| *** annabelleB has joined #openstack-tc | 00:05 | |
| *** mriedem_afk has quit IRC | 00:15 | |
| *** david-lyle has quit IRC | 00:34 | |
| *** annabelleB has quit IRC | 00:43 | |
| *** edmondsw has quit IRC | 01:19 | |
| openstackgerrit | Ian Wienand proposed openstack/project-team-guide master: Diagram explaining constraints layers https://review.openstack.org/568994 | 01:22 |
|---|---|---|
| *** ricolin has joined #openstack-tc | 01:45 | |
| *** rosmaita has quit IRC | 02:06 | |
| mnaser | ttx: link to the talk you’re giving that you’ve suggested to attend in regards to the focus areas etc? | 02:36 |
| *** lbragstad has quit IRC | 02:39 | |
| *** kumarmn has joined #openstack-tc | 02:39 | |
| *** kumarmn has quit IRC | 02:46 | |
| *** kumarmn has joined #openstack-tc | 02:46 | |
| *** kumarmn has quit IRC | 02:47 | |
| *** kumarmn has joined #openstack-tc | 02:47 | |
| *** kumarmn has quit IRC | 02:47 | |
| *** edmondsw has joined #openstack-tc | 02:59 | |
| *** edmondsw has quit IRC | 03:04 | |
| *** kumarmn has joined #openstack-tc | 04:24 | |
| *** kumarmn has quit IRC | 04:37 | |
| *** edmondsw has joined #openstack-tc | 04:48 | |
| *** edmondsw has quit IRC | 04:53 | |
| *** kumarmn has joined #openstack-tc | 05:08 | |
| *** kumarmn has quit IRC | 05:12 | |
| ttx | mnaser: https://www.openstack.org/summit/vancouver-2018/summit-schedule/events/20968/beyond-datacenter-cloud-the-future-of-the-openstack-foundation | 07:38 |
| ttx | mriedem: two remarks: some projects prefer to reinvent secret storage rather than depend on Barbican so there is clearly a catch-22 there... (2) The idea was to add "a castellan-compatible secret store" rather than Barbican, and have Castellan support Vault and other secret stores | 07:42 |
| ttx | There were two issues with that strategy: 1/ castellan had to be made more neutral, move under Oslo etc. -- I think that's mostly done now. 2/ Some projects (Designate ? Octavia?) depended on specific features of barbican that you don't find in other secret stores (like cert generation ?) | 07:44 |
| ttx | and we wanted to avoid some project having a strong dependency on Barbican, if the goal was to make Castellan the interface | 07:45 |
| *** edmondsw has joined #openstack-tc | 08:24 | |
| *** dtantsur|afk is now known as dtantsur | 08:26 | |
| *** edmondsw has quit IRC | 08:29 | |
| *** dims has quit IRC | 09:20 | |
| *** dims has joined #openstack-tc | 10:07 | |
| *** edmondsw has joined #openstack-tc | 10:13 | |
| *** edmondsw has quit IRC | 10:17 | |
| *** dims has quit IRC | 10:17 | |
| *** edmondsw has joined #openstack-tc | 10:52 | |
| *** jaosorior has quit IRC | 11:01 | |
| *** cdent has joined #openstack-tc | 11:01 | |
| *** jaosorior has joined #openstack-tc | 11:05 | |
| *** rosmaita has joined #openstack-tc | 11:07 | |
| mugsie | Designate wanted to use Barbican for signing DNSSEC hashes | 11:16 |
| mugsie | (a security review goes a lot easier if you can draw a line around an external service and say "the private key never leaves $HSM") | 11:17 |
| mugsie | and we didn't want to implment crypto signing in designate | 11:18 |
| mugsie | but with the deployment profile of barbican, we will have to just nice use castelen and retrieve the key and sign ourselves | 11:19 |
| cdent | Aye, there's the rub. | 11:21 |
| cdent | if we were a bit more monolithic (in definition, not design) it would be easiser to just say "the next version of openstack has barbican" done | 11:22 |
| mugsie | yeah - I have seen "new in Queens posts" that make it seem like OpenStack is one big monolith though | 11:30 |
| *** lbragstad has joined #openstack-tc | 11:33 | |
| *** mriedem has joined #openstack-tc | 11:57 | |
| *** cdent has quit IRC | 12:03 | |
| *** jaosorior has quit IRC | 12:12 | |
| *** jaosorior has joined #openstack-tc | 12:19 | |
| *** cdent has joined #openstack-tc | 12:35 | |
| *** dims has joined #openstack-tc | 12:43 | |
| fungi | i don't think it's necessarily just one "we" in that scenario | 12:47 |
| fungi | the project as a whole can declare that "openstack now includes barbican" but that doesn't mean deployers and operators will necessarily install it unless forced to do so | 12:48 |
| fungi | (either through projects they want growing a strict dependency on it thereafter, or insistence in trademark programs with which they want to comply) | 12:49 |
| fungi | and then we're back to the same "well sure we say it's part of openstack but can't really depend on it because almost nobody's installing it in their environments" catch-22 | 12:51 |
| fungi | but i do agree our strong focus on modular design has underscored for operators of just how many discrete services openstack is composed, and so every time we mention this class of features we've added they see it as yet another service which needs care and feeding | 12:53 |
| fungi | rather than necessarily seeing it as additional functionality | 12:54 |
| mnaser | i think it's because they're all separately configured, running their own processes, needing their own configuration, their own endpoint, etc | 13:01 |
| ttx | Dentist appointment 45 min before TC office hour today, might be late | 13:09 |
| dhellmann | tc-members: that API session mordred is organizing has been added to the forum schedule: https://www.openstack.org/summit/vancouver-2018/summit-schedule/events/21881/api-debt-cleanup | 13:32 |
| smcginnis | Wow, that time even works for me. Awesome. | 13:33 |
| mugsie | fungi: sure. But the trademark program has always been a trailing standard, rather than a progressive vision one | 13:45 |
