Monday, 2018-06-04

*** kumarmn has quit IRC01:10
*** kumarmn has joined #openstack-tc01:40
*** kumarmn has quit IRC01:43
*** gagehugo has quit IRC01:44
*** kumarmn has joined #openstack-tc01:45
*** kumarmn has quit IRC02:03
*** kumarmn has joined #openstack-tc02:04
*** gagehugo has joined #openstack-tc02:17
*** kumarmn has quit IRC02:28
*** kumarmn has joined #openstack-tc02:32
*** gcb has joined #openstack-tc02:43
*** kumarmn has quit IRC02:46
*** kumarmn has joined #openstack-tc02:47
*** kumarmn has quit IRC03:09
*** EmilienM has quit IRC03:22
*** EmilienM has joined #openstack-tc03:22
*** kumarmn has joined #openstack-tc03:40
*** kumarmn has quit IRC03:45
*** kumarmn has joined #openstack-tc04:19
*** kumarmn has quit IRC04:26
*** ricolin has joined #openstack-tc05:19
*** kumarmn has joined #openstack-tc05:36
*** kumarmn has quit IRC05:42
*** kumarmn has joined #openstack-tc06:45
*** kumarmn has quit IRC06:51
*** kumarmn has joined #openstack-tc07:18
*** kumarmn has quit IRC07:22
*** jaosorior has joined #openstack-tc08:05
*** frickler_ has joined #openstack-tc10:16
*** alex_xu_ has joined #openstack-tc10:17
*** purplerbot has quit IRC10:20
*** alex_xu has quit IRC10:20
*** frickler has quit IRC10:20
*** kumarmn has joined #openstack-tc10:54
*** kumarmn has quit IRC10:59
*** jrollinhatin is now known as jroll11:20
*** kumarmn has joined #openstack-tc11:35
*** kumarmn has quit IRC11:40
*** edmondsw has joined #openstack-tc12:12
dimso/12:22
*** jroll has quit IRC12:29
*** jroll has joined #openstack-tc12:30
*** rosmaita has joined #openstack-tc12:31
*** kumarmn has joined #openstack-tc12:35
*** frickler_ is now known as frickler12:39
*** kumarmn has quit IRC12:40
mnasermorning from rainy montreal o/12:46
*** kumarmn has joined #openstack-tc13:06
*** kumarmn has quit IRC13:09
*** kumarmn has joined #openstack-tc13:26
*** guvnah has joined #openstack-tc13:37
*** elbragstad is now known as lbragstad13:42
*** cdent has joined #openstack-tc13:54
*** mriedem has joined #openstack-tc14:00
*** annabelleB has joined #openstack-tc14:15
TheJuliao/14:15
*** TheJulia is now known as needsleep14:18
*** alex_xu_ has quit IRC14:20
*** alex_xu has joined #openstack-tc14:20
dhellmanno/14:32
*** openstackgerrit has joined #openstack-tc14:33
openstackgerritMerged openstack/governance master: Update Zun PTL email address  https://review.openstack.org/57160714:33
dhellmanntc-members: we need to decide what to do with late goal updates like https://review.openstack.org/55786314:37
dhellmannplease comment on the review14:37
smcginnisI've always thought of them as semi-working documents that would get updated over time.14:37
smcginnisSince we know not all projects are able to complete cycle goals on time, don't we want them to come back after the given cycle with updates?14:37
openstackgerritMerged openstack/governance master: Minor grammatical fixes to new project requirements  https://review.openstack.org/57134414:38
cmurphysmcginnis: ++14:38
fungiahh, yeah i thought i had weighed in on the overall idea, but i guess we only discussed it in here14:38
fungithe problem with that is at what point can we actually close them out as historical? we don't want to be taking changes to them until the end of time, and since we've switched away from using code review for tracking those things i think we'd rather freeze them sooner than later?14:39
openstackgerritMerged openstack/governance master: provide more detail about the expectations we place on goal champions  https://review.openstack.org/56406014:39
openstackgerritMerged openstack/governance master: charms: add Glance Simplestreams Sync charm  https://review.openstack.org/56695814:39
smcginnisfungi: That is a fair point.14:39
smcginnisMaybe a policy that if the work completes after the cycle, it's prepended with **cyle-name** to indicate it was not included in that cycle.14:40
smcginnisAnd then we do have some detail as to when the goal was completed for each.14:40
