Tuesday, 2018-06-19

*** annabelleB has joined #openstack-tc00:37
*** annabelleB has quit IRC00:47
*** ian_ott has quit IRC01:42
*** david-lyle has joined #openstack-tc03:00
*** mriedem has quit IRC03:02
*** ricolin has joined #openstack-tc03:06
*** david-lyle has quit IRC03:09
*** diablo_rojo has quit IRC03:18
*** kumarmn has joined #openstack-tc03:33
*** kumarmn has quit IRC03:56
*** kumarmn has joined #openstack-tc04:16
*** kumarmn has quit IRC04:21
*** jaosorior has joined #openstack-tc04:32
*** kumarmn has joined #openstack-tc06:23
*** kumarmn has quit IRC06:28
*** jpich has joined #openstack-tc08:03
ttxRe: not in the US, There is LF Open Source Summit EU in Edimburg, same week as AllThingsOpen.08:15
*** kumarmn has joined #openstack-tc08:24
*** kumarmn has quit IRC08:29
*** e0ne has joined #openstack-tc08:34
*** cdent has joined #openstack-tc08:37
*** ricolin has quit IRC08:47
*** jaosorior has quit IRC08:49
cdentahoy hoy tc-members and everyone else: it's office hours09:00
cdent#startmeeting tc09:00
openstackMeeting started Tue Jun 19 09:00:20 2018 UTC and is due to finish in 60 minutes.  The chair is cdent. Information about MeetBot at http://wiki.debian.org/MeetBot.09:00
openstackUseful Commands: #action #agreed #help #info #idea #link #topic #startvote.09:00
*** openstack changes topic to " (Meeting topic: tc)"09:00
openstackThe meeting name has been set to 'tc'09:00
cdent#chair ttx cmurphy09:00
openstackCurrent chairs: cdent cmurphy ttx09:00
cdentwe talked about maybe not using meetbot, but haven't decided yet, so there it is09:00
cdentanyone here?09:01
ttxI have a topic!09:02
ttxOpenStack Days China is this week, and they are as always hungry for directions on where to contribute to make a difference. So we need to make sure our help-most-needed list is up to date09:03
ttxIs there anything you'd remove from it ? add to it ?09:03
ttxStoryBoard is a bit struggling with contributions, at a point where we'd like more of OpenStack to migrate to it09:04
ttxnot sure it counts as strategic though09:04
ttxBarbican might also need some support, as we make it a base service option09:05
cdentI think if we ever want storyboard to be a thing it needs more of a sense of team and centrality to the org. The tools we use are as (or more) important than what we are building with them.09:05
ttxShould we keep infra/doc/glance/designate on the list ?09:05
cmurphyi think we brought up osc as a possible candidate for the list09:06
ttxyeah, I'm worried the current team is more of one person doing it on their free time, combined with a number of Foundation staff trying to help keep it afloat09:06
cdentI haven't got enough insight into those projects without speaking to the associated folk.09:06
cdentcmurphy: osc is a good one09:06
cmurphythe keystone team also keeps floating the idea of requesting to be on the list, though objectively i think we're in a better situation than osc or glance or designate09:07
ttxWe won't be able to change the list in time for the event (it's this week), but I can supply a list of "extra candidates"09:08
ttxosc is a good one yes09:08
cdentMatt R and I sent quite a bit of useful nova- and placement- related stuff to the associated hackathon09:08
ttxThe "doc owner" entry in the list sounds a bit past, related to the Doc transition that was completed since09:10
ttxIf we want to keep it there, we should probably refresh the text09:10
* cdent nods09:10
ttxAny other hot suggestion ?09:10
ttx(I'll ask again later today when Murica is up09:11
cdentcmurphy: I'm curious about the keystone situation: is it that there aren't enough people to do what's on the plate, or that there's a desire to have the plate bigger and thus more people are needed?09:11
cdentI ask because I would have thought that keystone, being, uh, key, would be less likely to suffer than others09:12
cmurphycdent: right, keystone being so integral to other projects means we're saddled with requests from not just end users but other openstack projects09:14
ttxI'm also not totally convinced by the "Goal champions" on that list, as it sounds less directly actionable than other items.09:14
cmurphyfixing policy is the main thing09:14
cmurphythe quota management is the other big request09:15
cdentah, so it is kind of the standard: hey we need a thing do it! without the accompanying "and here's some people to help"?09:15
