Friday, 2018-01-12

openstackgerritGhanshyam Mann proposed openstack/governance master: Add tempest-stress in Quality Assurance
*** liujiong has joined #openstack-tc01:53
*** mriedem has quit IRC02:42
*** rosmaita has quit IRC02:47
*** liujiong has quit IRC06:09
*** liujiong has joined #openstack-tc06:15
*** liujiong has quit IRC06:34
*** harlowja has quit IRC06:35
*** tdasilva has quit IRC07:07
*** harlowja has joined #openstack-tc07:10
*** harlowja has quit IRC07:10
*** tdasilva has joined #openstack-tc07:21
ttxdhellmann: the new ptgbot lets you dynamically allocate rooms to a track (beyond the pre-allocation) so we could use that flexibility to pick the best time on the spot07:51
*** jpich has joined #openstack-tc08:55
*** cdent has joined #openstack-tc09:30
*** dtantsur|afk is now known as dtantsur09:42
*** gcb has quit IRC09:55
*** dtantsur is now known as dtantsur|brb11:59
*** rosmaita has joined #openstack-tc12:24
*** jpich has quit IRC12:43
openstackgerritWitold Bedyk proposed openstack/governance master: Update goals status for Monasca
*** dtantsur|brb is now known as dtantsur13:30
rosmaitaEmilienM what's the deadline for community goals proposals?13:49
*** david-lyle has quit IRC13:49
*** mriedem has joined #openstack-tc14:18
dhellmannttx: we could. I'd like us to put it on the schedule, though, so that people plan around it. We tend to emphasize letting TC members spend time with other teams. Let's try the other way this time?14:53
dhellmannrosmaita : I'm not sure we set a deadline. ttx? We want to have things decided before the PTG, so sooner is better. Did you have something in mind?14:54
ttxdhellmann: works for me. Friday morning ?14:54
dhellmannttx: yeah, that's what I was thinking, to avoid any issues with folks leaving early14:54
ttxdhellmann: rosmaita: EmilienM coordinates goal collection -- not sure if he had a deadline in mind. I think we should select them before FF14:55
ttxi.e. stop collecting suggestion next week and discuss/vote which the next week or so14:55
dhellmannmakes sense. maybe we should list that as a deadline on the release calendar for next cycle14:56
rosmaitaok, cool, i will get something up over the weekend14:56
rosmaitai am going to propose a free-form goal14:57
dhellmannrosmaita : "free-form"?14:57
rosmaitayou'll see ... i need to do some stuff in glance that doesn't apply to everyone14:58
rosmaitaprobably everyone else is in same position14:58
dhellmannhmm, ok. you've piqued my interest.14:58
rosmaitaglance could use a cycle off from community goals, some of the previous goals have piled up some tech debt, but i saw the discussion that taking a cycle off would be bad, so this is my attempt to have it both ways15:00
EmilienMttx, rosmaita: deadline is PTG imho15:01
EmilienMeven before15:01
dhellmannEmilienM : we need to decide in time to have PTG time lined up to discuss the work, right?15:02
EmilienMvery true15:02
rosmaitaEmilienM thanks, it was the how-far-before part i wasn't sure of15:02
EmilienMI think right now we have some goals and some champions, we might have decided within the next 2 weeks15:02
EmilienMso I would say "in 2 weeks we have goals & champions" if things continue to go right15:02
cdentdhellmann, ttx: I agree with putting it on the schedule for the reasons dhellmann says15:03
ttxEmilienM: yes, well before PTG since we need to know if we want rooms to work on them or not15:09
EmilienMyeah ok15:09
ttxIdeally we'd decide before start of cycle, which arguably starts when we branch stable/queens and master swicthes to Rocky15:09
ttxwhich happens over the coming month15:10
cmurphyis there going to be a community-wide vote on goals? or ml discussion or ptg room? or is it just the tc that decides on our own?15:12
ttxML discussionm TC decision15:12
* ttx should not try to discuss in two channels at the same time15:12
*** dansmith is now known as superdan15:31
*** david-lyle has joined #openstack-tc15:38
*** kumarmn has joined #openstack-tc15:51
*** kumarmn has quit IRC15:57
*** hongbin has joined #openstack-tc16:02
*** kumarmn has joined #openstack-tc16:05
*** dtantsur has quit IRC16:37
-openstackstatus- NOTICE: Zuul has been restarted and lost queue information; changes in progress will need to be rechecked.16:46
*** kumarmn has quit IRC16:47
*** kumarmn has joined #openstack-tc16:50
*** kumarmn has quit IRC16:54
ttxInteresting data point from OVH on Keystone v3 adoption:
cmurphythat is interesting17:03
ttxwe could ask him for more raw data (this one in particular, but I'm pretty sure they would be open in sharing other type of information too)17:15
jrollttx: the most interesting bit is that only 26% of tools have sufficient documentation about what sorts of clouds they do and don't support17:55
persiaA related interesting question: how is that tooling determining how to interact with keystone?  Is there API discovery, are people writing their own clients based on docs, are people importing a known client?  Some of that 74% could be users that don't care about versions (like shade consumers).18:00
fungismcginnis: cdent: condolences on not getting board seats this time, but please run again next time if you're still up to it!18:04
EmilienMyes, ++ ^18:05
cdentThanks fungi. Next time seems a very long way away, who knows what the world will look like, but it's definitely on the radar for now.18:05
