14:19:16 <cody-somerville> #startmeeting storyboard
14:19:17 <openstack> Meeting started Thu Nov 28 14:19:16 2013 UTC and is due to finish in 60 minutes.  The chair is cody-somerville. Information about MeetBot at http://wiki.debian.org/MeetBot.
14:19:19 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
14:19:21 <openstack> The meeting name has been set to 'storyboard'
14:19:26 <mordred> o/
14:19:27 <ttx> o/
14:19:33 <NikitaKonovalov> hi
14:19:38 <mordred> lookie here. how exciting
14:20:00 <cody-somerville> #topic Introductions
14:20:05 <SergeyLukjanov> o/
14:21:10 <cody-somerville> So I thought we could first start by introducing ourselves, where we're based out of, why we're interested in storyboard, and how we're interested in helping.
14:21:44 <cody-somerville> SergeyLukjanov: Would you like to start?
14:21:50 <SergeyLukjanov> cody-somerville, sure
14:22:38 <SergeyLukjanov> I'm the Savanna PTL, I'm interested in cool issues/specs tracking system
14:22:41 <SergeyLukjanov> :)
14:22:59 <SergeyLukjanov> especially for releasing, milestones, series and etc.
14:23:31 * mordred is on infra-core, and employs a set of people who will be hacking on this. also wants to finish killing launchpad
14:23:48 <ttx> SergeyLukjanov: if everything goes well for Savanna, you might not have any time left anymore though.
14:24:00 <ttx> Hi! You know me, based in France. Mostly interested in making sure the end result stays in the original philosophy and serves OpenStack coordination needs. Might continue coding it in my copious free time, just because I like small projects.
14:24:21 <mordred> ttx: you have plenty of free time
14:24:28 <ttx> mordred: I do!
14:24:39 <ttx> all my time is free as in freedom.
14:24:57 <SergeyLukjanov> ttx, I know now how it works and I'd like to make this process better :)
14:25:11 <SergeyLukjanov> ttx, and we have a lot of other projects on stackforge
14:25:29 <ttx> SergeyLukjanov: ack
14:25:41 <cody-somerville> My name is Cody, I'm in Ottawa Canada, interested in helping realize ttx's vision, and particularly interested in helping out from architectural and implementation perspective.
14:26:08 <NikitaKonovalov> I have been developing a Horizon plugin for Savanna recently, and I'm interested in developing smart UI's, and since Storyboard is just starting, I see a great opportunity in it
14:26:38 <cody-somerville> I'm also in the process of hiring a set of people to help with storyboard development (I work for Monty).
14:27:24 <cody-somerville> Awesome.
14:27:25 <ttx> cody-somerville: do you have a vague timeframe for that hiring process ?
14:27:43 <cody-somerville> ttx: So far I have two people starting on December 9th.
14:27:56 <SergeyLukjanov> fwiw I'm dreaming about useful reporting, stats and etc...
14:28:19 <ttx> cody-somerville: are their names public yet?
14:28:30 <cody-somerville> Awesome. I think that's everybody here today. Thank you SergeyLukjanov, mordred, ttx, and NikitaKonovalov  :)
14:28:32 <ruhe> hi. i'd like to contribute all my available time to storyboard too. things i'd like to improve: search and reporting
14:28:47 <cody-somerville> Oh. Hey there ruhe!
14:28:48 <ttx> ruhe: awesome!
14:28:49 <cody-somerville> Welcome!
14:28:58 <mordred> cody-somerville, ttx: we sent out a third and fourth offer 2 days ago
14:29:10 <mordred> tentative start date dec 16
14:30:19 <cody-somerville> ttx: Haven't released any names publicly per-say.
14:30:29 <ttx> cody-somerville: ok
14:31:08 <cody-somerville> #topic Proposal: Use storyboard to develop storyboard - CD/CI from trunk to instance used track and plan storyboard development.