*** jaosorior has joined #openstack-tc09:15
ttxcdent: I've come to think it's a side-effect of the vertical silos we built. It's seen as "someone else's problem" to make it happen09:16
cdentttx I agree. I don't think that's sufficiently concrete. However the reason I think it is on there, if I remember right, was because we wanted to draw attention to the fact that goal champions are important, valued, required, etc. If we remove it from this list we should make sure it remains highlighted somewhere else.09:16
cdentPerhaps by the job reqs that got talked about at the last combined leadership meeting09:16
ttxcdent: yeah09:16
cdentyes, I agree with that interpretation of the silos09:16
ttxOne of the multiple reasons why I'd like to reorganize work around objectives rather than code boundaries09:17
cdentme too, but that's a _shift_09:17
ttxoh yes.09:17
ttxI think that's key to the long-term health of the project, but I'm also doubtful we can manage such a transition :)09:18
cdentan untrained ear would think you just said, "we're doooooomed"09:18
ttxorganizing work across code boundaries was great in the "abundance" period. Organizing work around specific objective / functions across code boundaries is a lot better when you have more limited resources09:19
ttxcdent: I'll keep on advocating for it. One soft way to drive it is to promote more of the SIG approach09:21
ttxBut we need to fix the ML situation first09:21
cdentI think in the meantime there's a lot more aspirational writing we could be doing (as individuals and as the official tc)09:22
ttxAlso reminds me of that discussion around having people caring about the compute node, since it runs much more than Nova code those days09:22
* cdent nods09:22
ttxThe SIG approach is gradual, but tricky as people see it as signing up for more work09:23
ttxbut we digress09:23
cdentwere we on a path?09:24
ttxhelp most needed list -- but maybe we are done with suggestions there09:24
cdentah, I felt like we were still on that path, at least tangentially.09:25
cdentFrom my standpoint when it comes to "help" we keep coming back to two things:09:26
ttxcdent: I welcome your input on https://etherpad.openstack.org/p/project-team-guide-basic-design-tenets-chapter09:26
cdentkey projects struggling to keep up with requests (what cmurphy just said, basically)09:26
cdentand, needing people with license to operate in a silo-violating (cross-project) fashion09:27
cdentwhile asking for help on specific instances of those issues moves things along, it's cultural change that will fix them09:27
cdentwill stick that etherpad in my look-soon list09:28
ttxcdent: re: TC get-together, saw my suggestion of Edimburgh above ?09:33
cdentah, no, haven't caught up yet09:35
cdentand just discovered I'm late for an appointment I thought was tomorrow :(09:35
ttxrun rnu09:36
* ttx does some project health review09:45
*** dtantsur|afk is now known as dtantsur09:46
*** openstack changes topic to "OpenStack Technical Committee office hours: Tuesdays at 09:00 UTC, Wednesdays at 01:00 UTC, and Thursdays at 15:00 UTC | https://governance.openstack.org/tc/ | channel logs http://eavesdrop.openstack.org/irclogs/%23openstack-tc/"10:08
openstackMeeting ended Tue Jun 19 10:08:16 2018 UTC.  Information about MeetBot at http://wiki.debian.org/MeetBot . (v 0.1.4)10:08
openstackMinutes:        http://eavesdrop.openstack.org/meetings/tc/2018/tc.2018-06-19-09.00.html10:08
openstackMinutes (text): http://eavesdrop.openstack.org/meetings/tc/2018/tc.2018-06-19-09.00.txt10:08
openstackLog:            http://eavesdrop.openstack.org/meetings/tc/2018/tc.2018-06-19-09.00.log.html10:08
*** dtantsur is now known as dtantsur|brb10:20
*** kumarmn has joined #openstack-tc10:25
*** kumarmn has quit IRC10:30
cdentthanks for killing the meeting cmurphy10:47
openstackgerritMerged openstack/project-team-guide master: Release management chapter updates  https://review.openstack.org/57543411:40
*** dtantsur|brb is now known as dtantsur11:44
openstackgerritMerged openstack/project-team-guide master: Clarify Code-Review 0 etiquette  https://review.openstack.org/57620711:45
*** edmondsw has joined #openstack-tc11:50
*** kumarmn has joined #openstack-tc12:21
*** kumarmn has quit IRC12:28
*** kumarmn has joined #openstack-tc12:32
*** kumarmn has quit IRC12:33
*** kumarmn has joined #openstack-tc12:33
*** kumarmn has quit IRC12:37
*** kumarmn has joined #openstack-tc12:39
*** mriedem has joined #openstack-tc12:53
cdent[t KYZ]12:58
