16:00:02 <harlowja_at_home> #startmeeting oslo
16:00:03 <openstack> Meeting started Mon May 16 16:00:02 2016 UTC and is due to finish in 60 minutes.  The chair is harlowja_at_home. Information about MeetBot at http://wiki.debian.org/MeetBot.
16:00:04 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
16:00:07 <openstack> The meeting name has been set to 'oslo'
16:00:09 <harlowja_at_home> courtesy ping for amotoki, amrith, bknudson, bnemec, dansmith, dhellmann, dims, dougwig
16:00:09 <harlowja_at_home> courtesy ping for e0ne, flaper87, garyk, gcb, GheRivero, haypo, ihrachyshka, jd__
16:00:09 <harlowja_at_home> courtesy ping for jecarey, johnsom, jungleboyj, kgiusti, kragniz, lifeless, lintan, Nakato
16:00:09 <harlowja_at_home> courtesy ping for ozamiatin, rbradfor, redrobot, rpodolyaka, sergmelikyan, sileht, spamaps, sreshetnyak
16:00:09 <harlowja_at_home> courtesy ping for sreshetnyak, stevemar, therve, thinrichs, toabctl, viktors, zhiyan, zzzeek
16:00:18 <ihrachys> o/
16:00:22 <nikhil> o/
16:00:41 <harlowja_at_home> btw for people that want to be in that ping list
16:00:47 <rpodolyaka> o/
16:00:50 <rbradfor> o/ \o
16:00:51 <amrith> ./
16:00:53 <kgiusti> o/
16:00:53 <harlowja_at_home> i just updated https://wiki.openstack.org/wiki/Meetings/Oslo#Ping_script with a crappy script, adjustments welcome (or add yor own)
16:00:59 <jungleboyj> o/
16:01:10 <bknudson> I have to get to the airport so won't be here.
16:01:16 <harlowja_at_home> k
16:01:53 <harlowja_at_home> #topic Red flags for/from liaisons
16:01:58 <flaper87> o/
16:02:04 <bknudson> none that I know of for keystone
16:02:15 <jungleboyj> Nothing from Cinder.
16:02:19 <harlowja_at_home> so i believe taskflow release caused some *existing* bugs to show up in some folks code (cinder, octavia)
16:02:19 <amrith> none for Trove.
16:02:42 <ihrachys> I am not aware of any for neutron, but hey, I am just back from 2week post-summit vacation...
16:02:52 <harlowja_at_home> jungleboyj, do u know if those cinder + taskflow things got fixed up
16:04:14 <redrobot> o/ ish
16:04:19 <harlowja_at_home> ihrachys, thx seems like mostly issue free :-P
16:04:32 <jungleboyj> harlowja_at_home: Yeah, I saw there was a patch pushed up for that at the end of last week while I was out.
16:04:38 <harlowja_at_home> k
16:04:39 <harlowja_at_home> cool
16:04:54 <jungleboyj> harlowja_at_home: So, I think we got beyond it for now.
16:04:57 <amrith> harlowja_at_home, heads-up on an eminent red-flag. I'm going to rake up the isotime() issue next week.
16:05:07 <harlowja_at_home> jungleboyj, sweet
16:05:09 <harlowja_at_home> amrith, lol
16:05:15 <harlowja_at_home> ah the ye-olde isotime
16:05:21 <harlowja_at_home> jd__, loves that one
16:05:43 <harlowja_at_home> #topic Releases for newton
16:05:54 <dhellmann> o/
16:05:57 <harlowja_at_home> so I haven't gotten a new release of all the things out yet, but i'll work on that in a bit
16:06:10 <harlowja_at_home> anyone need anything to be released explicitly?
16:06:46 <harlowja_at_home> (if not i'll just make a release for all oslo libs, and people can yell at me if its incorrect)
16:07:42 <rpodolyaka> :)
16:07:51 <harlowja_at_home> #topic Delimiter
16:07:55 <nikhil> o/
16:08:03 <harlowja_at_home> nikhil, do u have some time to describe the whole thing here :)(
16:08:09 * harlowja_at_home hands mic to nikhil
16:08:13 <nikhil> here's the agenda item:
16:08:15 <nikhil> Discuss the home of quota (delimiter) library http://lists.openstack.org/pipermail/openstack-dev/2016-May/thread.html#93959
16:08:30 <nikhil> please follow along the thread to know what exactly is happening here
16:08:45 <nikhil> but without going divergent into things, I can give a tl;dr;
16:08:49 <harlowja_at_home> sure :)
16:08:52 <nikhil> there are 2 school of thoughts:
16:09:19 <nikhil> 1) people who think this is a cross project library but the problem domain and solutions are not well defined hence it cant be x-prj yet
16:09:55 <nikhil> 2) another set of people who think that if we need to make it a successful x-prj library then we need it in a umbrealla that is already a cross project established domain
16:10:26 <nikhil> SO, to established the right apporach here
16:10:40 <nikhil> I need your vetted input on what are the tradeoffs to put this in oslo
16:11:02 <nikhil> what can be the good things (and of course I shouldn't be asking here :-)) what would be the warnings?
