Tuesday, 2018-01-23

*** kumarmn has quit IRC00:03
*** gcb has quit IRC00:05
*** openstackgerrit has joined #openstack-tc00:06
openstackgerritMerged openstack/governance master: Add charm-ironic  https://review.openstack.org/53161000:06
*** hongbin has quit IRC00:06
*** robcresswell has quit IRC00:12
*** kumarmn has joined #openstack-tc01:26
*** liujiong has joined #openstack-tc01:38
*** kumarmn has quit IRC01:43
*** robcresswell has joined #openstack-tc02:09
*** kumarmn has joined #openstack-tc02:13
*** kumarmn has quit IRC02:18
*** harlowja has quit IRC02:27
openstackgerritMasayuki Igawa proposed openstack/governance master: Add Cold upgrades capabilities  https://review.openstack.org/53354403:27
*** rosmaita_ has quit IRC04:00
*** coolsvap has joined #openstack-tc04:15
*** david-lyle has quit IRC04:30
*** dklyle has joined #openstack-tc04:31
*** harlowja has joined #openstack-tc05:12
*** harlowja has quit IRC05:13
*** kumarmn has joined #openstack-tc06:43
*** kumarmn has quit IRC06:48
*** tdasilva has quit IRC07:47
*** tdasilva has joined #openstack-tc08:08
*** cdent has joined #openstack-tc08:52
cdenttc-members and others, ahoy hoy09:00
*** jpich has joined #openstack-tc09:03
cdentno takers today?09:03
cmurphyoh hello09:08
* cmurphy was distracted09:08
cdentI guess the stuff that currently matters is refining and choosing goals.09:16
* cmurphy going through them now09:17
cmurphythere hasn't been that much community input on them so far09:21
cdentyeah, unfortunately that's pretty standard09:23
cdentI think we (the TC) underestimate how heads down people in the community often are09:23
johnthetubaguycdent: ++09:26
cdentI'm not sure if there's anything to be done about it. Or if anything should be.09:27
johnthetubaguythe digests are a great help from before when there were none09:28
cdentI do think that it can mean that we are obliged to interpret the lay of the land a bit more, to make policy.09:29
johnthetubaguyI don't find the UC meeting easy to attend, did they talk about TC goals feedback at all?09:29
cdentI don't watch that show.09:30
johnthetubaguycdent: as long as we are happy with the PTG turning everything upside down, which I think we should be, then that is probably a good call09:30
johnthetubaguyI would say the goals seem to be helping fixing things, so it seem well worth keeping pushing09:31
* cdent nods09:31
cdentI care a lot about this http://lists.openstack.org/pipermail/openstack-dev/2018-January/126486.html but I have no clear ideas on how to improve the situation09:48
persiaWhile folk are definitely heads-down, I wonder also how many folk feel they are not part of the group that decides the goals.  In other contexts, I find it requires quite a lot of training and reminding of both staff and team leads to cause folk to feel comfortable posting their intent to a project mailing list.  There may be a similar apparent barrier in OpenStack.09:48
cdentpersia: yeah, I think that's a factor too09:49
cmurphycdent: nod09:49
persiacdent: On Nova decomposition, do you know the relevant parties well enough to cause a group of them to run on a common platform, with a predefined breakdown?  The way our election system works feels like it would require a single candidacy from several folk to avoid the "there is one PTL" issue.  As precedent, there was a DPL election with a multiple candidacy a few years back (they didn't win).  I can probably dig up the relevant mail if you need09:52
persia documentation.09:52
johnthetubaguycdent: I noodled long and hard about that too, the Czar stuff reduced the stress but we had more folks stepping up back then09:53
cdentat the moment "relevant parties" is still pretty vague09:53
cmurphycdent: is your concern that no one will want to step up to fill in those shoes given the amount of work perceived to go with the job, or that there is no good system in place for delegation which leads to PTLs shouldering too much of the burden?09:54
cdentcmurphy: more like: whoever steps up may see (from histor) that they are obliged to do a ton of work, and that's not an okay way for what is a job to be09:56
cdentbut your two points are also mixed in there as well09:56
persiaHIgh-effort roles may also require employer consent, which can complicate volunteering.  For many projects, being PTL is only a few hours a week.  For our larger and more active projects, it is effectively full-time.09:57
cdentgreater than full-time for nova09:57
johnthetubaguyFWIW, its the emotional turmoil I found the worse09:58
* cdent nods09:58
*** dtantsur|afk is now known as dtantsur09:58
cmurphyjohnthetubaguy: can you elaborate?10:00
cmurphyI've never been a PTL or worked on nova10:00
johnthetubaguyits the sense of responsibility, and the number of people who want to talk to you about how to make it better, vs your ideas of what you want to change vs getting "the job done"10:01
persiaI think that comes with any product management role, although for PTL we merge product management, project management, and (sometimes) line management (where orgs contributing staff give them no direction).10:02
johnthetubaguyso I always went back in time a little, to when people spoke about creating the PTL role...10:04
