Tuesday, 2018-10-16

*** annabelleB has quit IRC00:29
*** jamesmcarthur has joined #openstack-tc00:39
*** jamesmcarthur has quit IRC00:47
*** jamesmcarthur has joined #openstack-tc00:51
*** jamesmcarthur has quit IRC00:52
*** jamesmcarthur has joined #openstack-tc00:52
*** diablo_rojo has quit IRC02:17
*** jamesmcarthur has quit IRC02:18
*** jamesmcarthur has joined #openstack-tc02:27
TheJuliattx: that clearly takes priority02:29
*** jamesmcarthur has quit IRC03:02
*** jamesmcarthur has joined #openstack-tc03:08
*** jamesmcarthur has quit IRC03:19
*** jamesmcarthur has joined #openstack-tc03:45
*** Bhujay has joined #openstack-tc03:51
*** Bhujay has quit IRC03:52
*** Bhujay has joined #openstack-tc03:52
*** Bhujay has quit IRC04:06
*** jamesmcarthur has quit IRC04:44
*** e0ne has joined #openstack-tc06:48
*** markvoelker has quit IRC07:29
*** tosky has joined #openstack-tc07:52
*** jpich has joined #openstack-tc08:00
*** dtantsur|afk is now known as dtantsur08:03
evrardjpo/08:04
*** e0ne has quit IRC08:37
*** e0ne has joined #openstack-tc08:44
*** cdent has joined #openstack-tc08:59
evrardjplet us know how it goes ttx :)09:01
* cdent waves09:02
gmanno/09:09
gmannoffice hour ?09:09
evrardjpit is09:09
cdentthis is the right time for it but this one has been pretty slow of late09:09
gmannk09:10
lbragstado/09:10
evrardjpI am surprised, the complaint's office should be busy during french business hours #funny_unoffending_joke_for_belgians09:11
gmanni was reading log in morning specially upgrade checker goal09:12
evrardjpgmann: and?09:12
gmannlbragstad: cdent evrardjp seems like we need more help on scaffolding CLI itself for upgrade check and we do not prefer to script that right ?09:12
gmanni am thinking to call the volunteer team to help matt on this and that team or project team can work to fill the actual upgrade checks09:13
gmanni have 2 member from my employer who can help this (one is congress developer and one started in cinder recently )09:14
gmannand like that if we collect more name then, it is possible to have the things up by Nov first week.09:14
cdentI was out last week so am not up to date on the issues. What the summary of the problem?09:15
evrardjpI am not sure all companies will have the same agility to "assign" people on community goals in a short timeframe. But that's another conversation.09:16
gmannit is question more than problem - how fast we would be able to finish this goal. and how many projects started this (i saw only 4-5)09:16
gmannevrardjp: they will from existing contributors only, like "hey, i have few bandwidth in this cycle and i can help on this goal under goal champion plan "09:17
gmannit just need we open up the formal volunteers need over ML etc09:18
evrardjpgmann: I believe it's better to promote the community goals as something to assign people to, at least in my company, instead of relying on extra time from individuals. But that's a personal opinion/experience, and I am working towards that.09:19
gmanntrue, we have assignee for goal, and extra time from others (with dedicated volunteer team) hell it to finish fast09:20
gmanns/hell/help09:21
*** persia has quit IRC09:27
gmannanother things i wanted to ask about start of health tracker  - https://wiki.openstack.org/wiki/OpenStack_health_tracker09:27
gmannare we going to start the tracking projects once we have all team filled with TC liaison ? and go ahead from mnaser dhellmann ?09:28
gmannor we can start anytime for projects already assigned09:28
*** persia has joined #openstack-tc09:28
cdentIt shoul be fine to start whenever you feel like it. As long as you date the information you put in the wiki.09:29
gmannok09:30
cdentThe process is supposed to be more of an ongoing conversation of "how's it going" rather than formal checkins09:44
gmannyeah.09:45
evrardjpcdent: I thought we were planning to discuss what to track together.09:51
evrardjpI thought that was the outcome of the PTG -- we needed a more exhaustive and shared list, but I might have misunderstood09:52
gmannyou mean to reiterate this - https://wiki.openstack.org/wiki/OpenStack_health_tracker#Health_check_list09:52
evrardjpI meant having a list of questions for element 409:53
cdentevrardjp: I'm not saying we don't want that, but that we don't want a strict time structure of engagement with each project. The stuff we watch out for should be similar and known up front, but how we gather the info is dependent on how each project operates09:53
evrardjpnot how, what :)09:54
evrardjpmy point in an email long ago was that some of the what, if we all agree with those items, can be automated, and therefore we can focus on the less automation-more human part09:55