14:31:42 <ttx> cody-somerville: I think that's a good idea, the trick is to arrive to the point where we can actually do that. Which would be.. a stable DB schema, I'd say
14:31:49 <mordred> so, last time jeblair and I talked, he said "fix the leave-comments bug and infra can start using it now"
14:32:01 <cody-somerville> I'd like to propose that our first goal should be to take what we have and get it up and running with CD/CI from trunk and use that instance to track and plan storyboard.
14:32:05 <SergeyLukjanov> cody-somerville, +1
14:32:18 <mordred> yah. I'd like to that too - and if we have to do db migrations, then so be it
14:32:32 <mordred> otherwise I fear we're going to sit in the land of not-quite-ready for ever
14:32:44 <cody-somerville> So in today's world, I'm not sure we'll ever have a stable DB and so we might as well solve the problem of db migrations now.
14:33:10 <ttx> mordred: since I added project groups I think the biggest obvious DB overhauls are done now
14:33:17 <mordred> yah. I agree
14:33:31 <NikitaKonovalov> agree, so we have stabilize db and all frameworks around as soon as possible
14:33:32 <mordred> also, when did I start saying "yah" instead of "ja" ... very weird
14:33:48 <ttx> and using it for tracking storyboard (instead of ahem github issues) sounds like an obvious first step
14:34:00 <cody-somerville> Part of that will be investing in some features/scaffolding that is usually more indicative of a more mature project such as feature flags and incrementally landing changes that involve large db changes.
14:34:07 <mordred> ttx: also, I think infra is game to be an aggressive first guinea pig too
14:34:36 <SergeyLukjanov> currently we have a django-based PoC and planning to move Pecan/WSME backend
14:34:48 <ttx> mordred: my only regret will be that it's two projects that won't make a lot of use of cross-project facilities, but meh
14:34:51 <SergeyLukjanov> it could possible change db schema
14:34:58 <mordred> SergeyLukjanov: so ... I pitched something to cody-somerville about that the other day
14:35:02 <ttx> although you could imagine storyboard+infra stories, for sure
14:35:06 <SergeyLukjanov> in auth/users part especially
14:35:15 <ttx> oh.
14:35:17 <ttx> oh.
14:35:20 <NikitaKonovalov> we still can use wiki to keep important data safe, which db model
14:35:20 <ttx> actually...
14:35:31 <NikitaKonovalov> which -> when
14:35:33 <mordred> which is that I _think_ we might want to think about pecan/wsme as a v2 idea
14:35:38 <ttx> mordred: infra would be split by repo and use an "Infrastructure" project group
14:35:46 <mordred> ttx: ooh. excellent!
14:35:53 <ttx> since tere is a strong project/repo relationship in storyboard
14:36:03 <mordred> yah. I think that's a great idea
14:36:09 <ttx> mordred: so it would make sense to dogfood that extensively
14:36:25 <SergeyLukjanov> my concern is that if we'll continue developing current code than migration to pecan/wsme will be full rewrite
14:36:27 <mordred> SergeyLukjanov: so, if we get storyboard up and going and CI/CD on the django
14:36:42 <mordred> and then we have a period of a month where we're not landing new features because we're redoing plumbing
14:36:55 <cody-somerville> with regards to Pecan/WSME, I think we need to evaluate that further.
14:36:55 <mordred> and then we have a downtime event to do a migration - I think I'm fine with that
14:37:21 <mordred> SergeyLukjanov: it might be a full rewrite - but it might also be able to be done incrementally
14:37:21 <ttx> mordred: you could also additionally use a specific projects for random "infra" requests like tickets from people who have no clue if it belongs to config or jeepyb (i.e. most people)
14:37:38 <ttx> i.e. infra-tickets or something
14:37:40 <mordred> SergeyLukjanov: if we started by just splitting out the current db model and putting a pecan/wsme api on top of the current model api
14:37:52 <ttx> we need support for fake projects anyway, for stuff like OSSA/OSSN
14:37:57 <mordred> ttx: ++
14:38:11 <ttx> (i.e. stuff where the fix is not a commit)
14:38:24 <mordred> SergeyLukjanov: I agree though about auth getting more complex
14:38:26 <mordred> :)
14:38:31 <ttx> the more I think about it, the awesomer it gets.
