Monday, 2018-07-09

*** alex_xu has quit IRC00:40
*** alex_xu has joined #openstack-tc00:41
*** edmondsw has joined #openstack-tc01:04
*** edmondsw has quit IRC01:08
*** edmondsw has joined #openstack-tc02:52
*** edmondsw has quit IRC02:57
*** diablo_rojo has joined #openstack-tc04:06
*** edmondsw has joined #openstack-tc04:40
*** edmondsw has quit IRC04:44
*** flwang1 has quit IRC04:54
*** ricolin has joined #openstack-tc04:59
*** diablo_rojo has quit IRC06:25
*** edmondsw has joined #openstack-tc06:28
*** edmondsw has quit IRC06:33
*** jpich has joined #openstack-tc06:58
*** tosky has joined #openstack-tc07:35
*** ricolin has quit IRC07:35
*** ricolin has joined #openstack-tc07:54
*** edmondsw has joined #openstack-tc08:16
*** edmondsw has quit IRC08:21
*** flwang1 has joined #openstack-tc08:30
*** rosmaita has joined #openstack-tc09:56
*** cdent has joined #openstack-tc10:17
*** cdent has quit IRC10:41
*** jaypipes has joined #openstack-tc12:06
*** cdent has joined #openstack-tc12:14
*** ricolin has quit IRC12:17
*** mriedem has joined #openstack-tc12:23
*** edmondsw has joined #openstack-tc12:24
*** edmondsw has quit IRC12:29
*** edmondsw has joined #openstack-tc12:31
*** edmondsw has quit IRC12:35
*** edmondsw has joined #openstack-tc12:37
*** edmondsw has quit IRC12:42
*** edmondsw has joined #openstack-tc12:45
*** edmondsw has quit IRC12:49
*** edmondsw has joined #openstack-tc12:51
*** edmondsw has quit IRC12:53
*** edmondsw has joined #openstack-tc12:53
*** lbragstad has joined #openstack-tc13:37
*** ricolin has joined #openstack-tc13:45
*** hongbin has joined #openstack-tc13:51
dimsfyi, ASF annual report -
*** pabelanger has quit IRC14:33
*** pabelanger has joined #openstack-tc14:33
scastaking some more time to look at how releases are being done in openstack-the-ecosystem, i don't see any real issue as to why chef can't use the releases repository with its own publishing mechanism. "i don't know what i don't know" applies especially true when working in this context. my trepidation has been largely out of lack of clarity. until a few minutes ago, i wasn't able to even say whether or14:41
scasnot moving chef's release process would work. i would be contributing to the amount that needs to be reviewed. if that's not a problem, then i don't have a problem with it14:41
dhellmannscas : there's no reason chef repos couldn't have their own release job14:59
dhellmannthe requirement is to have the tag application reviewed, not to use one of the existing release jobs to publish tarballs15:00
dhellmannwe just need to wire all the pieces together -- puppet and ansible already do something like that15:01
scasyep. that's what i've been sifting through gerrit to better understand. i don't see any major impediments after seeing what i've seen15:02
scaskolla publishes to dockerhub, which is analogous to my supermarket. i've come across the incantation to not publish tarballs, which works for me15:04
dhellmannhaving the supermarket publishing automated would be good, but it doesn't have to happen all in one step15:05
scasno, it doesn't. it's still a manual task right now, but i'll want it on a periodic15:06
scasbut i think that's all a me-thing15:06
scasENOTIME excepting, i should be able to find what i need there15:07
dhellmannwhy periodic?15:07
scassupermarket is a mirror like github. i could have it done after every version bump, but that feels like more work15:08
dhellmannah. ok. we usually treat tagging releases as the thing that triggers artifact uploads. if you're going to do it periodically, that's still a very different model than any of the other projects15:09
scas'simpsons^Wkolla did it' is my excuse. it can be done one-shot, which is more or less how it's done now15:10
dhellmannI think we're treating the kolla images as distinct from the kolla software15:11
*** cdent has quit IRC15:13
dhellmannbut you'd have to talk to smcginnis about that15:13
fungithe reason kolla is publishing snapshots is to prevent them from looking like officially supported versions, due to
fungichef would presumably be making source-based releases15:21
dhellmannah, right15:21
cmurphydiablo_rojo_phon: persia: fungi: I see there are no dates given for ptl elections on , should there be? aren't those coming up?15:22
cmurphyone of my health-check projects is wondering15:22
dhellmannit will be closer to the PTL and end of cycle15:23
jrollit's R-7 already :)15:26
cmurphyah somehow my brain thought it was R-5 already15:26
scasfungi: the artifacts on supermarket would be references back to the git projects as well as exports from supermarket as tarballs, but those tarballs would be outside the release process15:26
jrollfeels like we should at least be setting dates by now15:26
