Tuesday, 2018-07-31

*** diablo_rojo has joined #openstack-tc00:13
*** edmondsw has joined #openstack-tc00:42
*** edmondsw has quit IRC00:48
*** dklyle has quit IRC01:38
*** diablo_rojo has quit IRC01:51
*** gcb has joined #openstack-tc02:33
*** rosmaita has quit IRC03:40
*** dtroyer has quit IRC03:40
*** dtroyer has joined #openstack-tc03:52
*** ricolin has joined #openstack-tc03:57
*** gcb has quit IRC04:22
*** mriedem has quit IRC04:49
*** eumel8 has quit IRC04:51
*** eumel8 has joined #openstack-tc04:51
*** gcb has joined #openstack-tc05:36
*** e0ne has joined #openstack-tc05:59
*** zaneb has quit IRC06:34
*** zaneb has joined #openstack-tc06:34
*** dklyle has joined #openstack-tc06:39
*** e0ne has quit IRC06:51
*** dklyle has quit IRC07:06
*** cdent has joined #openstack-tc07:55
*** jpich has joined #openstack-tc08:10
*** tosky has joined #openstack-tc08:12
openstackgerritThierry Carrez proposed openstack/governance master: Add Tripleo Ansible repo  https://review.openstack.org/58341608:42
cdenttc-members: office hours09:00
ttxhi!09:00
cdentmornin'!09:01
ttxThanks for starting the PTG etherpad. Added one thing to it09:01
ttxhttps://etherpad.openstack.org/p/tc-stein-ptg for those following at home09:01
* ttx is still catching up from PTO week09:04
ttxThat cycle sure went fast09:04
cdentyeah, I'm still trying to get tasks done from Sydney09:06
ttxThat's the feel I (badly) tried to address by making the cycle longer09:06
ttxWe review progress faster than we do things :)09:07
cdentI don't think changing cycle length has any impact on the fact that some people are consistently trying to do too many things09:07
ttxright. Which is why it was "badly"09:07
cdentI have something like 5 upstream roles that could easily be enough to fill my time09:07
ttxAlthough having separate PTGs led to review progress every 3 months instead of every 6 months09:08
ttxwhich was probably a great idea in 2015, not so much in 201809:08
*** ricolin has quit IRC09:09
* cdent likes the 3 month checkins09:09
ttxsome of it is less resources, som eof it is just components maturing and things taking longer to change09:09
ttxChecking in every 3 months on progress is good, but 3-month in-person checkin was a bit of a luxury09:10
cdentDoes "openstack as base layer for k8s" fit into the constellation concept?09:11
ttxwhich less and less contributors can afford09:11
ttxhmm, I guess so09:11
ttx"What are the set of components you need to deploy to achieve that"09:12
ttxa minimal openstack install letting you take advantage of the openstack K8s cloud provider09:12
cdentyeah, that's what I was thinking. It has marketing cachet. More work for dims!! :)09:13
*** e0ne has joined #openstack-tc09:13
* ttx needs to book Denver travel today09:13
* cdent wonders if we need to rejigger the office hours hours09:46
*** jaosorior has quit IRC09:53
*** jaosorior has joined #openstack-tc09:53
ttxThe "European" hour is usually quite light, and if you add the Europeans propension to take PTO in July/August10:08
* cdent needs to learn to take more PTO10:08
ttxit's harder to reach critical mass10:09
ttxcdent: how much PTO do you have ?10:09
ttxttx has ~7weeks but has a hard time taking more than 410:10
cdentthere's some confusion on that because my official employment is UK based thus I have some standard approx 26 days of PTO plus holidays etc. But my group and mgt team is more US-based where it is much more ad hoc10:10
cdentttx your message just now about the PTG being with the next summit is the closest to official word I've seen on that. I think that change has been way underdiscussed and socialized.10:11
cdent(Not saying I disagree with it, just that the motions of the foundation in this case we're not very scrutable)10:12
ttxcdent: it was posted in that email I linked to, but i agree that it was not that widely or aggressively communicated10:12
ttxmostly because we were still working out the details of how to make it work10:13
