Monday, 2018-05-14

*** kumarmn has joined #openstack-tc01:42
*** zhipeng has joined #openstack-tc02:34
*** kumarmn has quit IRC02:44
*** kumarmn has joined #openstack-tc02:45
*** kumarmn has quit IRC02:49
*** zhipeng has quit IRC02:53
*** ricolin_ has joined #openstack-tc03:13
*** zhipeng has joined #openstack-tc03:44
*** kumarmn has joined #openstack-tc03:57
*** kumarmn has quit IRC03:57
*** kumarmn has joined #openstack-tc03:58
*** kumarmn has quit IRC03:59
*** kumarmn has joined #openstack-tc04:00
*** kumarmn has quit IRC04:00
*** zhipeng has quit IRC04:01
*** kumarmn has joined #openstack-tc04:50
*** kumarmn has quit IRC04:54
*** kumarmn has joined #openstack-tc04:54
*** kumarmn has quit IRC04:55
*** kumarmn has joined #openstack-tc04:55
*** kumarmn has quit IRC05:00
*** dtantsur|afk is now known as dtantsur08:17
*** alex_xu has quit IRC08:32
*** alex_xu has joined #openstack-tc08:33
*** cdent has joined #openstack-tc08:37
cdentttx, mugsie, cmurphy : any of you recall the subject or sender of the post-ptg email that included summit codes?08:45
ttxno... but maybe fungi does as he was involved08:48
ttxdhellmann: re Arch WG vs. championing change: I agree that most of the time we are missing people willing to push the change through, more than we miss up-front design or gap analysis. That said I don;' think it's either/or: I think we need both. You need to paint a direction and you need people willing to push such change through.08:51
ttxThe Arch WG discussion *were* useful. It just came at a time of resource constriction when everyone was trying to do the same with less. So there was limited participation and limited follow-up08:53
ttxSo we need to balance the two. Do *some* top architecture design, but not too much. I see the goals stuff as a tactical way to do that balance. Now we need to build a more strategic equivalent08:56
cdentI sympathize with doug's comments at [t 4jba] but I think it leaves out a critical aspect of things: we need to make it easier for people to stop doing the (potentially less useful) stuff they are already doing so they can do this other thing09:00
cdentpurplerbot sad?09:01
*** kumarmn has joined #openstack-tc09:01
*** purplerbot has quit IRC09:02
*** purplerbot_ has joined #openstack-tc09:02
*** purplerbot_ is now known as purplerbot__09:02
*** purplerbot__ is now known as purplerbot09:02
cdent[t 4jba]09:02
purplerbot<dhellmann> I think some of the frustration that I have with talking about this topic is that folks are focused on adding a layer of formality around a practice that ultimately comes down to an individual or small group picking an important topic, mustering support in a sufficient portion of the community, and then driving the work to completion. [2018-05-11 23:42:51.693008] [n 4jba]09:02
cdentIndividuals or small groups can't pick up additional work unless they can give some away09:02
cdentwe consistently fail to recognize this when we talk about contribution, making assumptions that there's either an endless supply of time or people09:03
*** kumarmn has quit IRC09:07
*** dtantsur is now known as dtantsur|brb09:08
*** jaosorior has joined #openstack-tc09:56
ttxcdent: right, our culture is built on an abundance model, and that is what we need to transition off from09:59
*** rosmaita has joined #openstack-tc10:00
cdentwell, it's not only that (although that's a big part of it). There are still quite a lot of people, but they are not well equipped/empowered to choose the right stuff to dismiss10:01
*** kumarmn has joined #openstack-tc10:02
*** kumarmn has quit IRC10:06
*** dtantsur|brb is now known as dtantsur10:38
dimscdent : because the agenda is set by their employer?11:00
dimso/11:00
cdentdims: no, I don't think so, or at least that's not all of it, I think because we get habituated to styles of doing stuff11:03
smcginnisIt is a factor though I think. Most of us our paid by employers who would like us focusing on certain areas. There are not a lot of people with the leeway to decided to ignore some of that stuff and work on other things.11:04