purplerbot<fungi> should i protest travelling into the uk until brexit is repealed? ;) [2018-06-18 17:34:14.715517] [n KYZ]12:58
fungireciprocation of moral support for each of our struggles against fascist oppression13:14
smcginnisIf there's a chance for Edinburgh, it would be good to decide that soon.13:25
smcginnisSelfishly speaking since I'm supposed to give a talk at ATO, but would much rather prioritize a TC gathering over doing that.13:25
ttxLF OSS EU CFP is still open fwiw13:32
smcginnisCFP ASAP FTW13:38
ttxacronym overload!13:46
mnaseridk my bffl jill?13:46
*** hongbin has joined #openstack-tc13:53
*** annabelleB has joined #openstack-tc13:55
*** ian_ott has joined #openstack-tc14:00
ttxdhellmann: thoughts re: help-most-needed pipeline ?14:01
dhellmannttx: OSC and Barbican seem like good things to add.14:22
dhellmannI'm on the fence about what to do for storyboard. I'm not sure I would want someone working on that instead of anything else on the list.14:22
dhellmannalthough I guess if someone was interested that wouldn't be a bad thing; it's just difficult for me to consider prioritizing it myself, and I think that's part of why we're having trouble recruting14:23
dhellmannit's really a shame we can't use another tool that is already supported well14:23
dhellmanns/can't/won't/ maybe?14:23
dhellmannas far as the goal champions, I have recruited some helpers for the python 3 work14:24
dhellmannI'm not sure what approach we would want to take for another goal14:24
dhellmannso maybe mentioning the goals as projects that the community is working on, rather than asking for champions in general, is the best thing?14:25
zanebI think the solution for storyboard would be to do a zuul14:25
dhellmannwe could, but there aren't people, right?14:25
*** annabelleB has quit IRC14:25
zanebget other people interested beyond just us, so they could share the development burden14:25
dhellmannI suppose. We need the existing folks to have the time, energy, interest, and funding to do that and I worry we missed the window of opportunity.14:26
*** e0ne has quit IRC14:27
*** e0ne has joined #openstack-tc14:27
smcginnisIt's a balancing act where you need some interest to get to where it can live on it's own.14:28
jrollI suspect there aren't many folks out there looking for a project management tool that needs work to meet their needs14:28
smcginnisI guess it would take some marketing type work to see how SB fills gaps that other existing tooling doesn't to get people outside of OpenStack interested in using it.14:29
ttxdhellmann: re: StoryBoard, one thing is that it's a slightly different skillset, so could attract different people14:29
smcginnisBut that seems like an already crowded space.14:29
ttxdhellmann: I agree I would not add it to "the list", but i can make a case to mention it in the OSD China-requested list14:30
ttxsince it's a bit more subjective14:31
dhellmannI feel like we would probably have been better off compromising on our "requirements" so we could make another tool work than building our own thing. And now we're in a state where we don't have anyone willing to operate another self-hosted tool *or* to finish the development work.14:31
dhellmannttx: yeah, I don't mind if you mention it, but I would put it last in case someone views the list as being in priority order.14:31
ttxoh sure. Make it a personal addition.14:32
jrolldhellmann: I agree with you14:32
dhellmannI *like* storyboard now that I've figured out how to use it. It has many nice aspects.14:32
ttxWhat's frustrating is that most of the rough edges are details, mostly in the presentation layer14:32
dhellmannwhich makes the situation even harder, since I want it to succeed14:33
ttxBut those are tricky to fix due to the framework used that's a bit alien to us14:33
ttx(same as if you asked me to work on Horizon)14:33
dhellmannyes, that's true. there are probably 1 or 2 data layer things, but most of the issues I encountered were in the API or UI14:33