smcginnisSo if someone comes to this doc looking to deploy pike and wondering if they can do it all in py3, they can at least see what they can and cannot do.14:41
smcginnisAnd maybe learn that if they update to a release or two later they can get what they are looking for.14:41
fungii'll leave a fresh comment, but i'm of the opinion that after the cycle is over there's little point in tracking future compliance. the final state is the projects which have that goal met at release time14:41
cdentI've never been clear what the reporting is for other than some cognitive scaffolding for the projects, so it is hard to have an opinion14:43
cdentbiab14:43
smcginnisfungi: My only concern there is if someone finds this doc as they are looking into py3 support and sees that a few projects had not completed the goal, where do they then go from there?14:44
*** cdent has quit IRC14:44
*** alex_xu has quit IRC14:47
*** gcb has quit IRC14:47
*** alex_xu has joined #openstack-tc14:49
*** gcb has joined #openstack-tc14:54
*** needsleep is now known as TheJulia14:56
*** ricolin has quit IRC15:01
zanebI'm inclined to say that if we're not expecting updates then we should delete it from the repo altogether15:09
zanebinaccurate out-of-date documentation is worse than no documentation15:10
*** cdent has joined #openstack-tc15:10
*** purplerbot has joined #openstack-tc15:17
dhellmanncdent : the reason for the reporting was to have a trigger for projects to engage (their planning docs) and then so the TC didn't have to be project managers (their completion docs)15:20
dhellmannI did at one point propose that we move all of the existing tracking to storyboard, which doesn't require reviews to be kept up to date.15:20
*** hongbin has joined #openstack-tc15:26
smcginnisdhellmann: We're doing that for current goals, right?15:26
smcginnisdhellmann: Maybe we need to remove or mark those old goal docs as legacy and out of date to make it clear they are not a current statement of ongoing status.15:27
dhellmannsmcginnis : yes15:27
dhellmannmarking the documents as "done" makes sense15:27
smcginnis"This document was used for tracking progress during the goal cycle and should not be used to determine the current state of the goal for all projects"15:28
smcginnisSomething along those lines.15:28
dhellmannI'd have to go back and look at what the existing goal docs say about late updates15:28
dhellmannI feel like we said we'd just keep updating the existing docs as we went along, but we may never have written that down15:28
cmurphyis there another document to point to for the current state of completion? i'm sure it would be useful to know what projects are py35 ready even if they didn't finish by pike15:29
smcginniscmurphy: ++15:30
smcginnisWe're lacking a way to get from the goal to discover current state.15:30
dhellmannthere's a wiki page for python 315:35
dhellmannhttps://wiki.openstack.org/wiki/Python3#Python_3_Status_of_OpenStack_projects15:35
smcginnisTrue. At least that is more discoverable than an etherpad.15:42
smcginnisMaybe we should update the py3 goal doc with a message stating its purpose and a pointer to that wiki page.15:42
cmurphyit might be good to have something similar for other goals, like policy-in-code15:44
dhellmannfor the newer ones, we link over to the storyboard story15:46
cmurphyah right15:46
dhellmannI'm just uncertain about what to do because we're in the middle of changing from one tool to another. I know what I would do, but I want to know what the rest of the team thinks before I move any data around15:48
* cdent reads the review again15:51
*** diablo_rojo has joined #openstack-tc15:52
*** gcb has quit IRC16:15
*** harlowja has joined #openstack-tc16:28
ttxI'll be on a plane at the TC office hour time for tomorrow, have fun without me16:32
*** ianychoi has quit IRC16:32
*** edmondsw has quit IRC16:41
*** annabelleB has quit IRC16:43
*** edmondsw has joined #openstack-tc16:46
*** ianychoi has joined #openstack-tc16:48
*** annabelleB has joined #openstack-tc16:52
zanebdhellmann: is migrating past goals to storyboard feasible? That seems like the best solution in principle, but I don't know how much work it is16:59