14:38:39 <SergeyLukjanov> mordred, I mean that if we'll continue developing current code for the *long* time than ...
14:38:43 <mordred> yah
14:38:46 <cody-somerville> We're getting a bit offtopic.
14:38:56 <mordred> I think if we hack on it even for a little bit so that we get our hands dirty with
14:38:58 <mordred> it
14:38:59 <ttx> indeed.
14:39:10 <mordred> it'll help to be able to make better long-term framework decisions
14:39:15 <mordred> (I think we're still on topic)
14:39:25 <cody-somerville> (ttx: I'd like to solve that problem more gracefully then with fake projects, I have some ideas)
14:39:30 <mordred> because we're talking about the pros and cons of deploying soon
14:39:37 <SergeyLukjanov> agreed
14:39:53 <ttx> (cody-somerville: ok, ping me anytime)
14:40:04 <mordred> but - in any case, I think getting one up and letting a few resilient projects, like storyboard and infra and UX beat on it for a while
14:40:08 <mordred> it'll serve us well
14:40:24 <mordred> and then even if we do a full rewrite for a v2 - so be it
14:40:42 <SergeyLukjanov> sure
14:40:48 <mordred> v1 took ttx a day or so - a team of us who all grok the problem space should be able to churn out a better v2 in a couple of weeks :)
14:40:59 <ttx> cody-somerville: so the on-topic answer is yes please. Anything we need to add to current storboard before going in CI/CD mode ?
14:41:01 <mordred> grokking the problem space as well as ttx is the hardest ramp-up part
14:41:06 <ttx> mordred: "or so"
14:41:08 <SergeyLukjanov> probably some more work on v1 could help us to make right decisions
14:41:09 <mordred> ttx: :)
14:41:12 <ruhe> we also can import data from LP to storyboard to see how it look on existing projects
14:41:37 <mordred> ruhe: yes. and even if that isn't the final import tool, I think we'll learn alot about it
14:41:41 <cody-somerville> mordred: I was thinking this initial goal would be to get something up for storyboard, not necessarily for -infra as well. Are you suggesting the latter?
14:41:57 <mordred> cody-somerville: I'm suggesting that infra is quite willing to also be early guinea-pigs
14:42:01 <ttx> cody-somerville: step 1 storyboard, step 2 infra, UI
14:42:14 <mordred> since this is ultimately part of the infra program
14:42:23 <mordred> s/quite willing/eager/
14:42:34 <cody-somerville> I just want to make sure we properly manage expectations.
14:42:38 <ttx> I think organizing milestones around what it takes to get new guinea pigs is a good way to prioritize
14:42:46 <ruhe> ttx, so the plan is to get working rest backend and only then start working on UI part?
14:43:17 <mordred> ruhe: nope. we're going to keep working on the current django codebase for now, and start deploying it asap
14:43:20 * cody-somerville curses ttx for having such a shiny proof of concept.
14:43:29 <ttx> ruhe: I think the rest decoupling (as well as the potential move to SemanticUI) would be parallel to the Ci/CD/onboarding-guinea-pigs effort
14:43:38 <mordred> ++
14:43:42 <ruhe> ok, i see. thanks
14:44:16 <mordred> I actually think it's great taht ttx's POC is distractingly shiny ...
14:44:21 <ttx> I don't think anyone in storyboard or infra will require REST or Semantic UI to be present before they start using it
14:44:28 <mordred> because it points out that at its heart we can keep this relatively simple
14:44:35 <mordred> that's correct
14:44:47 <mordred> we'll require that comments work
14:44:52 <NikitaKonovalov> then, should we have a list of things to do before deploying?
14:44:53 <mordred> and issues don'tget deleted
14:44:53 <cody-somerville> To put it another way I guess, storyboard today is so good people want to see and play with it. And it's strategic to do so to establish the project. CI/CD bit makes sense. Also still supports us being able to do fully rewrite still if we decide we need to go in crazy different direction.