| mugsie | but then again, I have opinions on that side interop, so others may see it as a net good thing | 13:45 |
| *** rosmaita has quit IRC | 13:48 | |
| *** rosmaita has joined #openstack-tc | 13:52 | |
| openstackgerrit | Merged openstack/project-team-guide master: Update dependency-management doc https://review.openstack.org/568883 | 13:56 |
| openstackgerrit | Merged openstack/project-team-guide master: Update Documentation per project team docs migration https://review.openstack.org/544514 | 13:56 |
| *** ianychoi_ has quit IRC | 13:59 | |
| *** annabelleB has joined #openstack-tc | 14:01 | |
| smcginnis | I need to check out and get ready to go, so I may miss office hours today. | 14:06 |
| cdent | smcginnis: https://www.youtube.com/watch?v=accEKtBebaU | 14:07 |
| smcginnis | Hahah | 14:07 |
| dhellmann | I think I'm glad I don't know what that clip is from | 14:10 |
| smcginnis | All I remember is thinking it was a very odd movie at the time. | 14:11 |
| cdent | dhellmann: donnie darko. One of those rare movie where the theatrical release is probably better than the director's cut. | 14:11 |
| dhellmann | ah | 14:15 |
| *** ianychoi has joined #openstack-tc | 14:18 | |
| ttx | I survived the dentist successfully and am back | 14:35 |
| *** rosmaita has quit IRC | 14:50 | |
| * fungi personally thinks darko is an excellent film | 14:51 | |
| * fungi likely has strange taste in cinema | 14:52 | |
| cdent | I really like it too, thus my confusion about the director's cut | 14:52 |
| *** dklyle has joined #openstack-tc | 14:54 | |
| mtreinish | I'm not sure if I've ever seen the director's cut | 14:56 |
| *** annabelleB has quit IRC | 14:57 | |
| *** annabelleB has joined #openstack-tc | 14:58 | |
| cdent | mtreinish: as I recall, it tries to flesh out the why of it all a bit more | 14:59 |
| cdent | which ruins some of the wtf and mystery of it | 15:00 |
| cdent | especially if you already had your own (different) interpretation | 15:00 |
| cdent | tc-members seems to be office hours | 15:00 |
| *** annabelleB has quit IRC | 15:00 | |
| TheJulia | o/ | 15:00 |
| ttx | it does | 15:00 |
| mnaser | o/ | 15:00 |
| mugsie | o/ | 15:00 |
| dhellmann | o/ | 15:00 |
| ttx | Do all TC members attend the Summit ? Or are we missing some ? | 15:01 |
| cmurphy | morning | 15:01 |
| ttx | I know cdent misses the Sunday meet for a good reason | 15:01 |
| EmilienM | o/ | 15:01 |
| ttx | I know EmilienM does not travel far | 15:01 |
| EmilienM | I have to walk 20 min :( (long walk really) | 15:02 |
| ttx | Sounds healthy | 15:02 |
| TheJulia | sounds like it would be invigorating | 15:02 |
| EmilienM | :) | 15:02 |
| ttx | EmilienM: when do you get the waterplane certification? | 15:02 |
| johnsom | FYI based on this mornings/last nights discussion. Octavia supports either Barbican or Castellan as of Queens. | 15:03 |
| mnaser | excuse me i have to fly 6 hours so :-) | 15:03 |
| EmilienM | hehe, not this year | 15:03 |
| ttx | johnsom: great! | 15:03 |
| EmilienM | so nothing happens on Saturday evening? | 15:03 |
| mnaser | johnsom: oh cool, i looked around the code, saw barbican references but wasn't sure about castellan | 15:03 |
| ttx | Enjoy the free evening | 15:03 |
| EmilienM | well if anyone wants to see a taste of Vancouver night life, just ping me | 15:04 |
| johnsom | mnaser Yeah, it is in the queens release notes. | 15:04 |
| johnsom | It wasn't much of a change for us because we were involved in the development of Castellan. | 15:04 |
| mnaser | johnsom: cool, that pretty much makes it that most projects i reviewed are making use of castellan | 15:04 |
| fungi | though per mugsie's point, some services certainly want an hsm and it sounds like not all castellan-compatible backends implement one? | 15:05 |
| ttx | Any last-minute question I can help answer ? I'm traveling toorrow | 15:05 |
| ttx | +m | 15:05 |
| mnaser | fungi: yeah, but castellan drivers can just have some supported features and others that arent | 15:05 |
| mugsie | yeah, I get squirrly about private keys being in memory of the service, but that may just be me | 15:06 |
| dhellmann | has everyone had a chance to review the slide deck on contribution stats? | 15:06 |
| fungi | i wonder whether castellan could also expose things like key generation backed by different solutions than key storage (so to the service the needs are still through a single pane even if behind the scenes your key is being shuffled between a couple subsystems) | 15:06 |
| ttx | If you attend a forum session that clearly would benefit from a continuation session, you can remind the moderator that they can schedule one if they manage to find a suitable free spot in the 3rd room | 15:06 |
| johnsom | Anchor was supposed to be the CA/generator, but I don't think it is still staffed | 15:07 |
| ttx | If they find such a slot, they can email speakersupport@openstack.org so that it gets added to schedule | 15:07 |
| *** annabelleB has joined #openstack-tc | 15:07 | |
| TheJulia | I did review the slide deck, but I have mixed feelings about the exclusion of some data because it can be gamed. It doesn't feel like a huge problem right now, but that is just my perception. | 15:07 |
| fungi | so if you're using castellan to talk to barbican then it asks barbican to generate and store the key, but if you're configured to use vault then castellan calls out to something else (openssl?) to do jey generation and you simlpy document that it's not a nearly so paranoid implementation under those circumstances | 15:08 |
