We have plenty of +2 power 14:55:49 Not enough eyes on the code 14:56:03 ok got it 14:56:18 HenryG: thanks for bringing that to attention 14:56:20 #topic Stadium update 14:56:30 dasm raised a good point 14:56:40 as you know we had stadium revisited somewhat lately 14:56:45 it's still in progress 14:56:53 the relevant patches are in https://review.openstack.org/#/q/topic:stadium-implosion 14:57:06 armax was assessing all stadium projects to decide which are governed well or not 14:57:10 some will go out this cycle 14:57:24 we will do changes in governance repos accordingly 14:57:27 to reflect the new world 14:57:51 dasm: to answer your question, it's up to PTL to interpret assessments accordingly 14:58:06 dasm: but I hope summaries in assessments reflect the upcoming decision 14:58:22 reach out to armax if there are still specific questions 14:58:50 any questions on stadium thing before we close the meeting (sadly without open floor this time)? 14:59:13 ihrachys: just wanted to know, if we're already finishing/finished stadium implosion. thanks 14:59:34 dasm: I think we imploded it previous cycles; now we are shrinking it a bit ;) 14:59:50 :D 15:00:10 700 15:00:14 ok thanks all stadium participants who are working on filling in identified gaps lately, it's worth some claps 15:00:19 ok we are out of time 15:00:21 thanks everyone 15:00:28 and may the Force be with you 15:00:30 thanks ihrachys 15:00:31 #endmeeting