cdentI can't tell if you are agreeing or disagreeing with me. My comment to gmann was simply "don't worry about time"09:55
evrardjpcdent: can't tell either :)09:55
cdent:)09:55
evrardjplet's say we agree then09:56
lbragstadi have a basic set of questions - but this is also my first rodeo09:57
evrardjplbragstad: could you share those? Right now I was taking the output of current status update and a few of my ideas to draft an email, but I guess I'd love working with you on said questions09:59
evrardjpI don't want to have a similar thing as "marketing" fatigue on people -- too many questions and a few will get ignored.09:59
lbragstadhow would you rate $PROJECT's health? are there things you're concerned about within $PROJECT? are there things you're concerned about within OpenStack? Is there anything the TC can help you with?10:00
lbragstad^ those are basically what i have, but i'm also massaging them to be specific to certain projects10:01
cdentI think I approach these things differently. I think the things we need to be discovering are the stuff that the members of the project are not aware of themselves and may be missing as a result of being within the project and not having an external project. I think it is critical to speak not just with the PTL and core or common members of the teams, but the irregular contributors.10:01
persiaI suggest that the goal should be to build a useful relationship between the project and the person doing the checkin, rather than to answer N questions, or similar.  To my mind, part of the measure of a project's health is "does this project feel that they have good connection to TC?".10:01
lbragstadthose are both good points10:02
gmanncdent: +1 to check not just PTL but core and other contributor too. this help to understand the complete situation10:02
cdentthe agenda there is to make sure that people who don't already have a strong voice are heard well10:03
evrardjppersia: that's why I think have the right questions are hard -- establishing a relationship vs inquiry10:03
persiaevrardjp: Makes sense.  I guess I think "having the right questions" is an impossible task, so going with "some questions that might be useful" becomes a useful target.10:05
gmannand even i would like to ask them -do they know about TC heath tracker & its objective and how TC can/will help them. to give them confidence that this exercise is not just formalities but we really want to help in any ways.10:06
*** jamesmcarthur has joined #openstack-tc10:56
*** e0ne has quit IRC11:18
*** e0ne has joined #openstack-tc11:33
dimsgmann : +112:23
*** e0ne has quit IRC12:24
*** e0ne has joined #openstack-tc12:27
TheJuliaGood morning13:00
smcginnisMorning TheJulia13:02
* TheJulia sips coffee13:02
smcginnisSo tosky and I were just chatting about the py37 situation, and I'm not sure I have the latest consensus on that. Should we be replacing 35 with 37, or is the question still up in the air?13:04
TheJulialbragstad: I last time we did healthchecks, I went with "how do you perceive your project's health?" Not asking for them to rate, but it kind of got them into a situation where they were writing about their perceptions instead of saying bad/good/great (for the most part)13:04
lbragstadthat makes sense13:04
lbragstadi've adjusted my wording a little bit13:04
lbragstadhopefully away from implying a rating13:05
TheJuliaI didn't go and poll some cores as well, but I did look through meeting logs to see if there was visible contention, looked at statistics, etc.13:05
smcginnistosky: I think we are going to be switching everyone to Bionic, so in that case based on fungi's comments, those patches do make sense.13:06
toskyso both py37++ and py35--13:07
fungithe addition of py37 jobs is unrelated to the removal of py35 jobs (except insofar as the platform to which we're moving will enable testing with py37 easily in addition to py36)13:08
cdentFrom a capacity standpoint, I would think that adding any jobs has to be related to removing some?13:08
smcginnisfungi: I think the reason they are tied together in a lot of conversations right now is there is the set of patches proposed out there that does both.13:08
fungiwe need to remove py35 jobs unless we want to continue testing stein (and beyond?) on ubuntu 16.04 lts13:09
smcginnisAnd cdent ++13:09
fungicdent: adding py36 jobs is related to removing py35 jobs, yes13:09
fungiadding py37 jobs is net new13:09
smcginnisfungi: This is what is being broadly proposed out there - https://review.openstack.org/#/c/610751/1/.zuul.yaml13:10
fungiyeah, i can see how that would be confusing13:10
toskyor better: https://review.openstack.org/#/q/topic:py37-job13:10
