Tuesday, 2018-08-21

*** cdent has quit IRC00:02
*** mriedem has quit IRC01:26
*** ricolin has joined #openstack-tc01:33
*** Bhujay has joined #openstack-tc02:00
*** gcb_ has joined #openstack-tc02:07
*** diablo_rojo has quit IRC03:21
*** Bhujay has quit IRC03:24
*** Bhujay has joined #openstack-tc04:45
*** Bhujay has quit IRC04:51
*** Bhujay has joined #openstack-tc04:53
*** jaosorior has quit IRC05:45
*** jaosorior has joined #openstack-tc05:47
*** tosky has joined #openstack-tc07:29
* ttx tries to catch up with backlog07:37
*** jpich has joined #openstack-tc07:39
ttxwow, that was something.08:10
*** e0ne has joined #openstack-tc08:10
ttxMy 2c: first I think it's critical that the people working on the code have control over its destiny, that is how everything is organized in OpenStack. So if the placement group is not the same as the nova group (and it is different, since the trust  rings are different) , I think it should have its own team08:12
ttxThat said, placement also needs to ensure its first and foremost customer (nova) is satisfied, so it should take a base feature list and red lines from nova. That can happen by delaying the extraction (as melwitt suggested) or more under a contract that the TC would ensure is filled08:14
ttxOtherwise we risk extracting a thing that nobody uses / duplication of code etc08:14
*** cdent has joined #openstack-tc08:33
cdentttx, yeah, you missed a very interesting time08:37
cdentI think it is far better to have the stuff surfaced then lurking08:37
cdentttx that this concern [t 1r0A] is present is doubly concerning08:38
purplerbot<ttx> That said, placement also needs to ensure its first and foremost customer (nova) is satisfied, so it should take a base feature list and red lines from nova. That can happen by delaying the extraction (as melwitt suggested) or more under a contract that the TC would ensure is filled [2018-08-21 08:14:28.108695] [n 1r0A]08:38
cdentgiven that if those things have been lurking all this time, yet people have still continued to strive to make placement (and nova) good, the commitment must be especially strong08:39
ttxIt's part of the TC role as the overarching governance body to ensure coordination and solve disputes between groups of people working on pieces (a.k.a. project teams)08:42
ttxSo I think I'm leaning toward supporting a team split (to solve the elephant) and then work at TC level to watch/ensure the common goals are met08:43
ttxwhich I have no doubt they will really, but if trust is not there from the other side, the TC can help08:43
ttxStill need to catch up on the ML thread08:45
ttxtc-members assemble for office hour09:01
* cdent waves09:01
cmurphyhiya09:01
cdentthere's a fair bit of good followup from matt in the email thread, he continues his trend of surfacing the hidding09:01
ttxwell I for one am thankful for the surfacing. it's hard to discuss the topic if everybody is just hiding an elephant09:02
cdentyes09:02
ttxalthough it's like a fracture reduction, it hurts when it happens09:03
ttxI'm reading the followup09:03
ttxlooks like splitting the repo is a good first step while we work to solve the trust issue there09:08
lbragstadis it just me or does zun have nearly the same requirements of placement as nova does (as opposed to services that are just reporting information to placement)?09:10
* cmurphy hands lbragstad some coffee09:11
lbragstadcheers :)09:11
cdentlbragstad: that was my read too09:20
cdentwith the complexity that they ideally want to share resources on the same host09:20
cdentwhich is placement can do09:20
cdentsorry that was two thought: which is what placement is designed for / which placement can do09:21
evrardjpThat's what makes sense to me (with no knowledge of the code though)09:22
lbragstadhas the TC ever had to solve something like #3 before? http://lists.openstack.org/pipermail/openstack-dev/2018-August/133559.html09:30
cdentthat's a good question. that it is considered a problem is weird (see the message I just sent on the thread: we're supposed to be disagreeing some of the time)09:31
ttxthe closest would be the magnum/Zun conflict09:32
ttx(in recent history)09:32
cdentttx are you talking generally or specific abourt issue #3?09:33
ttx#309:33
ttxalthough in that case the fear of having the TC step in was sufficient for them to work out their issues and decide to split09:34