cdentIf you were an average daily contributor a message with subject of " Vancouver Summit Joint Leadership MeetingUpdate" on the foundation list would be fairly invisible10:13
ttxI think we'll make a post-lunch presentation about it at the PTG itself10:14
ttxthat sounds like the best way to fully reach the targeted audience10:14
cdentI would think we need that, posts to os-dev, superuser article, at the bare minimum10:15
ttxagreed. Timing-wise, do you think doing that in September will work ? Doing it earlier might confuse people preparing for the September PTG10:17
ttxalso August is not the best month for wide comms10:17
cdentI'm not sure. As we've said many times: overcommunicating and being overly verbose is way better than under communicating10:21
cdentI think there are plenty of people who like to have good visibility over the next year as soon as possible10:22
ttxYeah, that's fair. I'll work on an article/blogpost that summarizes the rationale, survey results etc.10:24
cdentthat sounds good10:25
ttx(and clearly say that it's not a perfect solution at all)10:25
*** diablo_rojo has joined #openstack-tc12:05
*** rosmaita has joined #openstack-tc12:06
*** edmondsw has joined #openstack-tc12:24
scasre: tc report - 2015 had a conversation about letting chef 'gracefully die'. in 2018, the survey analytics suggest that some giants would have to undergo a cosmic shift as a consequence of losing an upstream path13:00
scasarguably, if i weren't mostly focused on the one thing, that would not preclude me from branching out more as an individual13:02
scason the other hand, refocusing is a signal to undergo that cosmic shift13:02
scassome are getting to that point on the enterprise time scale, finding natural break points to strategically refactor some tech debt13:03
*** mriedem has joined #openstack-tc13:05
scasas an operator for one of the aformentioned giants, that's a footgun and facemine waiting to happen. it's nothing short of mass retraining13:06
dimscdent : on top of the starter constellation i want to throw ironic (and other things needed) to be able to deploy bare metal with https://github.com/kubernetes-sigs/cluster-api too13:07
scaslosing out on valuable summit time hasn't helped my case any13:07
cdentdims: I was thinking in terms of a sort of "all your own stuff kubernetes and openstack starter kit" for what amounts to "hobbyists" as that could be a good source of contributors if we make it clear that the barriers to entry aren't too high13:08
dimsyes starter constellation can definitely be used for minimal k8s deployments (without Load balancers which need neutron-lbaasv2 or octavia)13:08
dimscdent : right13:09
scasone of my goals is to have chef deploy on top of kubernetes by default, be it through kitchen or zuul/ansible playbooks. my 'starter kit' is a fairly easy barrier to entry, although i won't be able to really showcase that until a) the gerrit downtime and 2) i fix my docs that somehow bitrotted13:09
dimscdent : i am yet to poke at what we have for all-in-one docker based deployments13:10
scaschef has a mostly working kitchen-dokken scenario, which is built on docker using the docker-api rubygem for glue. the standard scenario is all-in-one13:10
dims++ scas13:10
scasmostly working because i'm but one person13:10
dimsunderstood :)13:11
scasit mostly works even in integration, which smells promising if i can get the job to complete13:11
scasi've focused on aio as a result of my own personal pains, and it was something i could manage myself without having to actively recruit just to implement my grand delusions13:12
dimscdent : i like it!13:12
scasmy notion is that my downstreams have teams to implement the more complex architectures, so if i provide the basic primitives to get them there, they tend to go the rest of the way on their own13:14
scasin my operator life, this is true. i tend to rely on the work i do in the open to better dogfood things, without falling into that cosmic shift button13:17
dhellmannscas : given the number of times I hear about custom deployment tools, focusing on building primitives seems like a very good approach13:21