fungibasically there are three things in play at the same time. the python3-first goal, following ubuntu lts releases upgrading our testing from xenial to bionic, and the availability of an optional python3.7 package on bionic we can use to get additional coverage on newer-than-default python3 if we want13:11
*** e0ne has quit IRC13:12
smcginnisSo I think we need to be able to give projects some specific guideance on the whole thing.13:14
fungii agree13:14
smcginnisI _think_ we want to end up with just 36 and 37, but I also think there are some differing opinions on that.13:14
fungii think dropping python3.5 in stein is appropriate. we dropped 3.4 when we transitioned from trusty to xenial13:14
smcginnisIn the meantime, folks like tosky are left wondering what they are supposed to do, which isn't good IMO.13:14
fungii also think it's perfectly fine to push back on the 3.7 additions as they're less urgent13:15
fungipriority should be on switching from 3.5 (default and only python3 packaged in xenial) to 3.6 (default and lowest python3 packaged in bionic)13:16
toskyI think that py36 jobs were fully added in the previous set of (official) patches13:16
toskymore or less when job definitions were moved in-tree13:16
*** e0ne has joined #openstack-tc13:16
toskyso should I push back the changes, asking the author to just remove py35, or could this be done in a more centralized way?13:17
smcginnisYep, we should be good with py36 with Doug's set of patches.13:17
fungiyes, part of the plan, i thought, was to first add the 3.6 jobs and then remove the 3.5 jobs to give projects a (brief) window to make sure 3.6 was working for them13:17
toskyI suspect that some of those patches have been merged already13:17
smcginnistosky: I think you are right.13:17
toskyuh, someone must be reading, as the author just added -W to the patches13:18
toskyat least the sahara ones13:18
fungiand once projects which had been running py35 jobs had their py36 jobs voting we'd drop the py35 jobs and switch the default node type from ubuntu-xenial to ubuntu-bionic13:18
toskymore, to all of them13:19
smcginnisThere's the ML thread on this, but I think someone (us or infra or ?) needs to declare a position here so we stop this confusion.13:19
evrardjpsmcginnis: what about just in the middle and we ask fungi :p13:19
smcginnisevrardjp: He does meet multiple criteria. ;)13:20
evrardjpsorry to put you on the spot! :p13:20
evrardjpand knows so many things13:20
smcginnisBut in general, we've had the vague-ish statements about LTS versions and Python support, but I think a more clear declaration should be made by us. Or at least some public show of a concensus on the matter.13:21
fungilbragstad: persia: the idea was to flesh out https://etherpad.openstack.org/p/tc-health-checklist as a set of conversation starters for the health updates13:22
*** mriedem has joined #openstack-tc13:23
evrardjpfungi: I'd like to understand L23 there: https://etherpad.openstack.org/p/tc-health-checklist -- as anyone can get involved to producing help on community goals, you mean just the reviewing part/update of status, right? Or do you mean someone taking an active part? It can be read in two ways, and one is more forward thinking than the other.13:27
evrardjpfungi: do you mind me adding elements to said etherpad?13:28
fungievrardjp: of the community goals which affect that team's deliverables, have they completed ones for last cycle? are they working on the ones for this cycle?13:29
evrardjpI am thinking about the rate of changes in core contributors (losing cores or gaining them), also asking if they have a process of onboarding too13:29
fungievrardjp: adding stuff is great, thanks!13:29
evrardjpfungi: so two elements there: For new projects, if they met the requirement of reaching the previous goals into the one year timeframe, and for non-new projects, if they are keeing up.13:30
fungii don't understand. can you rephrase? what's the one-year timeframe for?13:32
evrardjpI thought this was a mandatory requirement13:33
evrardjpjust a second13:33
evrardjphttps://governance.openstack.org/tc/reference/new-projects-requirements.html13:33
evrardjp"Any existing goals that are not met should be prioritized and completed within the first year of a team joining."13:33
fungiahh, i'd forgotten about that. thanks!13:33
evrardjpI will write this in the etherpad for ppl to remember13:34
fungiso yes, for projects less than a year old i think the questions are still relevant (at least for the current cycle), but maybe just insofar as making sure they're aware of the goals process and what the current cycle goals are and how they might apply13:34
gmannmriedem: in case you have not read the log. i talked to 2 of my colleague (upstream developers) and they would like to help you on upgrade checker goal.13:37