persiaWhile I agree that we're in a terrible place regarding tracking systems, I assert that it isn't so much just the requirements that were the issue, but also the lack of consensus around the requirements: multiple development teams that were previously active on Storyboard have left in part due to confusion of multiple different parties requesting things, no clear roadmap, and a regular lack of adoption.  WIth the adoption that has happened, I would14:33
persia expect the next team to last longer (as there are actual users providing actual input on useful features, etc.)14:33
dhellmannpersia : that's a fair point, too14:34
zanebwhere is the next team coming from, though14:39
ttxdhellmann: did someone look into Barbican health already ? Maybe it's in no need of help, I don't know14:39
dhellmannI haven't had anyone report on it, yet14:39
dhellmannttx: that's assigned to you and fungi14:40
ttxah. ok :)14:40
ttxOK, so I propose to point them to the official list, say we are considering OSC for addition soon, Barbican might need help after adding base service, and a personal add with StoryBoard if you have spare JS developers (who should work on Horizon too)14:40
ttxI can make that a pledge for JavaScript/UI people14:41
ttxGiven the deadline, I think it looks good14:41
fungiin fairness, we evaluated (and even nearly pulled the trigger at least once) on switching to an already maintained solution written by someone else14:42
fungithe main challenge is that most communities/organizations who write defect and task trackers don't write them to be cross-project14:42
fungiso we had to try to come up with creative workarounds14:42
*** annabelleB has joined #openstack-tc14:43
mnaserwild idea: maybe storyboard's better place shouldn't live under openstack but under 'osf'14:43
fungithe trackers which have a concept of tracking multiple projects don't really let you have activities which span projects14:43
mnaser(it's an idea, i'm not supportive or against it)14:43
smcginnismnaser: That may get more visibility.14:43
fungimnaser: well, it's under infra which is working on splitting out from openstack14:44
mnaseryeah, i think that will help because storyboard probably feels like "an openstack thing" for the outsider14:44
fungiinterestingly, the people who resurrected development on it a few years ago (when the openstack community had all but left it for dead) have no ties to openstack at all and wanted it for its own existing features14:45
fungithe lead dev on it still really has no connection to the openstack community other than through work on developing storyboard14:46
fungibut he unfortunately has less and less time to dedicate to it, which is why diablo_rojo, i and a few others have been increasing involvement14:47
*** annabelleB has quit IRC14:53
*** annabelleB has joined #openstack-tc14:53
*** dtantsur is now known as dtantsur|afk14:54
*** dklyle has joined #openstack-tc15:41
*** e0ne has quit IRC15:44
*** e0ne has joined #openstack-tc15:44
*** dklyle has quit IRC15:46
*** dklyle has joined #openstack-tc16:11
*** knikolla has quit IRC16:19
*** jpich has quit IRC16:21
*** kumarmn has quit IRC16:44
*** kumarmn_ has joined #openstack-tc16:48
*** kumarmn_ has quit IRC16:49
*** kumarmn_ has joined #openstack-tc16:50
*** kumarmn_ has quit IRC16:54
*** knikolla has joined #openstack-tc16:56
*** annabelleB has quit IRC17:18
*** e0ne has quit IRC17:19
*** annabelleB has joined #openstack-tc17:24
*** e0ne has joined #openstack-tc17:29
*** e0ne has quit IRC17:34
*** hongbin has quit IRC17:56
*** dklyle has quit IRC18:10
*** kumarmn has joined #openstack-tc18:17
*** e0ne has joined #openstack-tc18:31
*** dklyle has joined #openstack-tc18:32
*** annabelleB has quit IRC18:32
*** annabelleB has joined #openstack-tc18:35
*** kumarmn has quit IRC18:51
*** kumarmn has joined #openstack-tc18:51
*** kumarmn has quit IRC18:56
dhellmannzaneb : is your impression from the adjutant folks that their current proposal de-emphasizes or removes the "grab bag" aspect that was causing objections in past versions? I see "extensible" in there and I'm wondering if that's just indicating that they plan to add to the API over time or if they still want deployers to install their own APIs.18:59
openstackgerritMerged openstack/governance master: add champion section to goal template  https://review.openstack.org/57593419:03
zanebdhellmann: yes, de-emphasising the grab-bag aspect was the goal19:04
zanebdhellmann: discussion here: https://etherpad.openstack.org/p/Adjutant_Features19:05