| mnaser | TheJulia: depends on the project, i can say that some projects get a lot of "tag along +1" .. but i do agree that it might take away from perfectly valid non-cores saying "i think this is good" | 15:08 |
| EmilienM | dhellmann: I did have a look and have no useful comment, it lgtm | 15:08 |
| dhellmann | EmilienM : thanks | 15:09 |
| ttx | dhellmann: One thing I noticed is the pie chart labels are a bit random, but I guess the goal is not to talk about individuals | 15:09 |
| johnsom | The big driver for us was Castellan support Vault as a backend | 15:09 |
| dhellmann | ttx : yeah, google docs doesn't give a lot of control over that. I considered dropping the names but thought it would be ok to highlight a few specific people | 15:10 |
| dhellmann | I think their layout algorithm tries to only show labels when lines or text wouldn't cross | 15:10 |
| ttx | yeah it's ok as long as you mention it I guess | 15:10 |
| cmurphy | dhellmann: the proposal and release bots are being counted as contributors, is that intentional? | 15:11 |
| dhellmann | TheJulia : that's fair. This is all an attempt to sift out some useful info, so different interpretations are valid | 15:11 |
| dhellmann | cmurphy : yeah, I decided to leave those in to show (a) automation does a lot (b) sometimes more than people (c) we still need people to deal with what the automation produces | 15:11 |
| cmurphy | ++ | 15:12 |
| dhellmann | ttx: do you think it would be better to remove names from the graphs? | 15:12 |
| TheJulia | I feel like names might help convey individual stories which makes data more relatable... possibly | 15:12 |
| fungi | johnsom: yeah, i mean i can understand the benefit to having multiple supported backends there, but for services which could benefit from a full-blown hsm it would be nice if they could at least count on the necessary features for that even if it means a combination of a couple of complimentary "backends" rather than a single integrated backend providing that full set of features | 15:13 |
| ttx | dhellmann: no.. I think it stresses the point better | 15:13 |
| dhellmann | ok | 15:13 |
| dhellmann | yeah, I thought seeing a few names over and over might help drive the point home, too | 15:13 |
| ttx | just say the chart shows a few names only | 15:13 |
| dhellmann | ok, yeah, that's a good point | 15:13 |
| fungi | "and some of those names aren't even people" | 15:13 |
| dhellmann | haha | 15:14 |
| fungi | i personally feel like that sends an even stronger signal | 15:14 |
| dhellmann | I'll call that out specifically on the designate slide, since it shows up well there | 15:14 |
| dhellmann | and the proposal bot is the 3rd highest contributor | 15:15 |
| fungi | you could consider the generated commits to be contribution from our nifra donors, i suppose | 15:15 |
| fungi | er, infra donors | 15:15 |
| jroll | is there a public link to the slides, out of curiousity? or TC-only for now? | 15:15 |
| mnaser | jroll: https://docs.google.com/presentation/d/1wsPfeGC83I5C8s6-9tlsxz49bBRY5PHdIdXGxutVqgA/edit#slide=id.g39ba90574e_0_84 | 15:15 |
| jroll | thanks | 15:15 |
| dhellmann | jroll : it's linked on the agenda but yeah, thanks mnaser | 15:15 |
| mugsie | dhellmann: yeah - slides look acurate | 15:16 |
| * jroll always has trouble finding the agenda, too | 15:16 | |
| mugsie | it doesn't look good for designate though :/ | 15:16 |
| mnaser | jroll: https://wiki.openstack.org/wiki/Governance/Foundation/20May2018BoardMeeting :) | 15:16 |
| jroll | mnaser: heh, thank you | 15:16 |
| dhellmann | mnaser a.k.a. link-bot | 15:17 |
| mnaser | i just have to dig through the 60 or so open tabs | 15:17 |
| dhellmann | mugsie, yeah, it doesn't look good. I'm hoping a visual will drive that home, since we've been saying so for a while now | 15:17 |
| mnaser | dhellmann: how easy is it to generate the data for another period of time? | 15:17 |
| dhellmann | the tools are there, they just take some time to run | 15:18 |
| mnaser | maybe showing 2016 vs 2017 might draw a much more important point | 15:18 |
| dhellmann | it's all in the goal-tools repo | 15:18 |
| dhellmann | undocumented, of course | 15:18 |
| jroll | mugsie: oof, it does not. probably also worth calling out that timsim doesn't work on designate anymore (I assume) | 15:18 |
| mugsie | yeah, neither does Kiall | 15:18 |
| jroll | jeeeez | 15:19 |
| jroll | so 2 out of the top 3 humans are out there | 15:19 |
| mnaser | i think having 2016 vs 2017 can help show both that some projects might be losing contributors (designate) and others that are headed by the same people year over year | 15:19 |
| mugsie | dhellmann: core group is actually 3 now - kiall + timsim are no longer there | 15:19 |
| mnaser | which could mean a pretty big dependency | 15:19 |
| dhellmann | mnaser : I thought about that. I wasn't sure whether focusing on a decline would be better or just saying "here's where we are today" | 15:19 |
| dhellmann | I want to leave lots of time for actual discussion about what to *do* about this, too | 15:20 |
| dhellmann | mugsie : ok, I got that from gerrit but I'll update the speaker notes, thanks | 15:20 |
| mnaser | i've had a really awesome experience on puppet-openstack picking up two new contributors which one of them recently brought to core. i think we should try and work more with PTLs about picking up early hints of new contributors and welcome/support them | 15:20 |
| dhellmann | I think that would be a good thing for us to work on, as a group | 15:20 |