mugsieyeah, you either allow for complete customisation, or build primitives that people glue together13:22
scasdhellmann: it works well enough for the clouds i operate to scale to 400k cores and above (presented at sydney), which i suppose is some form of validation13:23
mugsiee.g. If I hear "we want to do X but $TOOL will overwrite it, and we can't stop it" again, I think I may cry13:24
scasmy cookbooks are peppered with guards and conditionals to allow that flexibility. the complexity, of course, increases exponentially, but it works13:25
scasworks > lamenting about how some tool can't change -- at least to me13:25
mugsiescas: a lot of it assumes a uniform deployment layout - a lot of people will have some nodes set to 1:1, 3:1 and some for pinning. Some have SR-IOV on designated nodes, and others have wierd OS setups to support weirder vendor software13:26
scasmugsie: of course. like fingerprints, each cloud is differently tailored for its needs13:26
scashow to avoid 100% individual criticality is part of the goal, with everything laid out like that13:27
scasthe longer i solo it, the more it becomes a self-fulfilling prophecy13:27
dhellmanntonyb, persia, diablo_rojo : can you give us an idea of how the elections would be affected by approving this extra-atcs patch late? https://review.openstack.org/#/c/586751/113:28
persiadhellmann: Are the results intended to change voter eligibility, or just other extra-ATC permissions?13:28
*** zaneb has quit IRC13:30
scasthings like superuser may make for some clicks, if i'm able to be patient enough to get docs working and published. i'm largely awaiting the upcoming gerrit downtime before i can really tear into it13:30
*** zaneb has joined #openstack-tc13:30
diablo_rojodhellmann, I whouldn't think there would be an issue as we havent run the rolls to generate the electorate yet.13:30
dhellmannpersia, diablo_rojo : yeah, the effect on the voter roles was what I was looking for. There are some additions to the list, so I think those would be new folks for the i18n PTL election roll13:31
persiaAlso, I18n appears to be an uncontested PTL election, although there is time to go before nominations close.  In the event it remains uncontested, those people wouldn't need to vote anyway.13:31
openstackgerritDoug Hellmann proposed openstack/governance master: typo fix in voting description for chair-proposed typo fixes  https://review.openstack.org/58743213:33
scasgiven the discussions on the ml and here about entry points, i'm inclined to write something up going through the basic workflow to 'get to openstack'13:33
dhellmannmnaser : that one's for you ^^13:33
scasdocs are docs, but they're dry13:34
dhellmannpersia : I cannot imagine having a second candidate step up for that team, so I think we're safe. I just want to make sure we don't approve something after the deadline we set if it's going to cause a problem :-)13:34
diablo_rojoIf you can wait till nominations close to merge it we can be safe, but I also don't think there would be an issue dhellmann13:35
diablo_rojolol13:35
persiadhellmann: The potential problem is that at this point officials may have generated rolls (which takes a few hours), and if there was an I18N election, would need to regenerate.13:35
diablo_rojoIf we want these people included should there be an i18n election, merge now. If we don't want them included, wait till nominations close.13:36
scasbtw, i'm doing my musing here 'out loud' because it would otherwise be me talking to myself in my head, where everything is a great idea13:36
persiaYes, and try to avoid merging late if an election is expected.  Those election officials not participating in this discussion have generated rolls already.  In this case, we expect it not to matter.13:37
diablo_rojoWell, nominations close and rolls are genreate13:37
diablo_rojo*generated13:37
dhellmannI think, since the election for that team is a formality, it's safe to leave the patch open until the normal deadline for approving it13:41