16:11:16 <harlowja_at_home> everything is good when its in oslo :-P
16:11:27 <dhellmann> I think I fall in group 2, and I suggested creating an oslo.quota lib repo to host the work as a sub-project in oslo. I'm not entirely opposed to the project being stand-alone though.
16:11:38 <dhellmann> I just think the road to adoption will be longer that way.
16:12:12 <nikhil> (context: both the school of thoughts have pretty experienced and influencial people)
16:12:16 <dhellmann> we already have well-established processes for subteams like with messaging, db, and policy so we have the pattern of setting up a team like this
16:12:35 <harlowja_at_home> so if we as a group have it be a sub-project nikhil it doesn't guarantee adoption (i assume this is known)
16:12:39 <harlowja_at_home> just a word of warning there
16:12:51 <harlowja_at_home> adoption still hard (believe me, i know)
16:12:51 <dhellmann> yeah, that's true -- there are no guarantees either way
16:13:12 <dims> dhellmann : since we tried this oslo.quota before and there's a lot of resistance from folks who have that "feature" i thought it would be better to do it this way as a standalone lib
16:13:21 <SpamapS> What's the reason to not have it in oslo?
16:13:54 <SpamapS> oslo is all of the common "business logic" that we share in OpenStack, isn't it?
16:13:55 <dhellmann> dims : there seems to be a lot of support for sharing something now, though? what did we actually have to share in the past?
16:14:08 <dhellmann> SpamapS : that's more or less the idea
16:14:09 <amrith> dims, I believe that resistence in the past was more because of what the quota library sought to do and how. I believe that more consensus is being built in this process and therefore the odds of adoption may (and I stress may) be higher.
16:14:09 <harlowja_at_home> one reason i can think of (others may disagree) is limited people in oslo to maintain it (if the sub-team goes away) SpamapS
16:14:46 <dims> amrith : we can always adopt the library when it has some legs.. like os-profiler
16:14:47 <dhellmann> harlowja_at_home : that's true. we have the same risk for any piece of code anywhere in openstack.
16:14:59 <harlowja_at_home> dhellmann, good point
16:15:26 <SpamapS> harlowja_at_home: I do understand that. However, if it is standalone and the team fades away, the projects that adopt it are still in deep water, but without a known team to help them migrate off it while it is deprecated.
16:15:38 <dhellmann> dims : that's true. I think this team is doing a better job of collaborating than the os-profiler team did in its early days, though.
16:15:41 <harlowja_at_home> SpamapS, understood
16:15:48 <dims> dhellmann : yep
16:16:05 <SpamapS> Is it being done as a total new thing, or did they start with a forklift from some other project's quota system?
16:16:17 <harlowja_at_home> #link https://github.com/openstack/delimiter (if people want to look through the code that exists so far)
16:16:17 <dims> SpamapS : i'd give them a bit more time to come up with something that works with early adopters
16:16:24 <harlowja_at_home> yes, most of that code i put up over the weekend, lol
16:16:46 <nikhil> SpamapS: a bit of both
16:16:56 <nikhil> but the roadmap is still undefined
16:16:59 <SpamapS> if it is not a forklift, then I have a 3rd option: start over. Forklift from the best implementation. Otherwise we end up in the Neutron space again.
16:17:09 <nikhil> and we do not know who is going to work on this long term
16:17:14 <dims> SpamapS : this is starting over
16:17:24 <amrith> dims, that is true too.
16:17:26 <harlowja_at_home> SpamapS, i believe its mostly starting over (but doing it strategically/smartly)
16:17:38 <nikhil> harlowja_at_home: correct
16:17:48 <nikhil> this is more strategic effort
16:17:56 <harlowja_at_home> right
16:18:00 <nikhil> and we need some space for that to happen
16:18:05 <harlowja_at_home> right
16:18:09 <dims> nikhil : if we don't see good progress on early adopter project and or if the team loses steam, we don't want oslo core team to be burdened as well
16:18:13 <SpamapS> So, put me down as somebody who is very interested in cross-project harmony, and who does not want to see another Neutron situation (I know this is a library and not a service, but the fact remains that whole new efforts are far more painful than incremental migrations)
16:18:19 <nikhil> possibly not have a brand name tagged from the get go (mostly my worry)
16:18:32 <nikhil> dims: I agree and share the sentiment
16:18:59 <dims> At the moment the team has a space and talking to whole bunch of folks and iterating, i'd keep that going and revisit later personally
16:19:23 <amrith> dims makes a convincing case. I'll +1 his suggestion.