| mugsie | dhellmann: it would be worth noting that 2 of those 3 are new, and came up to core by digging in | 15:21 |
| dhellmann | I know we have onboarding sessions at the forum, but do we actually help teams prepare for those so they can be effective? | 15:21 |
| dhellmann | mugsie : that's a wonderful point, thanks, I'll add that | 15:21 |
| mugsie | One thing I was wondering as a topic for Sunday (more likely in the evening event) - the LCOO | 15:23 |
| dhellmann | my plan is to run through these as quickly as possible to highlight the general trends and then get to the conclusion slide and ask the room for advice about how we can bring focus on these areas | 15:23 |
| dhellmann | I know cdent also wanted to talk about giving folks time to be leaders and mentor, so maybe someone can talk a bit about that when we get to the discussion section | 15:23 |
| dhellmann | mugsie : yeah, the agenda is already very very full | 15:24 |
| dhellmann | that's a good topic, but it came up too late for this meeting | 15:24 |
| *** rosmaita has joined #openstack-tc | 15:24 | |
| dhellmann | I think we're going to want to work with the UC folks on that directly first, as ttx suggested | 15:24 |
| mugsie | dhellmann: yeah, the "allow people to grow" issue is very related to the presentation | 15:24 |
| dhellmann | maybe you can get together with mrhillsman to plan a meeting over lunch one day or something? | 15:24 |
| mugsie | yeah - it just seems everytime we ask about this it gets stonewalled | 15:25 |
| dhellmann | yeah, let's do it in person and let's talk with the UC leadership about how to move forward | 15:26 |
| *** eeiden has joined #openstack-tc | 15:26 | |
| ttx | mugsie: Might be an easier discussion to have directly with UC members rather than calling out companies at the Board | 15:26 |
| ttx | I'm also fine driving the point wit my Foundation hat on | 15:26 |
| * dhellmann would like to see ttx's hat rack | 15:27 | |
| TheJulia | dhellmann: So the foundation did supply a more specific list to PTLs for this cycle that suggested format an things to cover, also conveying what attendees should learn from the prior sessions | 15:27 |
| dhellmann | TheJulia : oh, good! I didn't realize that. | 15:27 |
| mugsie | Oh, I don't want to call out any companies - I would prefer to encourge them to move back to defaulting to more open | 15:27 |
| ttx | So that it does not appear like the TC is disciplining the UC on openness | 15:27 |
| TheJulia | dhellmann: it surely uses space folding technology | 15:27 |
| mugsie | ttx: sure, I get that - the optics may not look great | 15:27 |
| dhellmann | ttx: that's a good point | 15:28 |
| dhellmann | TheJulia : ++ | 15:28 |
| ttx | to be fair if the UC dropped them from the page nobody would probably notice/care :) | 15:28 |
| * ttx brb | 15:30 | |
| fungi | i gather mordred may also, in his capacity as a board member, be interested in an lcoo discussion | 15:31 |
| dhellmann | yeah. we should start with some facts gathering. maybe there are already plans afoot to fix things. | 15:31 |
| TheJulia | I'm not sure it is a matter of calling out companies, but helping companies understand that they may need to go back and evaluate if they are over burdening people and the effect of that (which as we all know, is ultimately burnout) | 15:32 |
| fungi | right, my biggest concern for the lcoo is that they're basically setting themselves up for irrelevance (but as their activities are fairly opaque, it's hard to know whether they're accomplishing the goals they set out for or not) | 15:33 |
| dhellmann | TheJulia : oh, I read the calling out comment as related to the LCOO thing, but your point relates well to the contributor stats discussion, too | 15:33 |
| smcginnis | I haven't been watching, but I haven't seen any changes either from that group. | 15:33 |
| dhellmann | I had originally put together charts showing which companies were contributing in these areas, rather than which people | 15:33 |
| dhellmann | and those charts were going to make for a very awkward conversation | 15:33 |
| smcginnis | And my impression from the last comments I saw was they were annoyed that anyone would suggest they need to change. | 15:33 |
| TheJulia | dhellmann: and here I wasn't even aware the topic had shifted so much... downside of.... 3 or 4 concurrent conversations I'm trying to track at the moment :( | 15:34 |
| * dhellmann would like to see TheJulia's hat rack, too | 15:34 | |
| mugsie | yeah, I don't want to call them (companies) out on people provided, but I think it is important that we know the stats, for planning and backchannel discussions | 15:34 |
| fungi | smcginnis: since the lcoo is made up of "member companies" and contributions tend to come from individuals, it's tough to map one to the other and say whether there's been any direct influence from the lcoo (positive or negative) | 15:34 |
| TheJulia | dhellmann: lol | 15:34 |
| smcginnis | TheJulia: Only 3 concurrent topics is pretty good for us. ;) | 15:34 |
| smcginnis | fungi: That's true. | 15:34 |
| dhellmann | I should write some user docs for that who-helped tool | 15:35 |
| dhellmann | maybe after the summit | 15:35 |
| smcginnis | dhellmann: Separately from specific project contributions, would it be useful to have a slide that at least alludes to the fact that there is a disproportionate number coming from certain companies? | 15:36 |
| dhellmann | maybe. that's a double edged sword. | 15:36 |