johnthetubaguythe bit I remember was, we elect someone to make decisions when there is no clear consensus, and we just have to get on with things10:04
johnthetubaguynot, that accidentally maps to, make sure everything keeps moving forward10:05
cmurphyI think it's probably gotten harder in the last year or so, there are fewer people working on openstack and the people who are here are more and more part-time, so fewer people step up and it all falls on the PTL to get things done10:06
cmurphythat's how it is in keystone10:06
johnthetubaguycmurphy: ++10:06
johnthetubaguyyeah, I was crazy lucky with people taking on things for me10:06
johnthetubaguythere are less folks with time to do that now :'(10:06
johnthetubaguyanyways, that just means its time to tweak things so it works10:07
johnthetubaguy... question is how?10:07
johnthetubaguyso my take it, we could get more aggressive splitting code out of Nova, but frankly that is a massive amount of work, and there is no one to do it vs getting fixes and features in10:08
johnthetubaguyplacement was always my best hope there10:08
johnthetubaguyI should say, what is now placement10:09
johnthetubaguy... going back to the week need help lists, do we need to add something about helping with project management10:10
johnthetubaguys/week need/we need/10:10
persiaI think product management is the larger load than project management for most of the teams in OpenStack.  PTL burden isn't so much about deciding what to do when , etc., but more about dealing with the 10K people who think they need PTL to bless them to do their jobs.10:12
johnthetubaguyyeah, it was a poor choice of words there10:14
cmurphyI think it's kind of the opposite. PTL burden is supposed to be about being a final decision maker and contact point but it's really about keeping track of everything in your head and keeping everyone moving10:15
johnthetubaguyPTL has now power or control on when something happens, only gets to say when something can't happen, it gets you down after a while.10:15
persiacmurphy:  What is the driver for keeping everyone moving?10:15
cdent"if you do this now, I won't say no to it later"10:16
cdentthus the PTL often needs to be aware of all the "this"10:16
johnthetubaguyyeah, turns out to you need to know all those things to make the decisions10:16
johnthetubaguylargely due to so few people having the visibility into things to be able to decide things10:17
*** liujiong has quit IRC10:17
johnthetubaguythat isn't because the info is private, its more because so there is so much of it openly available10:17
johnthetubaguyI found blueprint tracking hard to give away, for the reasons cdent mentioned10:18
persiaIndeed: even just as decider-of-last-resort, the PTL needs to keep context of nearly all activity current.10:18
johnthetubaguyits an interesting blend of product and project management10:18
johnthetubaguyI tell the story of the elephant and three blind men alot, I guess someone is tracking all the animals in the zoo10:20
johnthetubaguyI guess the ask is ideas on a postcard?10:20
johnthetubaguycmurphy: you asked about goals, the mutable configuration one from gcb excites me the most10:23
cmurphythat one does seem interesting10:25
cmurphyand achievable10:25
johnthetubaguywith cdent's suggestion I could see that getting done10:25
johnthetubaguyI like how it opens up operators to add patches for any mutable options they might want10:26
cdentHas there been any real motion on the interop testing situation since cmurphy re-booted the discussion? I've been tracking the thread and the new comments on the review, but I don't really see a clear picture emerging10:37
cmurphyit feels like it's still in a standstill10:40
cmurphyQA team seems to be saying they prefer option 2 but they haven't taken a hard line against option 110:41
cmurphywhich seems to contradict mugsie's experiences10:42
*** kumarmn has joined #openstack-tc10:44
persiaI suspect there is a difference between what some folk might say to an individual or in a review and what they might post to a ML thread that includes folk perceived as authority.10:46
*** kumarmn has quit IRC10:49
cdentcmurphy: in case it's not clear, I mostly agree with you on the upgrade situation, but I'm trying to push the point that we can't keep adding tasks without adding people, and the upgrades one is a relatively easy one to use for that10:57
cmurphycdent: that's fair10:58
*** dtantsur is now known as dtantsur|brb11:50
*** dtantsur|brb is now known as dtantsur12:23
*** kumarmn has joined #openstack-tc13:00
*** kumarmn has quit IRC13:04
*** diablo_rojo has joined #openstack-tc13:15
*** dims has quit IRC13:22
*** alex_xu has quit IRC13:24
*** dims has joined #openstack-tc13:25
*** alex_xu has joined #openstack-tc13:27
*** rosmaita has joined #openstack-tc13:31
fungimissed office hour due to travel, but i'm starting to wonder if the disconnect on the interop testing discussion might be that teams slated for inclusion are pushing for the okay to move their interop tests into the tempest repo even though the board resolution for those add-on trademark programs hasn't happened yet13:46