scasthe good news, to stop being so negative, is that i can get with the program on the releases repo going forward. i completed the general availability of the queens release of the cookbooks recently, which gives me a natural quiet time while i stabilize rocky once the packages are available13:44
diablo_rojodhellmann, works for me :)13:45
scasfor one person, four or so months to stabilize a release part-time isn't the worst. at least it wasn't like newton13:47
persiadhellmann: In practice, we don't actually run elections where they are uncontested: I'm forgetting the nomenclature at the moment but there is another word for how uncontested PTLs end up with their positions.13:47
dhellmannpersia : true13:48
scasthe sarcastic part of me wants to joke about the word being 'gullible'13:50
smcginnis:)13:50
jrollstockholm syndrome? :P13:51
ttxacclaimed13:51
scasjroll: :)13:51
persiattx: Thank you :)13:51
scasit's meant in jest, really13:51
ttxwe would say "plebiscited" in French but apparently that did not transfer to English13:52
ttx(although plebiscite did)13:52
scasin america, that would be interpreted as "you have a what?!?"13:52
scasin all seriousness, 'something' needs to happen with chef to either breathe new life into it, or signal that cosmic shift amongst the giants. as a dogfooder, i have somewhat of a vested interest in its success13:55
scasis superuser enough, at this stage?13:56
smcginnisIt's at least one way to get some publicity for it I think.13:57
scasvancouver was intended to have sessions talking about it. according to the emails back then, some people had added the sessions to their schedule before i had to change plans13:59
scaswithout someone else to present, it's down to me figuring out how to get to summits and conferences, while juggling everything else14:00
smcginnisscas: So you're saying it's all your fault?14:00
smcginnis:)14:00
scasit's my fault that mitaka didn't see the end of it :)14:01
scasat the end of the mitaka cycle, ibm and others pulled back. my operator ways kept feeding code into the machine long after it became apparent that it was just me14:03
scasas i've mentioned before, in the past, relocating chef openstack came up when there were more cores. the inertia of relocating it somewhere else can be even more deterimental than doubling down, so to speak. github and gitlab are not ideal places i want to hinge a deploy project like that, because they're funded ventures that can and will change course if the business agrees it should14:13
scasif the self-reporting is to be believed, the users that do exist by default agree with the current positioning14:16
*** hongbin_ has joined #openstack-tc14:18
ttxtc-members I put out the call for post-lunch presentations at the Denver PTG: https://etherpad.openstack.org/p/PTG4-postlunch14:29
scasit seems like it might be worthwhile for me to do whatever i can do make the denver ptg, if only to have conversations like this in-person for a change14:30
zanebno love for Zeppo?14:31
scasprovided there are no surprises in the next few months, i'm targeting being at the berlin summit. ptg never was too valuable for the chef cores internally due to the distributed nature of the team. irc worked until it didn't14:33
ttxzaneb: hard choices had to be made14:34
scaswe had more engagement when it was in stackforge, with bleary-eyed first-thing-monday google hangouts. irc made it harder to convey intent, including being overly transparent14:35
*** AlanClark has joined #openstack-tc14:46
mnaserttx: i'd be up to help out with some of them "welcome to the ptg" stuff if you need filling in for that question mark14:49
ttxmnaser: perfect, please add your name to it :_14:50
mnaserttx: cool, let me know when you want to sit and figure out what we have to do14:50
*** diablo_rojo has quit IRC15:03
*** ricolin has joined #openstack-tc15:13
*** e0ne has quit IRC15:23
*** AlanClark has quit IRC15:27
*** zaneb has quit IRC15:34
*** zaneb has joined #openstack-tc15:35
*** dklyle has joined #openstack-tc15:56
openstackgerritMerged openstack/governance master: typo fix in voting description for chair-proposed typo fixes  https://review.openstack.org/58743216:13