dhellmanncreating the stories is automated but someone would have to manually move the rest of the data. we only have a few.16:59
EmilienMnumber of goals x number of projects who worked on the goals16:59
EmilienMit can be some work16:59
*** harlowja has quit IRC17:10
*** harlowja has joined #openstack-tc17:49
kmalloco/ TC members, can I get a read on something related to Keystone. Keystone has dropped API v2.0 completely... except for 2 things that got missed in the initial deprecation messaging18:13
kmallocas it stands, we would need to keep that around until.. the "T" release [i think]18:13
kmallocthe rest of v2.0 is removed from our tree, this is the old path for ec2tokens [specifically]18:13
kmallocwould there be significant complaints about dropping this last lingering bit outside of the original agreed upon deprecation cycle only because of missed deprecation warning messages in keystone18:14
kmalloclbragstad: ^ please fill in if i missed anything18:14
cdentkmalloc: if the rest of the api is not there, is the ec2tokens stuff still useful?18:15
* lbragstad follows along18:16
kmallocyou can technically use it18:16
kmallochowever, we have a v3 version of the ec2tokens18:16
lbragstadit's literally the same API i think18:16
kmallocit's just the /v2.0/ec2tokens bit still lingering18:16
kmallocit's wired up to be the exact same API18:16
kmallocwith a different path, but it is holding over having to route any of /v2.018:16
kmallocit's a weird situation and it really is only around because a controller or two missed a deprecation warning decorator even though v2.0 was massively deprecated for a long time before18:17
cdentI would think that 2.0s death was sufficiently notified18:18
cdentbut I tend to be spirit rather than letter oriented when it comes to these things18:18
kmalloccdent: so do I, but I want to make sure I'm not running afoul of anything people have legit concerns about :)18:18
kmallocand api removal is... touchy :)18:19
cdentwe'll have to locate some alternate input. tc-members ^^18:19
kmallocWFM, i'm here to answer questions if anything comes up (lbragstad can answer too)18:20
mnaserdon't things like heat tools rely on it for some things (i think?)18:20
dimskmalloc : +1 to pull the plug (personally)18:21
kmallocmnaser: don't think they lean on the /v2.0 stuff since /v2.0 is gone.18:21
kmallocmnaser: we're keeping the /v3 version for sure.18:21
mnasermaybe we could reach out to the mailing list and hear if there are any users of it still?18:22
jrollwhat is the ec2tokens api useful for? can you use those tokens to auth to other bits in openstack?18:23
jrollor were they tied to the nova ec2 api somehow?18:23
smcginniskmalloc: My opinion, but if you announced /v2 was deprecated and then announced /v2 was removed, I think it's fine to remove these last little bits.18:23
TheJuliaIt seems to make sense to ask operators to possibly check logs? That seems like the true indicator if it is possible to do so sinner rather than later.18:23
TheJuliaSooner. Yay for autocorrect18:23
kmallocjroll: some folks utilize it for compat with aws tools since we can get auth with the ec2 tokens that way18:23
smcginnisTheJulia: Thought maybe Feudian slip. :)18:24
cdentsinners be using old apis18:24
kmallocjroll: it plays in heavily with the ec2api [external project]18:24
jrollkmalloc: what other openstack APIs... okay, right18:24
smcginnis*freudian18:24
kmallocjroll: i know heat used it at one point for stuff18:24
cdentsmcginnis: i subscribe to your goal newsletter18:24
fungiyeah, we say there's a deprecation period but we don't get into the weeds about how you implement deprecation that i've seen. announcing widely and documenting that the v2.0 api is deprecated in its entirety seems sufficient even if a couple of methods missed a decorator when it came time to implement18:24
smcginnis:)18:24
kmallocjroll: i really can't tell you who/what uses it18:24
jrollkmalloc: and you've reminded me that project still exists, just not in nova's tree, so my questions don't matter as much I guess18:24