14:45:09 <mordred> agree
14:45:22 <mordred> main thing we need is the puppet manifest written :)
14:45:33 <SergeyLukjanov> yup
14:45:58 <mordred> luckily - we have at least 2 other django apps we run (I think) so cargo-culting some puppet should not be hard
14:46:29 <ttx> but yeah. IMHO there are 4 different areas. Concept (like that fake project discussion), Architecture (REST/UI decoulpling), UI (Semantic?), CI/CD/onboard-guinea-pigs
14:46:44 <cody-somerville> mordred: How easy is it to simulate deployment in say hpcloud of what -infra will do to get it running?
14:46:46 <cody-somerville> ttx: agreed.
14:46:50 <SergeyLukjanov> and IMO we need to collect ideas/requirements, prioritize them and map to frameforks
14:46:51 <ttx> but by using the last to drive the main milestones we stay close to our users
14:47:04 <SergeyLukjanov> s/frameforks/frameworks/
14:47:32 <ttx> We need to tackle Concept asap due to its potential impact on CD
14:47:35 <mordred> cody-somerville: pretty easy. we've got docs somewhere on how to get started with that
14:47:45 <cody-somerville> mordred: Can you shoot me some links?
14:47:56 <mordred> yes
14:48:39 <ttx> UI is what I'm the least concerned about. Changing templates is NOT difficult.
14:48:50 <SergeyLukjanov> ttx, agreed
14:49:17 <SergeyLukjanov> the main concern for me - correct and stable model / workflows
14:49:23 <mordred> ++
14:49:46 <SergeyLukjanov> we can polish model on current poc except some auth/user related stuff
14:49:54 <SergeyLukjanov> and maybe some django stuff
14:49:59 <ttx> cody-somerville: so yes, agree on CI/CD as a primary target. ideally you would cut that into work items (can't use Storyboard for that yet)
14:50:41 <ttx> Then second target is to get Storyboard itself to use it
14:51:32 <cody-somerville> Aye.
14:51:40 <cody-somerville> #topic Storyboard Sprint
14:51:47 <SergeyLukjanov> looks like the #2 point will include some work on current code to improve it's usability for developing itself :)
14:52:14 <cody-somerville> I'd like to gauge interest in people attending an in-person storyboard sprint in the new year, hosted by HP.
14:53:01 <cody-somerville> If you're interested, please send me your name, e-mail address, location, and availability in the first three months of 2014 and I'll look into the possibility.
14:53:20 <ruhe> cody-somerville, which part of the globe? ;)
14:53:52 <cody-somerville> ruhe: Once I gauge interest and know where everyone is based, I'll determine a location with the lowest total cost for everyone.
14:54:02 <ruhe> cool
14:54:08 <SergeyLukjanov> great
14:54:21 <SergeyLukjanov> in which location are you hiring guys?
14:54:23 <ttx> also if you happen to go to a given conference, include that as well, we may co-host
14:54:33 <ttx> "location": syntax error
14:54:47 <cody-somerville> That's a great point re: conference.
14:55:06 <ttx> cody-somerville: yeah, my schedule is filling so fast co-location might not be a stupid option
14:55:27 <cody-somerville> Is anyone else going to conferences early 2014?
14:55:33 <mordred> it's all I do with myself
14:56:07 <ttx> I go to FOSDEM/Brussels. Might have a Foundation sprint in UT mid-January (ski!)
14:56:46 <ttx> cody-somerville: i'll send you the requested info
14:56:54 * SergeyLukjanov have no plans for the first three months of 2014 atm
14:56:58 <cody-somerville> Thanks.
14:57:06 <cody-somerville> My e-mail is cody.somerville@hp.com
14:57:17 <cody-somerville> We're now three minutes to the hour.
14:57:24 <cody-somerville> #topic Meeting Time
14:57:29 <cody-somerville> How does this meeting time work for everyone?
14:57:34 <cody-somerville> Should we change it or keep it the same?
14:57:37 <ttx> perfect for me
14:57:41 <SergeyLukjanov> ++
14:57:52 <ruhe> ++
14:57:54 <ttx> only issue is that it can't have #openstack-meeting* because both channels are booked
14:57:56 <NikitaKonovalov> +1 to keep
14:58:02 <ttx> so we could consider moving to Fridays
14:58:11 <ttx> but then that would be on Fridays.