cdentAnd a bad info economy. It's not easy to determine or point to artifacts which say what matters11:04
cdentsmcginnis: sure, it's factor11:04
cdentbut if we had more artifacts that says what's important it would be easier to say "I ought to be working on this instead"11:05
smcginnisThat's true.11:05
cdentof course the flip side of the coin, at least based on my experience, is that there are plenty of managers who say "it would be great if we could work on this $general improvement, but openstack won't allow it"11:05
cdentthat's probably more common in the nova-related universe. in this case I'm think of my recent tweet about an extracted nova-compute11:06
cdentvarious folk at various employers have said that would be keen11:06
cdentbut feel like openstack itself is the roadblock11:06
cdentthere are (and always will be) many factors11:06
cdentbut by having artifacts (rather than talking about it) we can help the discussion move forward, I expect11:07
cdent(and we're getting better about that)11:07
*** kumarmn has joined #openstack-tc11:08
*** needsleep is now known as TheJulia11:25
*** kumarmn has quit IRC11:31
*** kumarmn has joined #openstack-tc12:02
*** kumarmn has quit IRC12:11
*** cdent has quit IRC12:11
fungicdent: ttx: no idea what the subject line for those e-mails was since i neither sent them no received any. i can ask around and find out if you're still having trouble12:45
mnasero/12:53
dhellmanntc-members: we have 2 open reviews waiting to set up the new constellations repository. Please take a moment to look at them; all of you have +2 on the repo. https://review.openstack.org/#/q/project:openstack/constellations12:53
fungiooh, thanks for reminding me to subscribe to that one12:54
dhellmanncdent: your point about visibility of priorities is good. The release team used to encourage using blueprints, but I'm not sure how much anyone actually looked at those. Maybe with the move to storyboard we can come up with a new pattern of building priority lists or status boards or some combination thereof.12:56
dhellmannalthough I'm not sure what kinds of things you're thinking of when you say we're doing lots of things we could drop12:56
*** dmsimard is now known as dmsimard|off12:58
*** dklyle has quit IRC13:02
*** dklyle has joined #openstack-tc13:03
*** dklyle has quit IRC13:06
*** mriedem has joined #openstack-tc13:06
*** zaneb has joined #openstack-tc13:09
*** dklyle has joined #openstack-tc13:12
*** dklyle has quit IRC13:17
*** cdent has joined #openstack-tc13:18
openstackgerritMerged openstack/constellations master: initial repository setup  https://review.openstack.org/56616913:28
dhellmannfungi : thanks13:30
fungiyw. do we have docs jobs pending for that repo too?13:31
openstackgerritMerged openstack/constellations master: clean up sphinx settings  https://review.openstack.org/56691513:32
*** superdan is now known as dansmith13:33
dimslooks like both merged!13:33
dhellmannfungi : I thought the jobs patch was approved, but let me check13:33
dhellmannyeah, I don't see it on my open list any more13:34
fungiokay, cool. maybe i should have rechecked that second one13:34
fungibuilt in the gate though http://logs.openstack.org/15/566915/2/gate/build-openstack-sphinx-docs/e778ea0/html/13:35
*** annabelleB has joined #openstack-tc13:36
cdentfungi: yeah, I'm still interested in those emails, if you're able to find someone who knows that would be great13:37
cdentdhellmann: on the "things we could drop" it's more that any to do list cannot be infiniely long. If we want to do other things, we must drop something13:38
cdentand in order to drop things we need visibility13:39
smcginnisAnd prioritization.13:39
dhellmanncdent : ok, that makes sense13:39
*** jgwentworth is now known as melwitt13:57
fungii'm still curious which classes of "things" you would propose we drop13:58
fungihard for me to wrap my head around the nebulous argument, need some specific examples13:59