16:19:23 <dims> there's no rush in my mind to make it "oslo". i'd prefer to have something working first
16:19:42 <nikhil> Also, I do not want some people like (say the liaison from PWG) to stop contributing just because the space belongs more to developers
16:19:44 <harlowja_at_home> i'm personally also fine with that, like we can check-in a couple of weeks from now, see where things are at
16:19:52 <dims> Oslo is great at having very few people maintain a lot of code...not for this kickstart
16:20:03 <dhellmann> the main question in my mind is how are you going to know it works if it's called experimental and  being run not as an official project -- who is signed up to adopt what is built?
16:20:09 <SpamapS> dims: the point at which to push for a decision on namespace is probably when a project starts adopting it.
16:20:12 <harlowja_at_home> i think glance signed up :-P
16:20:17 <dims> SpamapS : yep
16:20:22 <nikhil> also, say I have a good intern working on it, she might be a little bit apprehensive on contributing to something like oslo with the experts
16:20:25 <dims> dhellmann : Nova too
16:20:25 <SpamapS> Keep the blast radius for a namespace change to the early adopters.
16:20:38 <harlowja_at_home> nikhil, we don't bite :-P
16:20:44 <nikhil> :)
16:20:45 <dims> lots of feeback from jaypipes and others
16:20:57 <dims> and amrith is keeping them honest
16:21:06 <nikhil> dhellmann: I have a bunch of people in glance who don't want to import code just because it's oslo too
16:21:17 <nikhil> so we have all sorts of ideas in openstack!
16:21:38 <nikhil> I do not know for sure who is going to sign up
16:21:50 <harlowja_at_home> i wrote a skeleton, my job is done :-P
16:21:50 <nikhil> and that's the learning process, I think
16:21:54 <dhellmann> great, so one of the reasons for doing this separately is because folks don't want to collaborate with the oslo team? happy days.
16:22:06 <amrith> i'm concerned about the comment that there may be people who won't import code just because it is oslo but that is a diversion
16:22:16 <harlowja_at_home> ya, let
16:22:19 <harlowja_at_home> *lets table that for now
16:22:24 <amrith> that (to put it mildly) sucks.
16:22:47 <nikhil> I agree
16:22:49 <harlowja_at_home> nikhil, so how about we have a check-up in 2 weeks, see where things are at?
16:22:57 <harlowja_at_home> u don't have to cough or anything ...
16:23:01 <nikhil> and I spend my 10% enerygy fighting that sort of ideas
16:23:19 <nikhil> energy*
16:23:33 <nikhil> so I do not know what the right ways are
16:23:36 <dims> nikhil : if they prefer hangouts or other forms of communication to make them comfortable, let us know
16:23:40 <nikhil> I am trying to be honest on my worries here
16:23:50 <harlowja_at_home> thx :)
16:23:53 <nikhil> and being open about getting feedback on tradeoffs
16:24:06 <SpamapS> That's weird because, they could.. join.. the oslo team..
16:24:15 <nikhil> dims: whatever makes the cross project experts comfortable
16:24:29 <SpamapS> Except it's an exclusive club and we only accept non-nerf-herders right? ;)
16:24:34 <dhellmann> nikhil : the only thing that would be different about this project if it was an oslo lib is the name, and the fact that the existing oslo-core team would also have +2 on the repo.
16:24:43 * SpamapS joins the oslo team as a "we" in that sentence whenever it suits him
16:24:47 <nikhil> all I worry about is the end goal - I do not care how we achieve it (we need  a successful quota implementation in openstack)
16:24:55 <dims> SpamapS : ++ :)
16:25:06 <dims> nikhil : agree.
16:26:03 <nikhil> dhellmann: I am okay with oslo or non-oslo as long as we won't be spending another 2 months discussing the weirness in the spec
16:26:10 <nikhil> that was 1)
16:26:22 <nikhil> 2) we are okay to put this on "probation"
16:26:33 <harlowja_at_home> soooo, checkup in 2 weeks ok with folks? that's somewhat my preference, let's get people to show up to code stuff (besides myself) and see where it goes a little (or that's what i feel)
16:26:43 <nikhil> 3) there's freedome to adopt different approaches and no issues with api of the lib version breaks
16:26:48 <dhellmann> harlowja_at_home : yeah, that's fine. If the team doesn't want to be part of oslo, we can spend our energy elsewhere.