jroll(the ec2api project)18:25
kmallocjroll: :P18:25
TheJuliasmcginnis: if so, I’ll blame the glass of wine I had with dinner. ;)18:25
lbragstadfungi: from my pov, we were trying to be consistent in wrapping the v2.0 routes with deprecation warning18:25
kmallocjroll: i know, i try and forget some of that stuff when i can too18:25
smcginnisTheJulia: Hah!18:25
lbragstadwarnings*18:25
kmalloclbragstad: but the warnings only go to operators18:25
lbragstadright18:25
jrollkmalloc: :)18:25
lbragstadwe just missed a couple of those paths unfortunately18:25
kmallocbut all of v2.0 was long since marked, it was truely a simple copy-paste oversight18:26
lbragstadright18:26
kmallocTheJulia: wine is good :) esp. with dinner!18:26
TheJuliaIt sounds like fire should be applied in that case18:26
cdentkmalloc I continue to think: it's okay to kill18:26
kmalloccdent: ++, i'm feeling like the sentiment is "it's ok to kill"18:27
* jroll thinks kill it with the rest but does not have a TC hat18:27
kmallocand maybe we should just fire an email to the mailing(lists) informing folks of this change.18:27
* cdent gives jroll a TC hat to try on18:27
kmallocand the reasoning.18:27
cdentgood plan18:27
mnaseri don't think anyone is opposed to it unless we hear of an operator that expresses concern that we missed18:27
kmalloccdent: i had a tc hat at one point... it didn't fit well18:27
mnaserit would be better to hear input from a bigger group18:28
jrollcould always do it in a separate commit that could be reverted, if that isn't too painful18:28
kmalloci'll wire up the removal as a commit we can revert18:28
kmallocmostly i'm reworking all of keystone to flask18:28
cdentyay!18:28
mnaserdon't forget a release note this time (ha)18:28
kmallocand i wanted to avoid wiring up old /v2.0 paths18:28
kmallocmnaser: hehe18:28
jrollalso makes it easy for people to revert downstream if they don't want to speak up18:28
kmallocjroll: not exactly :P18:28
kmallocjroll: move to flask is very invasive18:28
cdentseems like people who want to keep the functionality can stay old18:29
jrollkmalloc: well if downstream can't easily revert it, how would upstream?18:29
cdentthat's always been the option18:29
kmallocjroll: i can revert the removal and wire it up18:29
kmallocthe wire-up paths wont exist18:29
jrollah gotcha18:29
kmallocwiring it up isn't hard, but i want to avoid doing it at all and the potential bugs if i don't have to18:29
kmalloccdent: also, we're dropping paste.deploy on the cutting room floor18:30
cdentyay!18:30
kmalloccdent: which is a big win, imo18:30
cdentyes, very18:30
TheJulia It feels like we have consensus18:30
kmallocTheJulia: ++18:30
kmalloci'll get an email written to -operators and -dev18:30
kmallocor.. whichever mailing lists18:30
TheJuliaAwesome, and with that I go sleep()18:31
kmallocand i'll not wire it up, but isolate the removal to a single commit18:31
kmallocTheJulia: sleep well!18:31
kmallocTheJulia: sleep(time.until_morning_and_rested())18:31
TheJulia++18:31
kmalloccdent: thanks! :) btw, was really good to see ya at vancouver.18:31
zaneb+1 for killing it18:32
kmalloczaneb: thanks! :)18:32
cdentkmalloc: same18:32
zanebrelease notes are more important than decorators imho18:32
smcginniskmalloc: Would love to see a write up of the move to using flask at some point if that's at all a possibility.18:32
cdentsmcginnis++18:33
zanebif release notes said v2 was going away then v2 can go away18:33
*** annabelleB has quit IRC18:33
kmallocsmcginnis: totally. i'll have a big write up on what we did and the big changes going on with it18:33
smcginniskmalloc: \o/18:33
kmallocsmcginnis: the hard part was isolating keystone bits in a way to ensure we can move just one path at once e.g. /users or /projects18:33
kmallocsmcginnis: if you want to see the patch https://review.openstack.org/#/c/568377/718:34
smcginnisOh, I could imagine. That sounds tricky.18:34
smcginniskmalloc: Thanks!18:34
kmallocsmcginnis: the key bits are in keystone.server.flask.application and it's the middleware18:34