fungii agree, we're probably coming up on when we should be preparing for the next round of elections but i'm standing for a seat on the tc again so i won't be an election official at least for the tc election (which is going to be fairly adjacent to the ptl elections)15:27
scasi mean, the tool itself, stove, creates those tarballs and publishes them15:30
scasi had to do a code dive for clarity15:31
fungibut they're not compiled binaries integrating components other than what the chef team is producing, right?15:34
fungiso effectively a source-based release15:34
scasin that definition, yes15:40
fungimore to the point, the concerns in the binary artifacts resolution aren't applicable for chef releases15:42
pabelangerdhellmann: left a comment on using tox for docs builds16:01
pabelangergiven the history of tox -edocs, interesting we are considering reverting to it16:02
dhellmannpabelanger :
dhellmannin particular my response to zaneb's similar question:
pabelangerthanks! I missed that16:05
dhellmannnp. At least since there are questions I know people are reading my email. ;-)16:09
*** dklyle has joined #openstack-tc16:23
mnaserit sure would be nice if tox had some sort of inheritence model or something16:24
mnaserthat way the pti is defined in one repo16:24
dhellmannthat's an interesting idea16:27
dhellmannit sort of puts us back in a state where it would be difficult to change the jobs without breaking things, though, because of the centralization16:28
mnaserwell changing behaviour for doc jobs shouldn't really break if you're doing things the right way16:38
mnaseri'm torn on the adjutant stuff, i kinda was starting to be with it until I read " E.g. for Catalyst I'm right now building a custom Adjutant API for managing credit cards in our system/stripe, while yes the API is user accessible and requires Keystone auth"16:39
mnaserand again that totally sounds like that should be it's own service built by/for/only catalyst, dunno why it has to be part of adjutant16:40
mnaserif everyone is going to add their own logic to it, whats the actual point of it because no one will have any shared logic16:40
*** jpich has quit IRC16:52
scas'stove' has an option to not tag git repos, which is how i've been invoking it all along. it looks like i can have it exist as a playbook or a rake task with the same behavior17:06
*** ricolin has quit IRC17:09
*** cdent has joined #openstack-tc17:16
dhellmannscas : that sounds like it would be useful for building the job you want17:17
dhellmannmnaser : yeah, I took that to mean they were going to use adjutant as a framework for building those APIs rather than adding them to adjutant itself17:17
scasdhellmann: yeah, it's pretty much my only tool for publishing artifacts. i haven't done the wiring yet, but it *seems* doable17:23
openstackgerritDoug Hellmann proposed openstack/governance master: update the pti to use tox for building docs
openstackgerritDoug Hellmann proposed openstack/governance master: add
*** annabelleB has joined #openstack-tc18:01
*** mriedem1 has joined #openstack-tc18:19
*** annabelleB has quit IRC18:19
*** mriedem has quit IRC18:21
*** annabelleB has joined #openstack-tc18:22
*** mriedem1 is now known as mriedem18:22
*** annabelleB has quit IRC18:41
fungimnaser: yes, that basically echoes the concerns i reraised in my comment on the adjutant review. if nobody has the same actual api methods exposed, there will basically be no user-facing community forming around it (no real ecosystem client integration, no single sdk for it, et cetera)18:45
*** annabelleB has joined #openstack-tc18:56
fungibasically if you want to use the catalyst cloud payment api, you need to install catalyst-adjutantclient or catalyst-adjutantsdk18:58
fungiwhich presumably catalyst cloud's customer knowledge base or help desk will direct you to and document how it's used18:59
* fungi recalls rackspace directing customers to install their authentication client/library back when they didn't run keystone19:01
funginot a user-friendly experience19:02
*** tosky has quit IRC19:23
*** flwang1 has quit IRC19:35
*** annabelleB has quit IRC19:42
*** annabelleB has joined #openstack-tc19:48
*** annabelleB has quit IRC19:49
*** annabelleB has joined #openstack-tc19:50
*** dklyle has quit IRC19:52
*** dklyle has joined #openstack-tc20:35
*** e0ne has joined #openstack-tc20:36
*** annabelleB has quit IRC20:45
*** e0ne has quit IRC21:31
*** hongbin has quit IRC22:19
*** cdent has quit IRC22:47
*** ianychoi_ has joined #openstack-tc23:21
*** diablo_rojo_phon has quit IRC23:22
*** diablo_rojo_phon has joined #openstack-tc23:24
*** ianychoi has quit IRC23:25
*** edmondsw has quit IRC23:26

Generated by 2.15.3 by Marius Gedminas - find it at!