gmannmriedem: i am thinking on side of making a volunteer team who cal closly work with goal champion t finish the code soon.13:38
gmannmriedem: along with those 2 i would like to ask more volunteer on ML  to join if that approach looks ok to you?13:39
mriedemsure. what do you expect these volunteers to be doing?13:39
gmannstart working on pending projects work or project not yet started13:40
mriedemok yeah that's something we need13:40
mriedemsome projects don't even have release notes, like zun,13:41
mriedemso their upgrade check is probably just going to be a placeholder/noop13:41
mriedemand cyborg13:41
gmannyeah.13:41
mriedemcyborg is too new to really have any upgrade impacts13:41
gmanneven i saw few of project have no upgrade notes in current release like congress so it will be same for them13:42
gmannat least yet13:42
mriedemthat or teams just aren't good about documenting their upgrade impacts13:43
gmannfor project having upgrade content, would you like to implement the specific upgrade check also along with CLI placeholder right ?13:43
mriedemwe don't need a placeholder if the project has a real upgrade check13:43
mriedemlike the one i've identified for glance13:43
gmanni mean actual checks for upgraded items13:44
mriedemwhat i've seen a few projects doing, like keystone and cinder, is starting with a base patch that adds the framework, docs, entry point, etc, and then subsequent changes replace the placeholder/noop check with real checks13:44
mriedemso that pattern is probably fine for the volunteers13:45
mriedemother projects like designate just start with a real check13:45
gmannyea, that sounds good13:45
gmannand further you can schedule (if  you feel needed) hangout or office hour with volunteer to move things fast/particular direction etc13:46
mriedemit's been a long time since i've had minions, i might become drunk with power13:48
cdentwhat have you been drunk with all the rest of this time?13:50
mriedemtoilet wine13:50
cdentwoot!13:50
gmann:)13:52
*** e0ne has quit IRC13:53
*** e0ne has joined #openstack-tc14:01
evrardjpoh god I thought this term was only a joke but there _is_ something called toilet wine. I am learning everyday.14:01
smcginnisevrardjp: Important lesson - never, ever google something odd that mriedem says. It never goes well.14:03
*** annabelleB has joined #openstack-tc14:03
evrardjpI think I did fine. I landed on urban dictionary what can go wrong?14:04
scas'toilet wine' is also a phrase i use outside the urban dictionary definition for when my cats decide to have a tasting party of some vintage toilet water14:05
evrardjphahaha14:06
*** whoami-rajat has joined #openstack-tc14:28
*** annabelleB has quit IRC14:45
*** annabelleB has joined #openstack-tc14:45
clarkbsmcginnis: fwiw the TC has made what I thought were clear determinations in the past that should still apply today. The one that pplies here is OpenStack the software supports latest Ubuntu LTS and CentOS/RHEL. WIth that declaration openstack should continue to support python 2.7 and add python 3.6. Python3.5 is not on either of those two platforms and can be dropped14:55
clarkbAdding python 3.7 is completely orthogonal to that and the idea there is to catch the python 3.7 problems early before our statement of support requires we run on 3.7 and its a lot of effort to update14:56
smcginnisOK, I thought some were asserting that 3.7 was something that would be standard during this cycle.14:59
smcginnisIf that's the case, I think we should clearly declare 2.7 and 3.6 as the must support versions and anything else is optional.15:00
smcginnisWhich means the proposed update to the goal should not say either/or but just add something saying if teams would like to add 3.7 they are free to do so but it is out of scope for stien.15:00
smcginnis*stein15:00
fungiagreed15:11
dimssounds fair smcginnis15:13
smcginnisIt would probably also be good to recommend them as experimental jobs so we are not using extra gate resources for something that isn't required yet.15:13
*** openstackgerrit has joined #openstack-tc15:23
openstackgerritJeremy Stanley proposed openstack/governance master: Make Python 3 testing requirement less specific  https://review.openstack.org/61101015:23
fungithere's a stab at reducing the confusion15:23
smcginnisI was thinking of proposing a standard item to each cycle goal that explicitly states what our targeted runtimes are. Would that make sense? If so, then I think the pti should point to that.15:25
smcginnisExplicit over implicit.15:25
fungisure, though i don't want to be updating the (python) pti itself every time our primary platforms release new versions. if we want to put that in its own document that may be an okay middle-ground15:26