*** diablo_rojo has joined #openstack-tc16:33
*** jpich has quit IRC16:35
*** dtruong2 has joined #openstack-tc16:36
*** dklyle has quit IRC16:48
*** dklyle has joined #openstack-tc17:07
mnaserhttps://governance.openstack.org/election/17:11
mnaserare we failing to publish?17:11
mnaseroh wait17:12
mnaserpost queue is probably long17:12
dhellmannmnaser : I see 2 election patches in the post queue with ages over 3 hrs17:39
mnaserjust wanted to have a general figure of what projects dont have anyone elected (if any) yet17:39
dhellmannyeah, I was wondering that myself17:39
smcginnis6 hours left, so I'd like to see a little better picture too.17:40
smcginnisLooks like a bunch waiting to be approved yet.17:40
dhellmannsmcginnis , mnaser : render-statistics gives me: http://paste.openstack.org/show/72696317:45
dhellmannwhich of course doesn't include those waiting17:45
mnaserthe ones i identify with nothing proposed are: security (but that's unique), trove, RefStack, Loci, Winstackers, Packaging_Rpm, Freezer, Searchlight and Dragonflow17:47
mugsieTrove has a candiate17:47
mnaserhttps://review.openstack.org/#/q/project:openstack/election i didnt see one there :X17:48
mnaserhttps://review.openstack.org/#/c/587333/17:48
mnaseroh17:48
mugsieBut I am not sure if it will pass https://review.openstack.org/#/c/587333/17:48
mugsieI think that may need to be a post election appointment by us17:48
dhellmannhttp://paste.openstack.org/show/726966 is easier to read17:49
smcginnisSecurity should be dropped from the election once we approve the patch since it's now a sig.17:49
smcginnishogepodge: Were you planning on being RefStack PTL again?17:50
smcginnisI wonder if that makes more sense as a sig too though.17:50
mugsieYeah, or just rolled into InterOp17:51
smcginnisHmm, only one project needing an election so far.17:52
*** e0ne has joined #openstack-tc17:52
mugsieSenlin is going to need one as well I think17:54
dhellmannnothing for dragonflow?17:54
dhellmannor freezer?17:54
mugsienope17:54
dhellmannor Loci17:54
smcginnisNot too surprised on df or freezer, but a little surprised with Loci.17:55
mugsieor searchlight, but that might just be Steve doesn't realise that it is that time of year17:55
dhellmannI guess we need to start thinking about what we might do with some of these17:55
scasi keep putting my name on the hat so you all can prolong those thoughts a bit more, at least in my realm18:06
smcginnisSo scas for all projects that fail to get a PTL. ;)18:07
scasi've not perfected cloning yet18:07
TheJuliaDo we have transporters yet? Maybe that can create clones?18:18
scasquantum teleportation does that, but it's not ready18:19
jrollwe just need a time machine to go to the future and steal the other technologies18:19
scasi'm a bit larger than a photon18:19
*** dklyle has quit IRC18:23
*** dklyle has joined #openstack-tc18:40
*** ricolin has quit IRC18:53
TheJuliajroll: ++ but would that also mean that we would have teleportation?19:02
*** rosmaita has quit IRC19:03
*** e0ne has quit IRC19:06
*** cdent has quit IRC19:14
*** diablo_rojo has quit IRC19:22
*** e0ne has joined #openstack-tc19:48
*** e0ne has quit IRC20:51
*** amrith has joined #openstack-tc20:57
*** amrith- has quit IRC20:57
*** amrith is now known as Guest3949020:57
*** edmondsw has quit IRC21:00
*** edmondsw has joined #openstack-tc21:01
*** edmondsw has quit IRC21:01
*** Guest39490 has quit IRC21:01
*** edmondsw has joined #openstack-tc21:01
*** amrith has joined #openstack-tc21:33
*** amrith is now known as Guest4309321:33
smcginnisdhellmann: Was that your own script you ran this morning that showed election candidates?21:36
dhellmannsmcginnis : no, it was a command in the election repo (tox -e venv -- render-statistics -v)21:37
smcginnisOh, nice.21:37
* dhellmann is feeling all smug for looking to see if that existed before he wrote one21:37
smcginnisHah21:38