*** kumarmn has joined #openstack-tc13:47
fungionce the board okays the two proposed add-on programs, the prior tc mandate about interop test location indicates they should go in tempest proper as far as i can tell13:48
cmurphyI think the concern is that the past discussion around what will happen after the board approval has given the impression that the tc mandate is going to be ignored13:54
*** diablo_rojo has quit IRC13:55
mugsiefungi: well, according to that resolution, the interop team should not have even looked at tests not in the tempest repo13:55
mugsiecmurphy: ++13:55
fungibut _until_ the board gives the go ahead, the qa team is expected to include them at that point14:07
fungii'll need to reread it. does it say services with no tests already in the tempest repo can't be considered for a trademark?14:07
fungimeant to say, but _until_ the board gives the go ahead, the qa team is not necessarily expected to include them in anticipation of that14:08
fungii thought those services just needed to have the tc:approved-release tag to be considered for inclusion in a trademark program14:09
cdentit confuses me that there are people other than the tradermark/interop/defcore related people involved in writing or reviewing interop tests14:10
* persia feels that it is important to allow folk to look at things in an unlimited way, but only sharply limit the things they must examine14:11
*** dklyle has quit IRC14:22
*** diablo_rojo has joined #openstack-tc14:24
mugsiefungi: currently on the side of a mountain :) but I will double check, but afaik the resolution says to only consider tests in tempest for interop14:27
*** hongbin has joined #openstack-tc14:52
*** kumarmn has quit IRC15:00
*** kumarmn has joined #openstack-tc15:01
fungimugsie: if so, it sounds like we may have designed in a deadlock (need tests in tempest to be considered for a trademark program but need to be in a trademark program before there's a reason to include their tests in tempest)15:17
dhellmannmugsie, fungi : what we discussed at the time was that the interop group would identify tests from whereever they live and then they could *use* those tests after they were added to the tempest repo. So the move needs to happen before the board approval.15:25
*** diablo_rojo has quit IRC15:28
*** david-lyle has joined #openstack-tc15:40
*** coolsvap has quit IRC15:55
*** openstackgerrit has quit IRC16:03
*** openstackgerrit has joined #openstack-tc16:52
openstackgerritMerged openstack/governance master: Add charm-neutron-api-genericswitch  https://review.openstack.org/53160516:52
openstackgerritMerged openstack/governance master: Add tempest-stress in Quality Assurance  https://review.openstack.org/53302616:56
openstackgerritMerged openstack/governance master: Update policy goal for solum  https://review.openstack.org/53344016:56
openstackgerritMerged openstack/governance master: Add networking-generic-switch-tempest-plugin under ironic  https://review.openstack.org/53418016:57
*** jpich has quit IRC17:39
*** david-lyle has quit IRC18:07
*** diablo_rojo has joined #openstack-tc18:27
*** dtruong has quit IRC18:34
*** dtantsur is now known as dtantsur|afk18:48
*** diablo_rojo has quit IRC18:48
*** cdent has quit IRC18:55
*** david-lyle has joined #openstack-tc19:04
fungidhellmann: yeah, in that case i agree it likely needs a rethinking since we're basically saying the tempest team has to take additional tests in-tree in advance of a board decision which may or may not happen (and then possibly need to revert that if the decision is against or deferred?)19:34
*** harlowja has joined #openstack-tc19:34
fungithat also basically turns the temest team into gatekeepers of the list of projects for trademark programs, and puts additional pressure on them since they're the first stop on the road to approaching the interop team and the board19:35
fungiwhich i doubt is what they wanted19:36
dhellmannfungi: yeah. I'm starting to feel like the only one who thinks the current solution is the right approach and that this is all a lot of talk over something that should just be simple to do.19:42
dhellmannas long as we come up with something that ensures that we don't have folks who don't understand the extra criteria that need to be applied to interop tests making "bad" changes to them, I'll be happy19:43
dhellmannit's not clear that any of the other options give us that as they're currently described, though19:43
dhellmannmaybe we just need better reviewer guidelines for those and then we have to train all of the teams who are participating in any trademark program to follow them19:44
*** kumarmn has quit IRC21:01
*** kumarmn has joined #openstack-tc21:01
*** kumarmn has quit IRC21:02
*** kumarmn has joined #openstack-tc21:02
*** kumarmn has quit IRC21:20
*** kumarmn has joined #openstack-tc21:20
*** kumarmn has quit IRC21:21
*** kumarmn has joined #openstack-tc21:21
*** kumarmn has quit IRC21:27
*** kumarmn has joined #openstack-tc21:28
*** kumarmn has quit IRC21:32
*** kumarmn has joined #openstack-tc21:47
*** kumarmn has quit IRC22:41
*** kumarmn has joined #openstack-tc22:42
*** kumarmn has quit IRC22:47
*** kumarmn has joined #openstack-tc23:03
*** hongbin has quit IRC23:25
*** kumarmn has quit IRC23:32

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