dhellmannzaneb : ok, I got tripped up on that word "extensible"19:14
dhellmannI think we expect to be able to add to all of our APIs so I wasn't sure why that was included specifically19:15
zanebtbh I hadn't really given much thought to that word19:15
dhellmannI'm trying to work out if I'm being overly sensitive :-)19:15
zaneblot of projects are extensible though (Horizon, Heat, Mistral come to mind)19:16
dhellmannyes, well, there's history with that issue and this project, right?19:16
*** e0ne has quit IRC19:16
dhellmannwe also ask teams to remove things like "fast" and "performant" or whatever, because of course those are expected aspects of what they're doing19:16
dhellmannno one ever things to include "bug free" in these missions :-)19:17
*** e0ne has joined #openstack-tc19:17
cdentextensible is a difficult word. from a nova standpoint it is a "thing we don't allow any more"19:18
zanebthey may be fine with dropping that word. they're certainly not going to change how it works depending on whether that word's in there19:18
dhellmannok, well, having them actually change the way it works was the thing that was tripping me up. I thought we'd reached agreement on that.19:18
zanebthere's always a fine line between allowing and encouraging extensions19:18
zanebdhellmann: changing how?19:20
dhellmannI thought they were going to lock the API down to a well-defined set of things19:22
dhellmannperhaps I misunderstood the outcome of the discussion from the summit19:22
zanebdhellmann: I understood the outcome to be that they would lock down the *mission statement* to a defined set of things, and have another discussion with the TC before expanding to anything outside that set19:23
dhellmannI'm not sure what the point of saying they're going to do one thing and then doing another really is.19:24
*** e0ne has quit IRC19:24
zanebhold on, we're talking about two different things19:24
zanebone thing is what APIs do they support upstream as 'core' plugins19:26
zanebwe're committing them to a limited scope for those in the mission statement, and requiring them to come back to us before trying to expand into any other areas19:27
zanebthe other thing (what you're talking about) is whether it's easy or hard to add new APIs without getting them into upstream19:27
dhellmannI agree those are somewhat separate things.19:28
zanebI don't expect them to rewrite it to make it technically hard to create new APIs, whether or not it says 'extensible' in the mission statement19:28
zanebany more than they would try to make it slow or buggy because it doesn't say 'fast' or 'bug-free' in the mission statement ;) (though that may be a good argument for just not saying it)19:29
dhellmannLimiting the "upstream" scope vs. the actual feature set just makes it feel open core or something.19:29
dhellmannThe thing I have been objecting to is the emphasis on allowing deployers to use an official project to implement custom APIs that are somehow required to actually use a given deployment. Because that gives those deployers an opportunity to break all of the interoperability work we've done in the community while still claiming to run "OpenStack".19:31
dhellmannde-emphasizing that, without actually changing it, is just changing the marketing of the project and not actually changing the approach19:31
dhellmannnow I may have the minority view on this, and may be reading it more strictly than the rest of the tc19:32
dhellmannI think ttx, for example, pointed out it fit into the "operations" bucket rather than the "openstack" bucket, on his map19:32
dhellmannbut unless we say operations projects cannot be considered for the interop trademark program, I'm not sure that distinction makes any real difference19:32
dhellmannso I guess I'll wait and see what some of the rest of the tc thinks about it. I will support the majority, even if I vote against them.19:34
cdentI'm saving that (and most other contemplative reviews) for when I'm sitting19:34
zanebdhellmann: I guess a good question to answer would be: what technical change would you like to see to prevent people from writing custom workflows for it?19:35
zanebI honestly can't think of one other than using bad design to make it hard for people19:37
zanebwe're not allowed to use the GPL so that isn't an option19:37