| TheJulia | Likely the other thing that needs to also be kept in mind with contribution stats is while we are all scattered, realistically it also takes 2 more people to achieve anything, and those contributors may have lower stats that are hidden in the numbers because of all of the varying priorities | 15:36 |
| dhellmann | I imagine you could all guess which companies appear at the top of that list | 15:36 |
| dhellmann | you might be surprised at the degree of disparity | 15:37 |
| dhellmann | I was | 15:37 |
| dhellmann | and there are two ways that can be balanced: others doing more, or the top contributors doing less. | 15:37 |
| dhellmann | TheJulia : yeah, it would be interesting to pull just the approval numbers | 15:37 |
| dhellmann | I did that for the nova nfv stuff | 15:38 |
| dhellmann | although I didn't include it | 15:38 |
| smcginnis | Top contributors doing less would probably have a big impact on those numbers. | 15:38 |
| dhellmann | the core reviewers, of course, did most of the reviews. that didn't seem like a surprise | 15:38 |
| dhellmann | smcginnis : indeed | 15:38 |
| dhellmann | I like cdent's idea of having our more experienced folks work as force multipliers for some of the other contributors | 15:39 |
| dhellmann | I'm trying that with some oslo work this cycle with mixed results | 15:39 |
| cdent | I'm do mental experiments along the lines of "If you have +2 on a project you may not write features" | 15:39 |
| cdent | you can only review, bug fix, or tech debt clean | 15:40 |
| dhellmann | mixed in that the work is good, but taking longer than I expected. I have not yet worked out why that is the case. | 15:40 |
| cdent | I don't ever expect to see that happen but it is a good mental exercise | 15:40 |
| cdent | It would help, greatly, to create useful voids | 15:40 |
| TheJulia | cdent: interesting concept | 15:41 |
| fungi | i've been trying really hard to remind myself to spend my time providing guidance and mentoring, and not taking on tasks that i know others could do, but ultimately it's hard for me to not feel like i'm failing to pull my own weight so i still tend to give into temptation and volunteer to fix or run things | 15:41 |
| cdent | same | 15:41 |
| TheJulia | same | 15:41 |
| lbragstad | fungi: same | 15:41 |
| cdent | (except I don't have +2 on what I primarily code on, so I guess it doesn't apply) | 15:41 |
| dhellmann | something that extreme may end up incentivizing people to drop off the review team; asking for them to do *more* reviews than feature writing might be something we could actually do | 15:41 |
| * fungi is a victim of the "american work ethic" disease | 15:41 | |
| TheJulia | fungi: me as well | 15:42 |
| dhellmann | fungi : yeah, it's especially hard when you see exactly how to do it, but I'm trying to consider that I need to train my replacement(s) | 15:42 |
| fungi | we talk about the variety of cultural challenges contributors from different parts of the world face, and i suppose this challenge is endemic to my culture | 15:43 |
| dhellmann | we need a pithy slogan for that | 15:43 |
| TheJulia | Mentally I think, at least in the back of my brain, that it would be useful if we could delegate code approvals outside of a core team, distribute the workload a little better because there are always code paths that people may have looked at the week before and have extremely fresh in their memory where as a core may not have looked at a path in upwards of a year | 15:43 |
| dhellmann | TheJulia : what do you think stops us from doing that? | 15:43 |
| *** dklyle has quit IRC | 15:43 | |
| TheJulia | dhellmann: trust issues inherent to humans, lack of understanding of others context and motivations, lack of ability to do path acls in gerrit to say "This group can do whatever they want with y folder) or some sort of magical upvoting logic | 15:44 |
| fungi | delegating code review is fairly easy in that you can take +1 from a trusted mentee as if it were a +2. the infra team has done that from time to time but generally ends up turning those people into core reviewers on the specific repositories where they're reviewing eventually | 15:45 |
| fungi | (and "eventually" tends to be on the order of weeks, not months or years) | 15:45 |
| cdent | how refreshing | 15:45 |
| dhellmann | TheJulia : this feels like something one person needs to demonstrate as workable. Like, if *you* trust someone you could just take their +1 as approval and go put your +2 on the patch as-is. | 15:46 |
| cmurphy | ++ we're always free to make judgment calls on a case-by-case basis, we don't need to always have explicit acls and formality around it | 15:46 |
| dhellmann | when we try to make cultural changes like this we've done best to have examples to point to as successful | 15:46 |
| fungi | infra is probably in a more unique position though, insofar as it has lots of repos in a variety of subject areas, so those lighter-weight review silos are easier to build up | 15:46 |
| dhellmann | (or failures to learn from) | 15:46 |
| TheJulia | I think some people do it, but without fostering a culture to do it, it is easy to derail. The cycle counts we sometimes see also nukes that very quickly when people seek perfection. | 15:46 |
| cdent | fungi: that's (repo size) a huge part of it | 15:47 |
| fungi | yup | 15:47 |
| dhellmann | fungi : oslo has worked similarly in the past | 15:47 |
| fungi | i can see where it's tougher for monolithic codebases | 15:47 |
| dhellmann | although these days we're pretty liberal with drafting people into the main core team and trusting them to be careful | 15:47 |