kmallocsmcginnis: that does the PATH_INFO and SCRIPT_NAME rewriting18:34
kmallocsmcginnis: so we dispatch to webob/mapper for specific paths until we migrate each one.18:35
kmallocsmcginnis: we're also going to do a big restructure of keystone's code to show what is routable (REST) and what is not in the process.18:35
smcginnisExcellent18:35
kmallocsmcginnis: e.g. keystone.subsystem.identity will be the identity subsystem rather than keystone.identity18:35
kmallocbut things like keystone.exception (not RESTful routing things and shared) wont be moved18:35
kmallocand we'll be all flask + Flask-Restful/Restplus with blueprints when done18:36
kmalloc(we'll = keystone)18:36
kmalloci'll plan to write up a long blog post (hey, time to bring my blog back) on how to do this type of thing18:36
*** annabelleB has joined #openstack-tc18:37
smcginnis++18:37
*** mriedem has quit IRC18:40
cdentzaneb: if you'll have me, I'd like to help with the technical vision task18:41
zanebcdent: sure!18:42
cdenti'll put my name on the wiki18:42
zanebI need to start typing up some ideas, it hasn't made it to the top of the to-do list yet18:42
cdenti know how that can be18:42
*** mriedem has joined #openstack-tc18:43
*** cdent has quit IRC19:05
dhellmannkmalloc , et al: ec2-api is an official team, fwiw: https://governance.openstack.org/tc/reference/projects/ec2-api.html so maybe loop them into the discussion?19:06
kmallocdhellmann: sure, it still works with the /v3 bits, we're keeping the functionality.19:13
dhellmannkmalloc ok. I was keying on your comment calling it an "external" project19:13
kmallocyep19:13
kmalloc:)19:13
*** diablo_rojo_ has joined #openstack-tc19:19
*** jaosorior has quit IRC19:19
*** diablo_rojo_ has quit IRC19:20
*** diablo_rojo has quit IRC19:20
*** diablo_rojo has joined #openstack-tc19:25
*** harlowja has quit IRC19:29
*** diablo_rojo has quit IRC19:33
*** diablo_rojo has joined #openstack-tc19:35
*** diablo_rojo has quit IRC19:35
*** diablo_rojo has joined #openstack-tc19:37
*** cdent has joined #openstack-tc19:57
*** harlowja has joined #openstack-tc19:59
*** annabelleB has quit IRC20:07
*** annabelleB has joined #openstack-tc20:22
cdentmriedem: I believe, but am not fully certain, that python3 first simply means that the test jobs default to python3 and that developers running tox will run python3 when they don't otherwise specify20:56
dhellmanncdent, mriedem : yes, that's the idea. Right now we default to python 2 unless specified otherwise. "Python 3 first" means flipping that around, for all uses of python.21:00
dhellmannI'm working on writing up the tox.ini changes and project-config changes needed to do that21:00
*** annabelleB has quit IRC21:00
*** annabelleB has joined #openstack-tc21:02
*** harlowja has quit IRC21:04
*** edmondsw has quit IRC21:35
*** harlowja has joined #openstack-tc21:36
mriedemok, that doesn't seem like something that's going to be all that hard21:55
mriedemso, i'd suggest we put emphasis on something else as a community wide goal which will require more work/engagement21:56
smcginnisThere are differences with some of the tools. But if it really is something minor, maybe we should pick it as a third stretch goal.22:17
smcginnisI have a feeling it will uncover a few issues though.22:17
*** kumarmn has quit IRC22:20
*** annabelleB has quit IRC22:28
*** mriedem has quit IRC22:30
*** mriedem has joined #openstack-tc22:48
*** hongbin has quit IRC22:59
*** kumarmn has joined #openstack-tc23:01
*** cdent_ has joined #openstack-tc23:03
*** cdent has quit IRC23:04
*** cdent_ is now known as cdent23:04
*** annabelleB has joined #openstack-tc23:08
*** kumarmn has quit IRC23:13
*** cdent_ has joined #openstack-tc23:24
*** cdent has quit IRC23:26
*** cdent_ is now known as cdent23:26
*** kumarmn has joined #openstack-tc23:30
*** kumarmn has quit IRC23:34
*** kumarmn has joined #openstack-tc23:46
*** mriedem has quit IRC23:48
*** kumarmn has quit IRC23:53

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