dhellmannI haven't looked into the source to see how it actually loads the workflows, so I couldn't comment on implementation details.19:38
zaneb(me neither for the record)19:39
*** e0ne has joined #openstack-tc19:39
dhellmannit feels like, from what I've heard, that it's a fundamental detail of the implementation (if not, it seems to be a fundamental aspect of their design approach or goals or whatever)19:40
zanebbut in general, designing software to make it easy for you to extend is good, and when that software is written in Apache-licensed Python that almost invariably makes it easy for anyone else to extend as well19:40
dhellmannI'm making a distinction between building something to make it easier for a contributor or maintainer to extend it and building something to make it easy for an end user to extend it19:41
dhellmannand fwiw, if I set out with their initial goals, I imagine I would have done something very much like what they've done.19:42
zanebdhellmann: I think their major design goal is to make it easy to swap out the *implementation*, because all business processes are different19:42
zanebI imagine that making it easy to add a new API probably mostly fell out of that19:42
dhellmannchanging the implementation of a well-defined API seems OK. adding new arbitrary APIs is the issue I have.19:43
clarkbglance task api is agood example of why this is bad19:45
clarkb(if anyone needs an example)19:45
* cdent retires19:48
cdent'night all19:48
*** cdent has quit IRC19:48
*** annabelleB has quit IRC19:51
*** annabelleB has joined #openstack-tc19:56
*** dklyle has quit IRC20:00
*** diablo_rojo has joined #openstack-tc20:24
*** e0ne has quit IRC20:46
*** edmondsw has quit IRC20:47
*** ian_ott has quit IRC20:54
fungismcginnis: well, i just got an update from ato saying my talk has _not_ been rejected (yet) and they're still trying to finalize the lineup21:53
smcginnisfungi: Well, I guess that's kind of good news.21:54
fungithey're apparently "still in the process of contacting potential speakers" and expect to take another 1-2 weeks21:54
smcginnisSo kind of a wait list for speaking. That's an interesting way to do it.21:55
smcginnisI suppose there's plenty of time yet.21:55
fungimaybe i'm what our summits would consider an "alternate" and they're waiting to see who from the first round of picks bails on them ;)22:01
fungii mainly want to give this talk because i gave it a couple years ago at oscon and, unlike our conferences, o'reilly wants to charge people a subscription to be able to watch the video22:03
smcginnisI didn't realize that.22:03
fungithe first 3 minutes can be streamed but then you need a paid subscription to their digital content service to watch the full talk22:04
fungibecause, you know, they're the _open_ source conference22:04
smcginnisI had a past employer that paid for Safari subscriptions, but I'm still too old school and would rather just buy the dead trees.22:04
smcginnis"The open core conference"22:05
fungisame (though i've also stopped feeling comfortable supporting their publishing business in recent years after a variety of shenanigans)22:05
smcginnisIt has been awhile since any recent books of mine have had animals on them...22:05
clarkbhrm the kids are probably old enough for me to pull out all the old dead trees again22:05
fungiit was cool back in the days of "running linux" (mainly because the original author was a school chum who first introduced me to linux in the 0.95 download source from usenet and bootstrap it from minix days)22:07
fungibut as the years have gone on, they've seemed more and more money hungry there22:07
fungirunning linux sort of set the tone for their business model though, thinking back on it22:09
fungibasically it was mostly gfdl-licensed content from the ldp compiled into dead-tree form you didn't need to tear the tractor feed edges off22:10
fungi1. take open source/free content, 2. slap a cover on it, 3. profit22:11
smcginnisAnd companies have been trying to replicate that ever since. :)22:11
*** ian_ott has joined #openstack-tc22:41
*** dklyle has joined #openstack-tc22:43
*** david-lyle has joined #openstack-tc22:45
*** david-lyle has quit IRC23:04
*** annabelleB has quit IRC23:32
*** diablo_rojo has quit IRC23:35
*** diablo_rojo has joined #openstack-tc23:42

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