fungiwe did still get some board members elected who are strong allies of our technical community, even if we didn't get any current technical committee members elected to the board this round18:06
cdentIndeed. I've already informed mordred that he has to carry a torch.18:07
*** david-lyle has quit IRC18:12
mordredpersia, ttx: I just tweet-asked him if he has break-down numbers he can share .. we've also been trying to get at least all of the official python-based things to provide meaningful user-agent strings18:17
mnaserthis sounds like a fun topic18:33
* mnaser ssh's into keystone api servers18:33
mnaser82% of our requests are v318:37
mnaser supports keystone v3 (but disappointingly v1 image api only)18:38
* persia wonders if there is benefit in setting up a gamable stats publication site to allow operators to brag about how "the most advanced customers use this cloud"18:41
smcginnisThanks fungi! It was worth a shot.18:47
mtreinishmnaser: I'm a proud member of the 18% :p (mostly because I'm too lazy to update my scripts for daily snapshot)18:49
mnasermtreinish: i do admit it is a bit more tedious figuring out the game of bingo to get the right combo of vars configured18:50
*** dtantsur has joined #openstack-tc18:50
mnaserbetween project and user and project domain and user domain, etc18:50
mtreinishmnaser: yeah, that is a bit annoying18:52
mtreinishalthough in my case I'm using and the tempest clients in those scripts so there is a bit more work involved to make the switchover18:53
mtreinishmostly because I have to remember how I put it together :p18:54
mnasermtreinish: whats more fun is when you forget how you put it together, build a big plan in your head how you'll rebuild it even better with all these ideas, and realize its what you did in the first place18:54
mnaser"thanks old me!"18:54
*** mriedem has quit IRC19:25
*** harlowja has joined #openstack-tc19:27
*** mriedem has joined #openstack-tc19:32
*** david-lyle has joined #openstack-tc19:33
fungiyeah, i enjoy occasionally getting love letters from past fungi to present fungi19:45
fungiunfortunately, it's hard to send hatemail from present me to old me for making certain design choices19:46
fungi(or life choices for that matter)19:46
*** cdent is now known as future_fungi19:58
future_fungifungi: I have updated instructions, please set up the receiver19:59
*** future_fungi is now known as cdent19:59
mtreinishcdent, fungi:
*** harlowja has quit IRC20:04
dhellmanntc-members, I could use your input on this thread:
cdentdhellmann: skimmed it, looked reasonable, will look more closely by monday20:45
dhellmanncdent : thanks20:46
cdentI definitely agree that the amount of churn in the current way is subpar20:46
*** openstackstatus has quit IRC20:46
*** openstack has joined #openstack-tc20:48
*** ChanServ sets mode: +o openstack20:48
*** openstackstatus has joined #openstack-tc20:49
*** ChanServ sets mode: +v openstackstatus20:49
pabelangerdhellmann: I don't have any objections if people would like to try using SB myself20:49
pabelangerI've seem some internal teams do something like this using trello, and seem to work well for them20:50
dhellmannreading the SB client docs, I wonder where I get a token20:52
fungii'm intrigued by the proposal, but as a past and present sb proponent i'll wait for some more balanced feedback before providing my own21:08
fungidhellmann: great question wrt token generation. i don't have the answer but suggest asking in #storyboard and then we can update the docs21:09
EmilienMdhellmann: no objection21:09
dhellmannfungi : ack, thanks21:09
fungino more than one person should need to ask this particular question before the documentation answers it21:09
dhellmannEmilienM , pabelanger : could you reply to that effect on the thread (if you haven't already)?21:09
EmilienMof course21:10
dhellmannfungi: I suspect the answer is "it depends on your auth provider"21:10
dhellmannI imagine I'll need to get the launchpad libraries involved somewhere, for example21:10
dhellmannwhich makes me wonder if we already have a CLI somewhere21:10
dhellmannthere isn't one in the python-storyboardclient repo as far as I can tell21:10
fungii think sb tokens are independent of openid, but could of course be wrong21:13
dhellmannah, ok21:13
fungimy recollection is that api tokens are a workaround for the terribleness of trying to do openid in automation21:13
fungiand came about in the era of getting the storyboard-its plugin for gerrit working properly21:14
fungiso i _think_ is using a durable api token rather than openid to authenticate its calls to storyboard.openstack.org21:14
fungii'll grant it's probably gotten so little use that the docs are lagging shamefully21:15
dhellmannyeah, docs seem to be a major stumbling block for me21:15
dhellmannwrt storyboard that is21:15
fungias with most software, the docs could use some love and contributors21:16
dhellmannmaybe the foundation can hire an english major with an interest in technical writing as an intern to write some docs21:16
fungiwas not sufficient?21:16
dhellmannwell, I didn't look there, because I was looking at the client docs21:17
fungii'll warrant the url there likely needs editing21:17
fungiand yeah, that could probably stand to at least be linked from the client docs, if not moved into them21:17
dhellmannI've asked for more detail on that page in the past. I don't know what sort of search query syntax is supported, for example21:17