cdentif there's capacity to do 4 things and the list is already 10 long and we want to make it bigger, two things (at best) need to happen: we need to remove some things, we need to order things by importance14:00
cdentthere is nothing revolutionary or extraordinary about that. it's simple resource management14:01
fungie.g., a lot of teams have been dropping extra "paperwork" like tracking implementation detail minutiae in specs, requiring blueprints for any feature no matter how tiny, requiring a "bug report" for every fix or even sometimes for non-fixes/features, strict adherence to 2x +2 rules for obvious or trivial changes or where one of the authors is already a core reviewer...14:01
cdentwhich things get removed is the result of agreement amongst the participants14:01
fungiso looking for what you mean by "thnigs"14:01
cdentfeatures14:01
cdentcode that is being written14:01
cdentgoals that are trying to be reached14:01
cdentgetting rid of some of those14:02
fungioh! actually dropping features from the services/software, not just tropping tasks14:02
cdentdropping features which are under development14:02
cdentbut dropping some that already exist is a good idea too14:02
fungiyeah, i've long felt that we've fallen into a trap of making the software too "featureful" by including rarely-used options et cetera. the more of those you accumulate the more you're stuck supporting, plus it adds to the overall cognitive load14:03
fungithe "simplification" focus we came up with in our pow-wow with the bod and uc back in boston touched on some of that14:03
cdentover in #openstack-placement we're currently discussing that when the todo list is a year long, there's no point talking about new things14:03
cdentbut I'm losing that discussion14:03
fungibut the initial gains were mostly realized in dropping entire projects which were seen as dead weight or confusing the scope of openstack as a whole14:04
fungiand then the effort petered out due to lack of drivers14:04
cdentbecause all the drivers were already busy driving <- and that's exactly my point14:05
cdentnice wrapup14:05
fungiyup14:05
*** hongbin has joined #openstack-tc14:05
fungi"it's okay to drop some of these in-flight things on the floor, or even declare roadmap bankruptcy and start over"?14:05
fungigo back to the beginning, redefine the problem, determine your minimal featureset for viability, et cetera14:06
smcginnisThis is the hard part I think with project leadership.14:08
smcginnisEven if the roadmap is already over a year out, you do need to at least discuss new features.14:09
smcginnisAt least to see if a new direction impacts the current plan or to make sure things done now don't make future needs hard/impossible to meet.14:09
smcginnisBut saying no is probably the hardest part.14:09
smcginnisNo, or "not right now"/14:09
cdentthat long term planning is especially true with "older" projects. In the case of the placement discussion mentioned above, we still to haven't any real data on it being used to its full capacity (that is, with nested and shared working) and I suspect things are going to change, a _lot_, when we do. So trying to plan is ... distracting.14:10
smcginnisFair point. If it's early in its evolution, long term planning can definitely become a distraction if it's so far down a road that you won't be following.14:12
*** annabelleB has quit IRC14:13
jrolldhellmann: I agree with you re: centralized design groups and such. thanks for those twitter links. I made some effort around dropping old microversions a while ago that would be helpful there, I honestly haven't checked back in on it much since my hiatus and don't really have time to drive that now :(14:15
*** cdent_ has joined #openstack-tc14:16
jrollcdent: in case you haven't found it yet, "Your Vancouver Summit Code: OpenStack PTG" from OpenStack Foundation <summit@openstack.org> via e2ma.net14:16
*** cdent has quit IRC14:17
*** cdent_ is now known as cdent14:17
cdentjroll: yeah, mugsie's bottomless email archive provided. but thanks. turns out it is likely a spam problem14:18