ttxTo switch topics, we should confirm post-lunch talks this week, so please have a look at proposed strawman at bottom of https://etherpad.openstack.org/p/PTG4-postlunch09:35
ttxTrying to fit everythign on days people can do them09:35
* cmurphy looks forward to the email etiquette topic09:36
ttxcrosspost bad. toppost bad09:37
cmurphyhtmlonly bad09:37
ttxhave to doublecheck if diablo_rojo actually knows rockyg threw her under the bus09:37
cdentttx that schedule seems reasonable to me09:37
ttx(for Storyboard talk)09:38
ttxleaves some room on Friday for last-minute $stuff09:38
ttxalso makes sure everyone sees the Tech vision statement09:38
ttx(by placing it on Wed)09:39
cdenti rather think dhellmann's idea of a full day battle royale could be fun09:39
lbragstadthe third bit seems specific to not only building trust, but making sure things evolve and progress is made (the biggest fear so far), is there a way to "timebox" design things? has that ever been done before outside of something like the specs process?09:39
ttxcdent: consensus through shrinking maps ?09:40
*** mriedem has joined #openstack-tc09:42
* lbragstad is specifically curious because keystone has gone through similar drawn-out design discussions that last _releases_09:42
evrardjpcdent: taking the popcorn09:42
ttxtc-members: please check strawman agenda at end of https://etherpad.openstack.org/p/PTG4-postlunch -- there is room for everything, we just need to check that we WANT to have each of those. We could keep empty slots09:42
* ttx should check if there was anything planned for Dublin that we could not do that are still relevant09:43
cdentlbragstad: yeah, placement stuff never goes more than a few days, which is part of why I find the concern being expressed difficult to understand. placement velocity is insane09:43
cmurphythose all look like good topics to me09:43
ttxok, the one we could not do was around release management, and smcginnis re-proposed it09:44
cdentlbragstad: it does sometimes look like logjams but it breaks up in an eventually good way. I think the issue is more what I point out at the end of my message: the people doing the arguing and resisting are not the "right" people and that throws some people off09:44
*** dtantsur|afk is now known as dtantsur09:59
dtantsurmorning folks! Do you happen to know if people who attended either Dublin or Denver PTG will get free passes to the summit (like it used to be previously)? It's not obvious so far.10:26
*** Bhujay has quit IRC10:26
dtantsurttx: if you know ^^10:26
*** yolanda has joined #openstack-tc10:27
ttxyes, Denver attendees will get Berlin summit passes10:29
ttxYou have to attend the event though, registering is not enough10:29
dtantsurttx: what about Dublin attendees?11:01
ttxhmm, I'll have to ask.. fungi probably knows11:02
dtantsurpreviously it was two last PTGs. I highly count it since I cannot go to Denver..11:06
EmilienMwhat a pleasant backlog to read (sarcastic)11:12
EmilienM(hi)11:13
openstackgerritlei zhang proposed openstack/constellations master: Clean up .gitignore references to personal tools  https://review.openstack.org/59225911:13
*** Bhujay has joined #openstack-tc11:13
*** e0ne has quit IRC11:36
evrardjptalking about Berlin, is there room for Project Updates/Onboardings? I didn't see a call for that (yet?)12:02
*** rosmaita has joined #openstack-tc12:07
mnaserevrardjp: i go tthe update for that, there will be one :p12:39
mnasers/update/email/12:39
mnaseri mentioned we're going to be having both, you're welcome to host/run it if that's something you want to do :)12:40
evrardjplet's discuss that on the OSA channel :)12:40
*** efried is now known as efried_goatin13:03
*** dkrol_ has joined #openstack-tc13:05
fungidtantsur: ttx: yes, see the e-mail i sent with the code contributor discount code: "if you attended the February Dublin PTG or plan to attend the September Denver PTG, then you will receive a complimentary Full Access Pass. Those codes will be sent out late September."13:34
fungithat was the text provided to me by the summit event organizers, so i assume it's correct13:34
dtantsurfungi: oh, so you're saying I was supposed to read that email to the end? :D13:35
fungiheh13:35
dtantsurthank you fungi, my bad13:35
funginp13:35
fungiif i didn't have to proofread those, i probably wouldn't have read the whole thing either ;)13:35