16:27:02 <harlowja_at_home> incase anyone wants to review
16:27:03 <nikhil> 4) we do not have a complete re-implmentation of this lib in oslo just because it's not the oslo way
16:27:04 <harlowja_at_home> #link https://review.openstack.org/#/q/project:openstack/delimiter
16:27:14 <dims> dhellmann : i think of this as a POC, let them get some work done, see if they can make folks like amrith, jaypipes etc comfortable with the impl and they can apply for governance
16:27:18 <nikhil> 5) we test this in production before making it non on probation
16:27:22 * nikhil done
16:27:45 <dhellmann> nikhil : you will find, as we have learned and could help you work around, that projects adopting this library and trying to maintain CD compatibility will be upset if you start making lots of API changes that break their test systems.
16:28:07 <dhellmann> that's regardless of whether it's official, or who runs it, or what program it's in
16:28:22 <dims> dhellmann : ++
16:28:29 <amrith> nikhil, I think it is important (for the reasons that dhellmann just brought up) that we spend more time now and get the API right.
16:28:44 <dims> amrith : ++
16:28:48 <dhellmann> so if you're not prepared to deal with that situation, I think you need a new plan for getting the library into production
16:28:58 <amrith> FWIW, that's one of the reasons I disagree with jaypipes proposed API. We agree about the implementation now but the API is still a concern to me.
16:29:10 <nikhil> amrith: correct and that's the end goal I am talking about as well
16:29:12 <amrith> it will be time well spent if we can get that API acceptable to more, at this stage.
16:29:40 <harlowja_at_home> api i put up amrith @ https://github.com/openstack/delimiter/blob/master/delimiter/engine.py#L20 (feel free to change it)
16:30:05 <harlowja_at_home> main engine access point would be @ https://github.com/openstack/delimiter/blob/master/delimiter/__init__.py#L23
16:30:10 <nikhil> harlowja_at_home: I expect >10 changes tbh (once people from diff projects actually start giving input)
16:30:21 <harlowja_at_home> nikhil, likely :)
16:31:07 * amrith mumbles something about main entry point being a comment :)
16:31:10 <nikhil> plus we still haven't settled on the types of drivers
16:31:41 <harlowja_at_home> right, work to be done
16:31:52 <nikhil> yep, there's lot to go forward on here!
16:31:58 <amrith> nikhil, what do you mean about types of drivers? Should we take it offline?
16:32:05 <amrith> i.e. this is the #oslo meeting
16:32:08 <amrith> not the #quota meeting
16:32:11 <harlowja_at_home> ;)
16:32:13 <nikhil> amrith: yes, please
16:32:18 <amrith> ok, 2pm today
16:32:42 <harlowja_at_home> ok, soooo let's re-connect in a couple of weeks on delimiter, i think that's sane thing to do
16:32:59 <harlowja_at_home> seem ok to folks?
16:33:20 <amrith> +1
16:33:32 <dims> ++
16:33:50 <amrith> ok, fine. +=1
16:34:21 <nikhil> I am happy to reconnect wherever to get everyone on the same page
16:34:32 <harlowja_at_home> cool
16:35:14 <harlowja_at_home> #topic Stuck reviews/specs
16:35:27 <harlowja_at_home> any stuck reviews or specs that we can try to focus on for folks?
16:36:46 <harlowja_at_home> #link http://bit.ly/1oDpFQ5 (all oslo reviews)
16:36:59 <harlowja_at_home> specs @ https://review.openstack.org/#/q/project:openstack/oslo-specs
16:37:15 <harlowja_at_home> guess everyone getting everything reviewed then, nice :)
16:37:35 <kgiusti> harlowja_at_home: what did I mess up https://review.openstack.org/#/c/314603/
16:37:46 <kgiusti> harlowja_at_home: it doesn't appear in the spec list
16:38:03 <harlowja_at_home> https://review.openstack.org/#/q/project:openstack/oslo-specs ?
16:38:04 <kgiusti> harlowja_at_home: at least on that short url, that is
16:38:07 <harlowja_at_home> ah
16:38:14 <harlowja_at_home> maybe gotta update the short-url
16:38:31 <harlowja_at_home> ya, weird whats up with that
16:38:33 <harlowja_at_home> i blame gerrit
16:38:45 * kgiusti is good with that :)
16:39:38 <harlowja_at_home> i'll regenerate the short-link a little later, maybe something wonky with it
16:39:50 <harlowja_at_home> or ya, evil gerrit, lol
16:40:03 <kgiusti> Just offhand - when is the feature freeze date for oslo.messaging features in newton?