jrollah14:19
jrollcdent: re: placement usage data, just wait a couple years for people to actually go deploy it :) :( :( :(14:22
jrollthat's one of the harder parts in openstack leadership to me, knowing if you're doing the right thing when almost nobody is running the code for a year or two14:22
cdenttrue14:23
cdentbut it's not as bad as all that if we check with academia: a friend of mine runs a (big) openstack install at the university we both went to and they are upgrading to queens after summit14:23
cdent(the ptg code stuff is not for me, btw, colleague having issues)14:24
cdentand by usage of placement, I don't really even mean big installs, more "us". most of the code hasn't even merged yet14:25
jrollfair enough14:29
* jroll wonders if the uni is doing all the craziness NFV has driven into placement14:29
cdentthat's the craziness that's not done yet14:30
cdentand you have to be an optimistic think it will be "done" for rocky14:30
jrolls/rocky/a long long time/14:32
jroll(nothing against the people working on it, just seems like the target changes daily)14:33
cdentthus my desire to stop adding things to the list. the constant context and target switching means nothing happens14:34
cdentfinish a small piece, release it, see what happens. amazing, it's open source in action!14:34
jroll++14:34
* dhellmann looks forward to seeing the placement API using the cycle-with-intermediary release model for frequent releases14:37
cdentdhellmann: if we ever get it extracted I would _love_ to do that14:38
cdentbut there are currentyl no other resources besides me on that task14:39
*** annabelleB has joined #openstack-tc14:42
mnaserjroll: re: nobody is running the code for a year or two -- i get so excited when i hear about a customer who "needs queens"14:45
jrollheh14:45
jrolleverybody needs queens, only a select few will do it soon :P14:46
mnaserso i guess tech people don't really care as much as marketing, but we're pretty much running queens except for keystone (i have a few customers moving off v2 right now)14:47
cdentmnaser: you are a credit to the industry :)14:48
jrollthat's awesome14:48
mnaserthat's supposed to be announced next week (queens).. but it's pretty interesting that there are people who are actually looking for *latest releases*14:48
* jroll cares greatly and is decidedly not a marketing person14:48
mnaserand it's an enterprise customer so i'm about to hop on a call to know what features they need14:48
mnasersuper useful insight and i'd love to bring it back here to find out reasons we can get people to run newer code14:49
mnaserjroll: yeah, but the fact we're running queens is "embargo'd" by marketing but /me shrugs14:49
jrollheh14:49
openstackgerritDoug Hellmann proposed openstack/governance master: move the team badge generation to the end of the build  https://review.openstack.org/56830914:52
openstackgerritDoug Hellmann proposed openstack/governance master: Add os_blazar to OpenStack-Ansible  https://review.openstack.org/56654114:57
*** gcb has joined #openstack-tc15:01
fungicdent: your placement todo backlog is sounding similar to what we went through with zuul v3... it was in the works for so long with a fairly specific scope of target features/tasks that we had an increasing number of people already asking for features which would need to come after the release; at one point we basically just said to come back to us after the release to discuss them because trying to15:09
fungiplan for the next version while the current version wasn't even done yet was more of a distraction than anything15:09
cdentyeah15:09
cdentsounds very familiar15:09
fungiand i personally think that's fine15:10
cdentme too, but there's not agreement within the group (which is also fine, but a pita)15:10