fungievrardjp: yeah, the project update and onboarding sessions were determined by diablo_rojo sending a form link to ptls13:37
*** efried_goatin is now known as efried13:37
fungiper the lunch session straw-man schedule, the technical vision statement presentation is a "smaller talk?"13:37
fungigiven the amount of discussion so far on the ml and in the review, i anticipated we'd want time for q&a13:38
*** e0ne has joined #openstack-tc13:40
smcginnisI can give up the release management time so we can have more time for the technical vision.13:44
smcginnisIt's useful info I think, but I can just write a blog post and/or record a video or something.13:44
cdenti'm not sure 10 minutes over 5 minutes is sufficient. Either 5 minutes or an >hour13:46
cdentin between will be weird, maybe?13:46
smcginnisIf we can't devote more time, maybe have a quick "this is what we're now discussing" overview with a plan to have a full Forum session in Berlin?13:48
evrardjpsmcginnis: yeah or we can complete the docs together :)13:48
smcginnisevrardjp: That could be useful! ;)13:48
evrardjpsmcginnis: I mean we can call for volunteers, and start working on it13:48
*** annabelleB has joined #openstack-tc14:23
*** lbragstad has quit IRC14:26
* cdent skips tc report this week. I'm writ out14:30
zanebcdent: I tend to agree. it can be "here's a thing that's happening, please review it" or "here's a 15 minute spiel about what it all means, now let's have 30 minutes for questions", but I don't know what you could usefully do in between14:31
fungii feel the same14:32
fungithe election officials had also expressed an interest in a tc election panel in a lunch session14:33
*** annabelleB has quit IRC14:33
fungii know there were some positives and negatives laid out on the ml14:34
dhellmannwould it make sense to introduce the technical vision and then mention we will be discussing it in more depth on friday afternoon?14:34
fungii like that14:34
zaneb+114:35
ttxyeah14:36
ttxFriday afternoon sounds like a great time14:36
ttxthen the lightning talk can be a teaser14:36
dhellmannit's already on our agenda in the tc room for friday afternoon, so we just need the announcement to come earlier than that14:37
dhellmannkeeping the lightning talks on wednesday gives folks to ask some questions before friday14:37
*** annabelleB has joined #openstack-tc14:41
ttxIt will be more "short talks" than lightning talks on Wednesday, since one of them will have 10min14:43
dhellmannok14:43
ttxbut yes, earlier will ensure that people see that content14:44
smcginnisI'm assuming we won't have specific time slots but just have it when one talk ends the next start?14:44
ttxsince Friday is a bit of a hit and miss. I tend to go to the Taco truck on Fridays14:44
ttxsmcginnis: I'd still recommend having a timeline, but not the "rip laptop off your hands after 5:00" style14:45
smcginnis;)14:46
* cmurphy looking forward to tacos14:46
zanebdo we still have time in the lunch schedule to fit in a session where the Nova team does trust falls in front of the whole PTG?14:49
ttxzaneb: too soon?14:50
zanebit was mriedem's idea14:50
ttxalthough the atrium setup makes it tempting14:51
mriedemit's the only way to heal the wounds14:53
smcginnisCan I sponsor this portion of the PTG?14:54
zanebquestion: does the distrust that apparently exists between the nova core team and the placement team also extend to distrust within the placement team, given that there is some overlap?14:55
dimso/14:56
*** hongbin has joined #openstack-tc14:56
zanebyesterday's IRC discussion seemed to suggest that it did, but the mailing list thread seems to be assuming that it doesn't14:57
cdentzaneb: I somewhat reject that it is a question of trust.14:57
dhellmannis it really an issue between the teams? or between individuals?14:57
cdentI think that's a factor, clearly14:57
cdentbut it is more about power14:57
cdentwhich obviously is informed by trust14:57
zanebright, people generally don't mind giving power to people they trust14:58
cdentsure, but the apparent mechanism for acquiring trust is "agree with me"14:59
cdent(I'm trying to be as honest and open here as mriedem was last night, so if this is uncomfortable, my apologies, but we gotta work through this)15:00
ttxI'd say "control" rather than "power", but yes, it's interconnected15:02
*** lbragstad has joined #openstack-tc15:02
cdentyes, that is probably a better choice of words15:03