fungiyep, it's vague. i believe it to the the same search syntax the webclient supports for lookups, not that the comparison is necessarily helpful21:18
smcginnisdhellmann: By CLI, I assume you are looking for something other than boardtty (or whatever it's called)21:24
smcginnisAnd +1 for storyboard for goals. That might be a good way to get more folks exposed to it and make the eventual transition away from launchpad easier.21:26
*** harlowja has joined #openstack-tc21:26
smcginnisfungi: IIRC, the scripts to migrate bugs from launchpad to storyboard use a special range of numbers/IDs. Is it possible to start using storyboard, then later migrate the existing bugs over?21:27
fungismcginnis: in theory, i suppose. yes freshly-minted sb story numbers start at a high enough offset to preclude collision with imported lp bugs21:31
fungifor future compatibility with bug references we reserved the low story numbers exclusively for imports from lp bugs21:32
smcginnisIt could cause some confusion perhaps, but I wonder if that is a way we could ease some migrations without all needing to take a giant leap at once.21:32
fungiwell, we've already mostly concluded that waiting for projects who share bugs in lp to migrate together will result in none of them ever migrating21:33
fungiand that it's probably easier to just take the divergence hit between them for the bugs where they overlap (and the import script can still import new comments on already-imported lp bugs when rerun)21:33
smcginnisThat's what I'm afraid of with cinder, nova, os-brick, etc.21:34
fungii mean, it's already the case that other relatively closely-related software (outside openstack) may already use different defect trackers, and those communities/users/devs seem to manage to get by21:35
TheJuliaI have the same fear with ironic, and all of it's various locations of lp bugs21:35
fungicomes down to cat-herding, at its most basic level21:36
persiaSome history on storyboard CLI: there was some initial work done, but none of the people worling on it cared about CLI (they were mostly API consuming folk), and even python-storyboardclient doesn't support some things in the API.  Once boartty was developed, I think just about all the work on both CLI and python-storyboardclient mostly ended.21:36
TheJuliaWell, cat herding from an authoritative cat herding position. I've already perceived resistance to change, which means that to get it done, it will have to be a "you will migrate" or a "Here is the massive value gain, now lets migrate!" situation.21:37
smcginnisOtherwise the law of inertia rules.21:38
TheJuliaYeah, the positive thing is it will be minimally interruptive to work in progress.21:38
persiatracking release goals is the best "massive value gain" I've heard proposed for SB.  Most of the other advantages of SB over LP are fairly minor (and the OpenStack community has a number of different ways to handle those, e.g. prioritisation etherpads)21:39
TheJuliapersia: well said21:40
dhellmannsmcginnis , persia : yeah, I'm looking for a way to script creation of stories and tasks for the goals stuff, so a general purpose cli I could wrap would be good but an interactive tool is less so21:41
dhellmannwe could write a custom script just for that purpose, too, I suppose21:41
persiadhellmann: If you're fine with scripting in python, python-storyboardclient can do most of that.  It's more around boards, worklists, permissions, etc. that the python client is less well tested.21:42
dhellmannyeah, we'll want it to create a board, too, but that's less onerous to do by hand than creating 50+ tasks for tracking21:42
persiaAnd you'd probably want to subscribe folk to the board, so they get notifications of changes there (as story subscribers don't get notified just because someone made a board).21:43
dhellmannI wasn't able to find a way to subscribe to notifications for a board in the UI; is that even supported?21:43
persiaI had thought so, but don't see a UI element.  Folk in #storyboard will probably know for sure.21:46
dhellmannIt's EOD for me here, but I'll follow up next week21:47
dhellmannthanks for your help, persia & fungi21:49
* dtroyer wakes up in time to see everyone start weekending…21:49
dtroyerI love the idea to put the goals in SB, as said above, having a reason to use it is a great way to get people exposed to it and make them want to migrate the other things21:50
persiadhellmann: Eventually we'll have sufficient automation and interface that we can all know everything.  Until then, we'll just have to share what we know :)21:50
fungihonestly, the biggest "massive value gain" to me is openstack still having working task/defect tracking even if lp falls over21:51
* dhellmann makes a note to print more "write it down" buttons for dublin21:51
* fungi would totally war a "write it down" button21:53
fungiwear too21:53
dhellmannnow I'm down a rabbit hole looking at sites that let you have custom pins made22:10
dhellmannthough I think I want a release ring: https://www.ringsource.com22:10
dhellmannthat's even better than a challenge coin22:10
dtroyerClass of 2018 RULZ!22:11
*** rosmaita has quit IRC22:16
fungihere's to all the 2018 queens!22:20
*** rosmaita has joined #openstack-tc22:23
*** mriedem has quit IRC22:54
*** harlowja has quit IRC23:01
*** hongbin has quit IRC23:53

Generated by 2.15.3 by Marius Gedminas - find it at!