fungiyou have only so much time and energy to invest in it. you can get the current stuff done, but discussing what comes next can get kicked down the road as long as you at least do what's necessary to try and avoid painting yourself into a corner (know vaguely what you might be thinking about beyond that event horizon)15:11
fungitoo much time spent designing what comes next will simply delay what you need to do now15:11
fungiit's an all too easy trap to fall into though. excessive planning can be a way to procrastinate on what you already know needs doing while rationalizing it away as productive15:13
*** dklyle has joined #openstack-tc15:14
* cdent nods15:14
*** dklyle has quit IRC15:22
*** gcb has quit IRC15:22
*** cdent has quit IRC15:43
*** annabelleB has quit IRC16:55
*** annabelleB has joined #openstack-tc17:00
*** kumarmn has joined #openstack-tc17:01
*** mugsie has quit IRC17:06
*** mugsie has joined #openstack-tc17:06
*** mugsie has quit IRC17:06
*** mugsie has joined #openstack-tc17:06
*** mugsie has quit IRC17:08
*** kumarmn has quit IRC17:10
*** mugsie has joined #openstack-tc17:12
*** dtantsur is now known as dtantsur|afk17:14
*** kumarmn has joined #openstack-tc17:14
*** kumarmn has quit IRC17:19
*** kumarmn has joined #openstack-tc17:19
*** kumarmn has quit IRC17:25
mordredtc-members: over the weekend we surfaced a desire to have a discussion at the forum about moving our APIs forward, bumping microversion mins and adding new major versions17:27
mugsieyeah - I have some thoughts on that :)17:27
mordreda) are people cool with adding such a topic and b) how does wednesday at 16:40 sound?17:27
mordredmugsie: me too :)17:27
dhellmannmordred : yes, let's get together17:28
dhellmannthat's a good time fo rme17:28
mugsiethat works for me as well17:28
fungimordred: the opendev ci/cd join session overlaps that, as does the session on requirements for contributing organizations (i'll likely be in the latter), but i suppose conflicts are likely to abound. i doubt i'm needed in the api discussion anyway17:29
mordredfungi: yah - missing the opendev ci/cd joint session was the 'best' conflict I had - and I figure we'll have coverage in that one anyway17:29
*** ricolin_ has quit IRC17:32
jrollmordred: glad to see y'all working on that :)17:35
*** cdent has joined #openstack-tc17:48
*** annabelleB has quit IRC17:58
*** annabelleB has joined #openstack-tc18:00
*** kumarmn has joined #openstack-tc18:04
*** kumarmn has quit IRC18:09
*** harlowja has joined #openstack-tc18:12
TheJuliamordred: that time also works for me, sounds like a good topic imho18:12
mordredjroll: ++18:15
mordredTheJulia: \o/18:15
jrollmordred: I won't be there but I fully support being able to kill a microversion18:16
jrollbonus points if you can kill the "micro" in the word "microversion"18:16
mordredjroll: oh - well ... I'm not sure that's *exactly* the stance I'll be advocating for, but having argued with you on the topic before I believe I can represent your point of view as well :)18:17
mordredjroll: (also, yeah - the name microversion is really unfortunate)18:17
mugsiejroll: my view is going to be, if we do a global major version cut, lets forget microversions ever existed, and say if OpenAPI cannot model it, its not an OpenStack API18:18
cdentmugsie: gee, I dunno, that's too small a step ;)18:19
cdentwhat's 16:40 vancouver in utc?18:20
jrollheh18:20
mugsieone small step for mugsie, a giant leap for OpenStack?18:20
clarkbcdent: 234018:20
clarkbor is it 0040 now18:20
clarkbone of the two18:20
cdentthanks clarkb, wasn't sure of the offset18:20
cdentha!18:20
clarkb234018:21
cdentyeah, figured it out too18:21
cdentmordred: can you assign me a roll or position for this conversation. I no longer know how I really feel18:23
jrollcdent: rabble-rouser work for you?18:23
mordredcdent: totally18:24
mordredcdent: you wanna be jroll?18:24
cdentwell that's the thing. usually it would, but these days, I dunno. I'm tired.18:24
mordredfair point18:24
jrollI like the idea of cdent being me for vancouver18:24
mordredmugsie: I'm going to disagree with you about that - but I think it'll be more fun to disagree with you on that in person in vancouver18:24