zanebI feel like we need to build trust, and absolutely not by saying 'stop thinking for yourselves'15:03
ttxzaneb: i think the key is to document common culture/expectations so that you don't have to control and review everything to prevent things from going "wrong"15:04
ttx(i.e. define "wrong")15:05
zanebcdent: but I don't know how to do that given that the stuff you listed in your email that you're doing are the right ways to go about building trust, and they don't seem to be getting reciprocated :(15:05
cdentthanks for noticing :)15:05
*** Bhujay has quit IRC15:27
*** dklyle has quit IRC15:41
*** dklyle has joined #openstack-tc15:44
*** e0ne has quit IRC15:44
openstackgerritHongbin Lu proposed openstack/governance master: Add openstack/os-ken to neutron  https://review.openstack.org/58835815:45
*** rosmaita has quit IRC15:57
evrardjpcdent: (non constructive comment from a Belgian) I'd say if you want to build more trust, you can talk around beers.15:59
smcginnisMmm, a big foamy pint of trust.16:04
*** ricolin has quit IRC16:06
evrardjpsmcginnis: :)16:06
efriedsmcguinness <== fun friday nick16:13
ttxtc-members: diablo_rojo confirmed she can give a StoryBoard talk, but only has 15-min worth of it, so i combined it with apiers's "Simplifying backports with git-deps and git-explode" which is also a 15-min thing16:15
ttxThat way the three things on Wednesday have a little more wiggle room16:15
ttxplease let me know (on the etherpad) if that sounds good to you and I'll socialize16:16
ttxhttps://etherpad.openstack.org/p/PTG4-postlunch16:16
dhellmannwhen we had this idea originally we talked about having project teams talk about some of the things they were planning (or planning on planning). it's a shame we can't seem to make that happen. maybe next time.16:17
ttxdhellmann: I still hope some will use lightning talks for that16:17
ttxon-site16:17
dhellmannmaybe we can encourage that in the monday session16:17
*** yolanda has left #openstack-tc16:17
ttxgood idea16:18
ttxi'll note it down16:18
dhellmanntc-members: based on the votes cast, I have approved the searchlight PTL nomination and abandoned the proposal to drop it from governance16:18
cdentanother thing we had talked about, early on, for the monday session, was setting themes for the week "these are the ideas that may help you shape what matters"16:19
ttxcdent: yes, that's what I mean by "set the stage for Stein"16:19
ttxmnaser signed up to do that part16:19
ttxso maybe you can influence the content with him16:20
* mnaser is in OSA meeting right now but can chat about that later16:20
openstackgerritMerged openstack/governance master: Adding Trinh Nguyen candidacy for Searchlight PTL  https://review.openstack.org/59060116:20
ttxcorrected to "set the stage for the week" which is more accurate16:20
cdentI actually don't have strong opinions about what the stage shoud be, just that it needs to be set16:21
* dhellmann tosses a roll of gaffer's tape in his suitcase16:21
fungiooh, yeah i like the idea of unconferencing the lightning talks and giving teams the week to prepare for any they want to present. could be a good way to showcase what they worked on during the event even16:22
ttxcdent, zaneb: I've removed myself from the Tech vision talk since 3 people for a 8-min thing sounds overkill16:22
cdentayeaye16:22
fungii'm happy to heckle from the peanut gallery ;)16:23
openstackgerritMerged openstack/governance master: mark Trinh Nguyen's appointment as an appointment  https://review.openstack.org/59147816:23
openstackgerritMerged openstack/governance master: Add qinling-dashboard under Qinling project  https://review.openstack.org/59155916:24
openstackgerritMerged openstack/governance master: Update Puppet PTL IRC nick  https://review.openstack.org/59363716:24
openstackgerritMerged openstack/governance master: import zuul job settings from project-config  https://review.openstack.org/59370116:25
*** annabelleB has quit IRC16:33
*** jpich has quit IRC16:34
*** dtantsur is now known as dtantsur|afk16:38
*** annabelleB has joined #openstack-tc16:43
*** tosky has quit IRC17:24
zanebttx: so, uh, what _is_ our position on dependencies between OpenStack components?17:38
*** e0ne has joined #openstack-tc17:42
*** portdirect has quit IRC17:54
*** diablo_rojo has joined #openstack-tc17:56
fungiwhat do you mean by "dependencies"?18:28
fungiand "components" for that matter?18:29