| dhellmann | the oslo code is also a good bit more stable, so there tends to be less churn to be concerned about | 15:48 |
| *** dklyle has joined #openstack-tc | 15:48 | |
| dhellmann | that article I linked to a day or two ago talked about the amount of change as a stress factor leading to burnout | 15:48 |
| fungi | there's occasionally been the suggestion that we should arrange acls by subtree within a repository, but if your code is already that cleanly delineated then it's barely a step away from being able to decompose it into different repositories anyway | 15:48 |
| TheJulia | cmurphy: I guess my fear is some contributors see any case by case judgement call not aligning with policy because it was not an explicitly two or more core reviewers involved. | 15:48 |
| dhellmann | for a completely unrelated presentation I pulled some numbers about how many code files were unchanged in nova in the last few years, does anyone want to guess the %? | 15:49 |
| lbragstad | dhellmann: do you remember where you dropped the link to that article? | 15:49 |
| dhellmann | lbragstad : let me find it for you | 15:49 |
| dhellmann | it was here in channel, but I should have it handy | 15:49 |
| fungi | dhellmann: i'll guess 95% | 15:49 |
| fungi | if you're going by file count | 15:49 |
| dhellmann | lbragstad : https://www.fastcompany.com/40569522/these-are-the-reasons-why-your-whole-team-is-burning-out | 15:49 |
| lbragstad | dhellmann: thanks | 15:49 |
| dhellmann | fungi : files unchanged, or changed? | 15:50 |
| fungi | unchanged | 15:50 |
| * TheJulia is going to guess for 83% | 15:50 | |
| dhellmann | over the time period I looked, and eliminating empty files, and focusing only on .py files, 0 of them had not been changed in some way | 15:50 |
| fungi | wow | 15:50 |
| cdent | you have to change files all over nova because it is massively coupled | 15:51 |
| cdent | massively | 15:51 |
| TheJulia | wow is right | 15:51 |
| TheJulia | dhellmann: out of curiosity have you done the same with ironic? | 15:51 |
| dhellmann | so yeah, the next thing I want to do is gather more detailed stats over the lifetime of projects, per release, how much change they've seen | 15:51 |
| dhellmann | I didn't. I needed 1 example for a RH presentation to customers to explain to them why carrying patches downstream is bad practice. | 15:52 |
| TheJulia | dhellmann: ahh! | 15:52 |
| fungi | granted, i was thinking more like changed sufficiently that the copyright year merited updating, but i suppose most teams are really bad about ever updating copyright headers | 15:52 |
| dhellmann | let's say the stats were more in my favor than I expected | 15:52 |
| cdent | ha! | 15:53 |
| dhellmann | yeah, this was a very coarse grained check: how many of the same git shas appear in the tree for the current version vs. the older version | 15:53 |
| cdent | they may have been in your favor, but they point out some other issues | 15:53 |
| smcginnis | dhellmann: "why carrying patches downstream is bad practice" - that would be a very good thing to publish. I've had the conversation more than once. | 15:53 |
| dhellmann | git diff --stat was producing numbers that showed more files being changed than existed, so I need to work that one out | 15:53 |
| TheJulia | Agreed, it would be a great thing to publish | 15:54 |
| smcginnis | Maybe including removed files? | 15:54 |
| dhellmann | smcginnis : kernelcdub's keynote at the RH summit this year touches on it | 15:54 |
| dhellmann | smcginnis : yeah, that was my theory, and maybe I was interpreting the output incorrectly | 15:54 |
| * TheJulia puts on her cat staff hat and wishes it was feasible to take her cat with her to Summit | 15:55 | |
| dhellmann | I was also asking for the diff stat over a rather long period of time | 15:55 |
| cdent | TheJulia: the couch sig encourages the presence of cats | 15:55 |
| dhellmann | TheJulia : I was just thinking yesterday that it would be nice if we had a quiet room | 15:55 |
| TheJulia | couch sig ++ | 15:56 |
| dhellmann | which isn't the same thing, but is more easily portable | 15:56 |
| smcginnis | Couches and cats and quiet. Sounds good to me. | 15:56 |
| dhellmann | and a faraday cage | 15:57 |
| ttx | Last time in vancouver the outside areas were pretty good for that | 15:57 |
| dhellmann | yeah, I'm hoping they're not too crowded this time around | 15:57 |
| mugsie | yeah, the design summit balcony was ++ | 15:57 |
| TheJulia | Yeah, I seem to remember a nice dead zone by the tap and barrel | 15:57 |
| cdent | Does working on the couch sig count as helping to create leaders? | 15:58 |
| dhellmann | cdent : definitely | 15:58 |
| * TheJulia has effectively derailed the discussion \o/ | 15:58 | |
| TheJulia | Well, my cat did | 15:58 |
| ttx | Oh well, time to pack | 15:59 |
| ttx | TheJulia: my cat usually does its best hiding in my cabin luggage, so we might see it attend summit one day | 15:59 |
| TheJulia | The suitcases are in the bedroom, the cats are already upset :( | 16:00 |
| johnsom | Roughly following the conversation due to competing meetings, but wanted to throw my two cents in. The biggest issue I have in the Octavia projects is the reviews are highly weighted to the cores. People don't seem to understand that +1 votes/reviews matter. | 16:01 |
| cdent | our cat develops a limp every time we threaten to travel | 16:01 |
| johnsom | I have been trying to get the word out that +1 votes make a huge difference for the core team and the project. | 16:02 |