mugsiemordred: I figured, and yes, in person will be more fun :)18:25
cdentI'm thinking maybe I'll take the position of "whatever, but let's do _something_"18:25
TheJuliajroll: but if cdent is you for vancouver, that means he might be able to claim your beers18:25
cdent\o/18:25
jrollthat is fine18:25
mordredmugsie: although maybe we should find some beers before wednesday and disagree with each other over alcohol first18:25
cdentdouble \o/18:25
mugsiemordred: ++18:25
jrollI'll have 3 months before PTG to build up my beer IOUs again18:25
* cdent is looking forward to being as young and attractive as jroll18:25
cdentand hopefully as capable of handling beer18:25
cdentIf I do anything stupid it's jroll's fault18:25
* jroll :o18:26
mordredmugsie: I guess we could just sit next to each other at whatever mixer happens after the board/tc/uc meeting - abuse the foundation's alcohol for the purposes18:26
mugsiecdent: "jroll" can be your alterego18:26
cdentHello stackers, I'll be your jroll today.18:26
TheJuliamordred: this sounds acceptable ;)18:26
mugsie"what did jroll do last night?"18:26
mugsiemordred: yeah - I think that is a fine way to abuse the foundations alchohol :)18:27
jrollwhat have I done18:27
cdentworse: what will you do?18:27
* jroll starts writing his apology letters18:27
TheJulialol18:27
*** annabelleB has quit IRC18:28
TheJuliajroll: in the words of jayf, 'doooooommm'18:28
*** amotoki has joined #openstack-tc18:29
*** annabelleB has joined #openstack-tc18:30
cdentOkay, I've adjusted my schedule so that both jroll and cdent will be present at an api bump session at 16:40 on wednesday18:31
TheJuliacdent: wait... both?!?18:32
mugsiecdent: the balmer peak?18:32
cdentI'm a complicated communicator18:32
*** Guest16323 is now known as mgagne18:40
*** mgagne has joined #openstack-tc18:40
*** annabelleB has quit IRC18:47
*** dklyle has joined #openstack-tc18:49
*** annabelleB has joined #openstack-tc18:52
*** diablo_rojo has joined #openstack-tc18:57
*** annabelleB has quit IRC19:00
smcginnisTheJulia: That was you wondering about Sunday registration, right?19:06
smcginnisJust saw the checklist email "Registration is open on Sunday 1-6pm"19:06
TheJuliasmcginnis: I suspect you might have me mixed up with someone else19:07
smcginnisTheJulia: Ah, OK. Sorry. :)19:07
smcginnisSomeone was in here wondering about Sunday registration last week I thought.19:07
*** kumarmn has joined #openstack-tc19:12
*** annabelleB has joined #openstack-tc19:17
*** kumarmn has quit IRC19:19
*** guvnah has joined #openstack-tc19:22
*** guvnah has quit IRC19:22
*** cdent has quit IRC20:26
*** kumarmn has joined #openstack-tc20:27
*** dklyle has quit IRC20:31
*** kumarmn has quit IRC20:32
*** kumarmn has joined #openstack-tc21:03
*** diablo_rojo has quit IRC21:06
*** kumarmn has quit IRC21:08
*** diablo_rojo has joined #openstack-tc21:21
*** annabelleB has quit IRC21:27
*** kumarmn has joined #openstack-tc21:30
*** diablo_rojo has quit IRC21:32
*** kumarmn has quit IRC21:34
*** annabelleB has joined #openstack-tc21:39
*** harlowja has quit IRC21:56
*** dklyle has joined #openstack-tc21:59
*** diablo_rojo has joined #openstack-tc22:00
*** dklyle has quit IRC22:08
*** kumarmn has joined #openstack-tc22:14
*** kumarmn has quit IRC22:19
*** mriedem has quit IRC22:22
*** hongbin has quit IRC22:40
dhellmannthey usually have our badges and swag ready at the board meeting22:43
dhellmanns/board/joint leadership/22:43
*** dklyle has joined #openstack-tc22:50
openstackgerritMerged openstack/governance master: move the team badge generation to the end of the build  https://review.openstack.org/56830922:56
openstackgerritMerged openstack/governance master: Add os_blazar to OpenStack-Ansible  https://review.openstack.org/56654122:56
*** dklyle has quit IRC22:56
*** kumarmn has joined #openstack-tc23:23
*** kumarmn has quit IRC23:28
*** dklyle has joined #openstack-tc23:35

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