fungijust be aware we're not testing, e.g., "python 3.6" but rather "the python 3 provided with ubuntu 18.04 lts which they call 3.6.something but also has some custom patches"15:27
smcginnisfungi: Yeah, I was thinking the PTI could then just point to the governance cycle documentation and stay out of saying any versions there.15:27
smcginnisfungi: Are we testing any other platforms as official in gate, or are things like RHEL left up to Red Hat's own testing?15:28
fungiwe're testing latest centos on the assumption that it's close to the same as latest rhel15:28
fungithough we run fewer jobs on it than ubuntu lts, i'm not sure what the exact breakdown is15:29
clarkbfungi: actually15:29
fungiwe could consider switching our python2.7 jobs to centos15:30
clarkbif you consider tripleo to be centos testing its really close to even right now I think15:30
fungifair, tripleo does consume more nodes than any other project15:30
fungiand is exclusively on centos15:30
smcginnisI want to make sure I'm clear on wording if I document this. Should it state the current Ubuntu LTS and RHEL? Or is there such thing as a CentOS LTS release?15:30
clarkbhttp://grafana.openstack.org/d/rZtIH5Imz/nodepool?orgId=1 the in use nodes graph15:30
fungismcginnis: see the note in the linux distros section of the pti15:30
fungiwhich states the relative choice of centos as a stand-in for rhel15:31
clarkbits like sin and cos graphed for ubuntu vs centos. We do a bunch of one then a bunch of the other (has to do with how zuul schedules work It hink)15:31
smcginnisOK, so Ubuntu latest LTS and CentOS latest stable.15:31
*** e0ne has quit IRC15:50
*** jpich has quit IRC16:29
*** diablo_rojo has joined #openstack-tc16:33
openstackgerritSean McGinnis proposed openstack/governance master: Explicitly declare Stein supported runtimes  https://review.openstack.org/61108016:36
smcginnisAn attempt at clarity. ^16:37
*** zaneb has quit IRC16:41
*** annabelleB has quit IRC16:50
*** annabelleB has joined #openstack-tc16:51
*** jamesmcarthur has quit IRC17:00
*** jamesmcarthur has joined #openstack-tc17:00
*** dtantsur is now known as dtantsur|afk17:03
*** annabelleB has quit IRC17:07
*** e0ne has joined #openstack-tc17:08
*** whoami-rajat has quit IRC17:13
openstackgerritSean McGinnis proposed openstack/governance master: Fix format errors in PTI docs  https://review.openstack.org/61109817:16
*** zaneb has joined #openstack-tc17:33
*** e0ne has quit IRC17:59
*** annabelleB has joined #openstack-tc18:01
*** diablo_rojo has quit IRC18:03
*** diablo_rojo has joined #openstack-tc18:10
*** e0ne has joined #openstack-tc18:15
*** e0ne has quit IRC18:22
*** annabelleB has quit IRC18:29
*** annabelleB has joined #openstack-tc18:37
*** annabelleB has quit IRC18:37
*** jamesmcarthur has quit IRC18:38
*** jamesmcarthur has joined #openstack-tc18:52
*** jamesmcarthur has quit IRC18:56
*** annabelleB has joined #openstack-tc19:00
openstackgerritSean McGinnis proposed openstack/governance master: Update sphinx extension logging  https://review.openstack.org/61113219:09
*** jamesmcarthur has joined #openstack-tc19:25
*** jamesmcarthur has quit IRC19:35
*** jamesmcarthur has joined #openstack-tc19:54
*** jamesmcarthur has quit IRC19:56
*** jamesmcarthur has joined #openstack-tc19:56
*** jamesmcarthur has quit IRC20:12
*** annabelleB has quit IRC20:16
*** annabelleB has joined #openstack-tc20:18
*** jamesmcarthur has joined #openstack-tc20:19
*** cdent has quit IRC20:24
*** mriedem is now known as mriedem_away21:02
*** annabelleB has quit IRC21:05
*** jamesmcarthur has quit IRC21:20
*** jamesmcarthur has joined #openstack-tc21:22
openstackgerritSean McGinnis proposed openstack/governance master: Fix format errors in PTI docs  https://review.openstack.org/61109821:35
*** lbragstad is now known as lbragstad-50321:42
*** mriedem_away has quit IRC21:55
*** smcginnis is now known as smcginnis_vaca21:55
openstackgerritGage Hugo proposed openstack/governance master: Retire project Anchor - Step 5  https://review.openstack.org/61118722:08
*** jamesmcarthur has quit IRC22:26
*** jamesmcarthur has joined #openstack-tc22:38
*** jamesmcarthur has quit IRC22:43
*** jamesmcarthur has joined #openstack-tc22:54
*** jamesmcarthur has quit IRC23:07
*** jamesmcarthur has joined #openstack-tc23:07
*** tosky has quit IRC23:10
*** jamesmcarthur has quit IRC23:18
*** jamesmcarthur has joined #openstack-tc23:21
*** annabelleB has joined #openstack-tc23:27
*** jamesmcarthur has quit IRC23:32
*** annabelleB has quit IRC23:39
*** jamesmcarthur has joined #openstack-tc23:44

Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!