| *** annabelleB has quit IRC | 16:02 | |
| mugsie | johnsom: yeah. or people don;t get that *good* +1s matter | 16:02 |
| dhellmann | ttx : safe travels, and see you sat/sun | 16:02 |
| smcginnis | johnsom: I've been trying to get that message out for years to limited success. | 16:02 |
| smcginnis | Or good -1's too. | 16:02 |
| ttx | so looking forward to getting up at 4am tomorrow morning | 16:02 |
| johnsom | Yep | 16:03 |
| johnsom | I will totally not merge a patch with a -1 from anyone | 16:03 |
| smcginnis | ttx: That sounds fun. | 16:03 |
| cdent | From personal experience, I stopped making +1s because I was told that it seemed like I was being too generous[1] and I stopped making -1s because my comments were ignored. | 16:03 |
| johnsom | We are a team and it matters | 16:03 |
| dhellmann | yeah, I counted -1 as a review but not +1. I want to make the tool smarter to count a +1 following a -1 but haven' tyet | 16:03 |
| cdent | [1] I just wanted to see code that was "okay" merge. | 16:03 |
| mugsie | cdent: well, that is a bigger issue we have in the community | 16:04 |
| smcginnis | I love seeing patches with +1/-1's. | 16:04 |
| smcginnis | As long as they aren't from people that only seem to blindly apply +1. | 16:04 |
| cdent | also [1] and too many +1s that disagree with cores is considered statistically naive | 16:05 |
| jbryce | I think your presentation is really good, dhellmann. I'm also not going to be in the meeting Sunday unfortunately (ttx, lsell and mark will be), but I'm happy to catch up with any of you during the week on foundation-related topics that you might want to dive into | 16:05 |
| cdent | so ┻━┻ ︵ ¯\_(ツ)_/¯ ︵ ┻━┻ | 16:05 |
| *** annabelleB has joined #openstack-tc | 16:05 | |
| dhellmann | for those of you following the API cleanup discussion: http://lists.openstack.org/pipermail/openstack-dev/2018-May/130617.html | 16:05 |
| cdent | no jbryce at a summit board meeting. I'm glad I won't be there, because otherwise I'd feel all weird and wrong :( | 16:05 |
| dhellmann | jbryce : thanks for the feedback, and the offer. I'm sure we'll catch up at some point during the week. | 16:06 |
| jbryce | mugsie: I'm also interested in finding out about the lcoo, so if you do get something set up, please let me know when | 16:06 |
| mugsie | jbryce: will do | 16:06 |
| jbryce | cdent: I feel weird. This will only be the second one I've missed | 16:08 |
| cdent | I'm going to pretend like you are there. | 16:08 |
| * smcginnis pictures a cardboard cutout | 16:08 | |
| ttx | We do have that in stock | 16:09 |
| dhellmann | ttx: do we have time to get the artists to do a rendition like the super user personas? | 16:09 |
| smcginnis | Haha, nice. | 16:10 |
| ttx | https://photos.app.goo.gl/WnwpMysZVKVBSdyG3 | 16:11 |
| ttx | Just needs a bit of cutout work | 16:11 |
| smcginnis | Oh wow | 16:11 |
| dhellmann | oh, you can't unsee that | 16:11 |
| ttx | I know, right | 16:12 |
| smcginnis | Just the hair alone. | 16:12 |
| TheJulia | lol | 16:12 |
| ttx | The moose is so much on-topic, too | 16:13 |
| fungi | frightening | 16:13 |
| fungi | i had somehow successfully repressed prior memories of that | 16:13 |
| jbryce | I'll add one of my favorite ttx staff photos https://usercontent.irccloud-cdn.com/file/TpK9F9Cf/Framed.png | 16:35 |
| *** eeiden_ has joined #openstack-tc | 16:40 | |
| openstackgerrit | Zane Bitter proposed openstack/governance master: Note that the old incubation/graduation process is obsolete https://review.openstack.org/569164 | 16:41 |
| *** dklyle has quit IRC | 16:42 | |
| *** eeiden has quit IRC | 16:42 | |
| *** eeiden_ is now known as eeiden | 16:42 | |
| *** hongbin has joined #openstack-tc | 16:53 | |
| lbragstad | dhellmann: what does bus factor mean? | 16:58 |
| mugsie | lbragstad: if someone is no longer availible to work on something, how bad will it affect said thing | 17:00 |
| * jroll likes the less morbid "lottery factor" | 17:00 | |
| mugsie | jroll: yeah, thats better alright | 17:00 |
| lbragstad | oh - like getting hit by a bus, got it | 17:01 |
| *** dtantsur is now known as dtantsur|afk | 17:04 | |
| cdent | lbragstad: if you don't like that violent imagery, and you go back far enough to remember it, you can also think of it as not terminating your scsi bus | 17:06 |
| cdent | but really that's just callback to the hit by a bus, so not actually a replacement | 17:07 |
| lbragstad | lol - i was just reading the notes and couldn't resolve the mapping | 17:08 |
| mugsie | basically, I better look both ways crossing the street on the way home :P | 17:08 |
| cmurphy | the proposal bot on slide 15 also better be careful | 17:10 |
| *** kumarmn has joined #openstack-tc | 17:17 | |
| zaneb | cdent: stop making me feel old | 17:22 |
| *** kumarmn has quit IRC | 17:22 | |
| cdent | I'm here for you | 17:23 |
| cdent | dhellmann: did you see https://twitter.com/jessitron/status/997105157979758592 ? It's related to both the burnout stuff you mentioned as well as the constraints we put on ourselves because we learn to be our own guards | 17:35 |
| cdent | you have to read the handwriting in the pics for it to make any sense | 17:35 |
| dhellmann | lbragstad , jroll : yeah, it's maybe not the best phrase, but it's well known | 17:45 |
| jroll | dhellmann: indeed, I just like to spread the optimism of the lottery factor, I don't really have a problem with the bus factor phrase :) | 17:45 |
| dhellmann | or at least I thought it was -- maybe that's a bad assumption | 17:45 |