smcginnis7 without candidates, one of which is Security.21:39
dhellmannwhich 7?21:40
dhellmannor 6 I guess21:40
smcginnisDragonFlow, Freezer, Loci, Packaging RPM, Searchlight, Winstackers21:43
smcginnisWait, refstack too.21:43
smcginnisMaybe I didn't collate that right in my head. We have a few candidate patches not accepted yet.21:43
smcginnisAssuming all of those are qualified candidates.21:43
*** mriedem has quit IRC22:14
tonybOn July 25th I sent an email to openstack-tc requesting the governance repo be tagged.  I failed to notice it got caught in moderation and is still there.  The TL;DR is can we, the election officials, please have the governance repo SHA cefd60f94f77066bf33d396811c6320de03eaa3e tagged as aug-2018-elections22:21
* tonyb can generate at list of open projects but it'll just be based on render-statistics22:23
*** edmondsw has quit IRC22:23
tonybdhellmann: can I graft your reviews into my tools-fix series mostly to get https://review.openstack.org/#/c/587649/122:23
*** edmondsw has joined #openstack-tc22:23
tonybZun and Trove will need TC attention22:26
dhellmanntonyb : go for it22:29
tonybdhellmann: Thanks22:29
dhellmanntonyb : I will tag the governance repo tomorrow22:30
tonybdhellmann: Thanks22:30
dhellmannactually, let me see if I can just do it now22:30
dhellmannI assume you need a signed tag22:30
*** edmondsw has quit IRC22:30
tonybdhellmann: Yes please22:31
*** edmondsw has joined #openstack-tc22:31
dhellmanntonyb : done22:31
tonybdhellmann: Thanks, that's perfect!22:32
fungiyeah, having a tagged point in history of the governance repo makes it easier to see later what teams/deliverables/extra-atcs were considered22:33
fungias opposed to just "whatever the head of master was at the time we ran stuff"22:34
tonybI've hacked up a tool that makes it a little quicker to setup an election so once the PTL elections have concluded we can keep the momentum up and set the TC elections up22:34
fungiawesome22:35
tonybpart of that will be to add an item to the release scheule for 'tag the governance repo'22:35
fungiwith ptl and tc elections moving around relative to one another, that's hard to do until we know the release schedule22:36
tonybWe can *probably* even set the 'T' dates ... at least pencil them in22:36
tonybfungi: Yeah a little but the next TC election is fixed to the Berlin summit, after that it gets hard to do anything other then pencil them in22:37
scasfungi: i resemble that remark about the head of master! that is, until after the next gerrit downtime22:38
smcginnistonyb: We can add something to the releases process doc to note it as a task that needs to be done at a certain point in the cycle.22:44
tonybsmcginnis: Thanks, that was my plan. to include it as a x-project item when we add the otehr election details22:46
smcginnistonyb: That's a little different than what I'm saying. We have the process.rst doc that we follow that notes at what points in the cycle certain tasks should be performed. I wonder if that is a better place for it than in the actual cycle schedule itself.22:47
tonybsmcginnis: Oh yeah that could work.  It feels a bit like 'crossing the streams' to me in that PROCESS.rst is tasks the realease team needs to do to get a release out the door (in general) where as this is definately a thing for the election officials to do with the help of the TC.22:49
tonybsmcginnis: I see it mostly the same as 'extra ATC updates' but the bottom line is as long as it happens in a reasonably controlled fashion it doesn't matter where it's tracked ;P22:50
dhellmanntonyb : we have several other items in that list to remind other teams already22:51
dhellmann#keepersofthecalendar22:51
tonybdhellmann: Okay.22:51
* tonyb will come up with a chnage to process.rst22:54
*** hongbin_ has quit IRC23:00
*** edmondsw has quit IRC23:11
*** tosky has quit IRC23:35
mnaserMaybe for23:58
mnaserFor tomorrow’s office hour we need to23:58
mnaserFigure out about the projects needing attention23:58
mnaser(Sorry extra enters)23:59

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