fungiare you talking about components like keystone or components like oslo.db? and are you talking about literal import dependencies or are you talking about functional dependencies expressed through communication over some protocol?18:30
smcginnis"The single biggest problem in communication is the illusion that it has taken place.18:30
fungi"18:30
*** ttx has quit IRC18:37
*** bodgix has quit IRC18:37
*** fungi has quit IRC18:37
*** TheJulia has quit IRC18:42
*** tobberydberg has quit IRC18:42
*** hrybacki has quit IRC18:42
*** mnaser has quit IRC18:42
zanebsmcginnis: ironically a netsplit happened right after you typed that18:44
*** TheJulia has joined #openstack-tc18:44
jrolllet's collapse everything back into a monolithic service so that won't happen to openstack deployments18:45
smcginniszaneb: That is ironic.18:48
*** fungi has joined #openstack-tc18:48
*** rosmaita has joined #openstack-tc18:53
*** e0ne has quit IRC18:57
TheJuliajroll: I'll just shake my head19:01
jroll:)19:01
*** mnaser has joined #openstack-tc19:10
*** bodgix_ has joined #openstack-tc19:12
*** cdent has quit IRC19:16
*** bodgix_ has quit IRC19:17
*** bodgix has joined #openstack-tc19:18
*** bodgix has left #openstack-tc19:18
*** bodgix has joined #openstack-tc19:19
*** melwitt has quit IRC19:33
*** melwitt has joined #openstack-tc19:34
*** evrardjp has quit IRC19:39
*** scas has quit IRC19:39
*** mriedem is now known as mriedem_afk19:39
*** dirk has quit IRC19:46
*** samueldmq has quit IRC19:46
*** dirk has joined #openstack-tc19:47
*** samueldmq has joined #openstack-tc19:49
cmurphyzaneb: we're paired up on both rpm-packaging and mistral health checks, i went ahead and added notes for rpm-packaging and i'm kind of hoping you'll do mistral :)19:53
zanebcmurphy: ack, thank you and I will take Mistral19:54
*** jaosorior_ has joined #openstack-tc20:14
*** jaosorior has quit IRC20:17
*** edmondsw_ has joined #openstack-tc20:18
*** edmondsw has quit IRC20:19
*** edmondsw_ is now known as edmondsw20:19
*** dkrol has quit IRC20:34
*** scas has joined #openstack-tc20:38
*** annabelleB has quit IRC20:41
*** rosmaita has quit IRC20:43
*** annabelleB has joined #openstack-tc20:55
*** harlowja has joined #openstack-tc21:02
*** ildikov has joined #openstack-tc21:02
*** mriedem_afk is now known as mriedem21:04
efriedIf anyone has this at their fingertips and can save me searching...21:27
efriedIs there a doc somewhere that describes the process for appointing cores?21:28
mriedemthere is....something21:39
mriedemhttps://docs.openstack.org/nova/latest/contributor/how-to-get-involved.html#how-do-i-become-nova-core21:40
mriedempretty old though21:40
mriedemlike, from when john/mikal were ptl i think21:41
efriedFrom an OpenStack perspective, is the nomination/appointment process governed entirely per-project? Is there no standard?21:45
smcginnisefried: I think so.21:45
smcginnisI don't think I've seen any document declaring or even recommending how each team needs to handle it.21:46
efriedk, thx mriedem and smcginnis21:46
smcginnisSome just make a decision in their weekly meeting. Some have public posting on intent with a wait period for feedback. I think it's really what works best for the team.21:46
mnaserhttps://redislabs.com/community/commons-clause/21:53
mnaserInteresting. I wonder if we use them inside OpenStack anywhere.21:54
dimsefried : correct21:54
dimsmnaser : some hits - http://codesearch.openstack.org/?q=%5Eredis&i=nope&files=%5E.*%5C.txt&repos=21:55
pabelangermnaser: yah, I think there are some projects with a redis dependency21:56
pabelangertooz comes to mind21:57
mnaserI wonder where that put us given that it’s not really 100% open anymore21:59
pabelangerI would guess the same place as influxdb, they are open core last I checked22:04
*** tosky has joined #openstack-tc22:07
*** annabelleB has quit IRC22:21
*** mriedem is now known as mriedem_away22:22
*** annabelleB has joined #openstack-tc22:27
*** diablo_rojo has quit IRC22:44
*** hongbin has quit IRC22:44
*** diablo_rojo has joined #openstack-tc22:48
*** tosky has quit IRC23:07
fungiyeah, if we require that dependencies be packaged on major distros and those distros drop their packages of those dependencies (likely in this case) then i think our hand is forced anyway23:27
*** annabelleB has quit IRC23:51
*** annabelleB has joined #openstack-tc23:52

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