| dhellmann | I usually use "abducted by aliens" instead | 17:45 |
| jroll | heh, I like it | 17:46 |
| * cdent is ashamed | 17:48 | |
| dhellmann | cdent : I had not seen that tweet | 17:48 |
| cdent | I hadn't connected "lottery factor" with winning the lottery and packing up to a new life | 17:48 |
| cdent | I still saw death | 17:48 |
| dhellmann | as in The Lottery? | 17:48 |
| dhellmann | https://en.wikipedia.org/wiki/The_Lottery | 17:49 |
| cdent | effectively, but not directly | 17:49 |
| cdent | that was probably in the back of my mind though, if it really is "one of the most famous short stories in the history of American literature" | 17:49 |
| dhellmann | thought provoking ideas in that tweet | 17:52 |
| dhellmann | the first couple of points of https://doughellmann.com/blog/2017/08/24/stop-working-so-hard-scaling-open-source-community-practices/ relate, although they're less flashy than "panopticon" | 17:53 |
| cdent | yeah, the stop working so hard stuff aligns, as well with the do-less, generate-more ideas too | 17:56 |
| jroll | cdent: I also liked the following tweet, in a different way: https://twitter.com/jessitron/status/997106274948059136 | 17:56 |
| cdent | indeed. I think maybe I wanted to be at this confernce | 17:57 |
| jroll | ditto | 17:58 |
| *** ricolin has quit IRC | 17:59 | |
| *** kumarmn has joined #openstack-tc | 18:12 | |
| openstackgerrit | Merged openstack/governance master: Setup Release Goal Folder for Stein https://review.openstack.org/567700 | 18:14 |
| openstackgerrit | Merged openstack/governance master: fix stein goals build https://review.openstack.org/568947 | 18:14 |
| *** kumarmn has quit IRC | 18:14 | |
| *** diablo_rojo has joined #openstack-tc | 18:45 | |
| *** annabelleB has quit IRC | 18:50 | |
| *** dklyle has joined #openstack-tc | 18:56 | |
| *** annabelleB has joined #openstack-tc | 19:03 | |
| *** zaneb has quit IRC | 19:26 | |
| *** dtantsur|afk has quit IRC | 19:29 | |
| *** eeiden has quit IRC | 19:33 | |
| *** dtantsur has joined #openstack-tc | 19:36 | |
| *** dtantsur has quit IRC | 19:40 | |
| *** zaneb has joined #openstack-tc | 19:48 | |
| *** dtantsur has joined #openstack-tc | 19:51 | |
| *** annabelleB has quit IRC | 20:00 | |
| *** annabelleB has joined #openstack-tc | 20:07 | |
| fungi | cdent: zaneb: similar to the scsi analogy, not terminating 10base2 was equally problematic on the network side of things | 20:15 |
| * fungi had plenty of opportunity to make references to the terminator films both on the networking and storage sides of systems administration | 20:16 | |
| zaneb | 10base5 4eva | 20:17 |
| fungi | oh, fo sho thicknet was where it was at | 20:20 |
| * zaneb is not actually *that* old | 20:20 | |
| fungi | thinnet was just out at the edges | 20:21 |
| * fungi is a little embarrassed to have managed 10base5 to 10base2 bridges | 20:21 | |
| fungi | also all manner of assorted serial concentrators, multiplexers and terminals | 20:21 |
| zaneb | sometimes I regret never having had the chance to vampire-tap a thicknet cable | 20:22 |
| zaneb | but other times I just carry on with my life :) | 20:22 |
| fungi | i can at least say i've only ever vampire-tapped pots | 20:22 |
| fungi | (plain old telephone system, for the acronym-impaired) | 20:23 |
| *** annabelleB has quit IRC | 20:26 | |
| cdent | I enjoy playing with masses of coax t connectors | 20:27 |
| cdent | Well, enjoyed | 20:27 |
| cdent | I have no more | 20:27 |
| *** annabelleB has joined #openstack-tc | 20:29 | |
| * smcginnis has been around long enough to have used token ring | 20:35 | |
| * dhellmann remembers sneakernet fondly | 20:54 | |
| *** dtantsur has quit IRC | 21:02 | |
| *** dtantsur has joined #openstack-tc | 21:03 | |
| fungi | cdent: yeah, at one point i had a giant sculpture of random coax connectors (adapters, tees and terminators) on my desk. no idea what eventually happened to it | 21:06 |
| fungi | odds are i decided between one of my moves that i was unlikely to ever need anything besides a type n for 50Ω media | 21:07 |
| fungi | er, 75Ω | 21:09 |
| fungi | though ultimately both, i suppose | 21:11 |
| *** notmyname has quit IRC | 21:15 | |
| mtreinish | smcginnis: the ibm office I used to at had the whole building still wired for token ring, with the old data connectors (the big ones) | 21:17 |
| mtreinish | they just had adapters in each office to convert it to ethernet | 21:17 |
| *** annabelleB has quit IRC | 21:18 | |
| smcginnis | mtreinish: I was in the Rochester office not long ago, and I can state for fact that they *still* have things wired up for it. | 21:18 |
| *** notmyname has joined #openstack-tc | 21:18 | |
| mtreinish | I haven't been to the poughkeepsie office in several years, but I would assume they didn't change anything | 21:19 |
| smcginnis | Probably a safe assumption. ;) | 21:19 |
| *** annabelleB has joined #openstack-tc | 21:28 | |
| *** edmondsw has quit IRC | 21:46 | |
| *** annabelleB has quit IRC | 22:19 | |
| *** annabelleB has joined #openstack-tc | 22:22 | |
| *** mriedem is now known as mriedem_afk | 22:36 | |
| *** mriedem_afk has quit IRC | 22:38 | |
| *** hongbin has quit IRC | 22:40 | |
| *** cdent has quit IRC | 22:54 | |
| *** edmondsw has joined #openstack-tc | 23:03 | |
| *** edmondsw has quit IRC | 23:08 | |
| *** annabelleB has quit IRC | 23:29 | |
| *** edmondsw has joined #openstack-tc | 23:46 | |
| *** annabelleB has joined #openstack-tc | 23:48 | |
| fungi | pretty nuts | 23:53 |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!