16:40:06 <rbradfor> harlowja_at_home, re specs, working around the context policy things,  https://review.openstack.org/#/c/290907/  has been discussed. I think it's impact needs input from other projects
16:40:33 <harlowja_at_home> let's see here http://releases.openstack.org/newton/schedule.html
16:40:35 <rbradfor> - Standardize Context Arguments spec,  especially if projects abuse putting name values into id releated fields.
16:40:56 <amrith> I have a q about this spec
16:41:08 <amrith> Will this change the oslo.context object as it is today?
16:41:19 <amrith> Trove currently uses it and sends it in an RPC message to guests.
16:41:29 <amrith> if the object changes, that'd be bad for Trove.
16:41:32 <harlowja_at_home> kgiusti, i believe thats 'Aug 29-02' for this
16:42:04 <kgiusti> harlowja_at_home: looks like it's going to be a busy summer....
16:42:24 <harlowja_at_home> rbradfor, agreed, it'd nice if some liasons could look at 290907 (for those that use oslo.context)
16:42:40 <amrith> rbradfor, harlowja_at_home ... question abotu that spec ^^
16:42:52 <rbradfor> amrith, well it's about providing a backwards compatible solution, (at least one full cycle), but moving from the N projects that subclass just to rename an attribute from the the context values
16:43:24 <rbradfor> amrith, I'd like to hear of all problems, so we don't cause issues, such as with Trove
16:43:40 <amrith> rbradfor, I will chat with you about this.
16:43:45 <amrith> once I understand the spec better
16:43:48 <amrith> thanks for the heads up
16:43:54 <harlowja_at_home> rbradfor, if u want me to send out a ML email, i'd be willing to as well
16:44:06 <harlowja_at_home> get people to start thinking about that ...
16:44:13 <rbradfor> amrith, trove does none of the mess we are trying to improve on
16:44:17 <harlowja_at_home> (people that may not be here that is)
16:44:26 <rbradfor> harlowja_at_home, ML will do.
16:44:29 <amrith> rbradfor, give us time :)
16:44:35 <amrith> rbradfor, don't challenge us :)
16:45:18 <harlowja_at_home> #topic Open discussion
16:45:20 <rbradfor> amrith, it's just a spec, I wrote it over 2 months ago, and it will probably be a long time before enough people comment on it, so time is on your side
16:45:33 <harlowja_at_home> lol
16:45:37 <amrith> ha
16:45:46 <amrith> let me go buy a lottery ticket
16:45:55 <harlowja_at_home> time is on your side for those as well
16:46:24 <kgiusti> but the odds sure as heck are not
16:46:28 <harlowja_at_home> ;)
16:46:40 <amrith> kgiusti, that's true ...
16:47:12 <harlowja_at_home> anything else anyone wants to bring up?
16:47:19 <harlowja_at_home> rbradfor, when we gonna start that blog stuffs ;)
16:47:39 * dims runs and hides
16:47:40 <rbradfor> harlowja_at_home, yeah, I keep dreading that question.
16:47:44 <harlowja_at_home> lol
16:47:55 <rbradfor> I know it's important for Oslo and all
16:48:20 <rbradfor> but I proposed a plan and made a schedule, so time to man up
16:48:43 <harlowja_at_home> i'd be up for trying at least
16:48:48 <harlowja_at_home> if it doesn't work out,that's ok
16:49:08 <rbradfor> harlowja_at_home, let me dig up my half done kick of post and run it by you Wednesday
16:49:16 <harlowja_at_home> cool
16:49:31 <rbradfor> harlowja_at_home, thanks for bringing up the topic
16:49:56 <harlowja_at_home> np, at least we can try, that's what matters to me
16:50:11 * harlowja_at_home needs to hire a ghost-writer, ha
16:50:44 <rbradfor> you can't ghost write technical, e.g. code blocks.
16:51:04 <harlowja_at_home> hmmm, good point, lol
16:52:28 <harlowja_at_home> okie dokie, back to your regular scheduled programming in 3
16:52:31 <harlowja_at_home> 2
16:52:38 <harlowja_at_home> 1
16:52:47 <amrith> thanks harlowja_at_home
16:52:51 <harlowja_at_home> np :)
16:53:03 <harlowja_at_home> #endmeeting