14:58:12 <cody-somerville> It might be difficult for people on the west coast to attend at this time.
14:58:30 <mordred> hrm. I'm tempted to say "life is hard"
14:58:43 <mordred> and we make it hard for people in other timezones to attend openstack meetings all the time
14:58:44 <ttx> cody-somerville: there will be no way to get California and deep Russia.
14:59:24 <cody-somerville> My new team is going to hate me. They're all on the west coast. 6am meeting every Thursday. :)
14:59:32 <ttx> maybe we'll have alternating meetings if you get west coast people on board
14:59:49 <mordred> are there any times which are late-night US west that are morning for russia?
14:59:51 <ttx> I'm fine with one or two hours later
15:00:06 <cody-somerville> One or two hours later also works for me.
15:00:09 <cody-somerville> Let's defer.
15:00:12 <ttx> SergeyLukjanov, NikitaKonovalov: later means late for you right ?
15:00:12 <mordred> ++
15:00:13 <cody-somerville> and keep current time.
15:00:25 <SergeyLukjanov> ttx, it's ok for me
15:00:29 <mordred> this is fine for me - new employees at hp will just learn to love it
15:00:34 <SergeyLukjanov> ttx, project meeting is much later ;)
15:00:40 <ttx> no kidding
15:00:50 <cody-somerville> #topic Quick Last Minute Things
15:00:52 <NikitaKonovalov> forrks for me
15:00:55 <mordred> one last question
15:00:59 <ttx> cody-somerville: we can set it to the next hour. that way we'll have a meeting channel
15:01:00 <mordred> what url should we use?
15:01:02 <NikitaKonovalov> forrks - > works
15:01:14 <mordred> bugs.o.o ? stories.o.o? issues.o.o? or?
15:01:26 <ttx> mordred: or storyboard.o.o
15:01:35 <ttx> (too long ?)
15:01:36 <cody-somerville> bugs.o.o for production instance
15:01:41 <cody-somerville> storyboard.o.o for our instance?
15:01:42 <SergeyLukjanov> ttx, mordred, +1 for stroyboard.o.o or sb.o.o
15:01:51 <mordred> well, we use review.o.o for gerrit, not gerrit.o.o
15:01:55 <ttx> cody-somerville: or stories.o.o ?
15:02:03 <mordred> so that we have it named after function, rather than software implementation choice
15:02:03 <ttx> "bugs" sounds a bit... reductive
15:02:14 <ttx> tasks.o.o ?
15:02:37 <NikitaKonovalov> I like stories.o.o it's much less depressing than bugs
15:02:45 <mordred> yeah. I thnk I like stories the best
15:02:46 <ruhe> i'd name this domain the same way this project is named - stroyboard.o.o
15:02:55 <SergeyLukjanov> tasks sounds good, but stories is the better name
15:03:04 <cody-somerville> what if marketing wants stories.o.o for something?
15:03:15 <ttx> cody-somerville: don't we control dns ?
15:03:46 <ttx> heck, I have summit.o.o and then never coimplained.
15:04:21 <cody-somerville> Right.
15:04:34 * SergeyLukjanov likes short aliases like sb.o.o :)
15:04:51 <mordred> cody-somerville: marketing can suck it
15:04:53 <mordred> :)
15:05:11 <mordred> openstack is a technical meritocracy, they work for us ...
15:05:28 <cody-somerville> I'd suggest we have two domains though - one for our cd/ci instance (which -infra and other beta testers can be Guinea pig on) and then another for when we actually deploy this for realz.
15:05:30 <mordred> also, I bet they'd want openstack.org/stories in any case (the less antagonistic answer)
15:05:33 <SergeyLukjanov> mordred, are you sure that they know it? :)
15:05:47 <mordred> SergeyLukjanov: I'm sure that they don't :)
15:05:52 <cody-somerville> (as they should remain two separate instances)
15:05:56 <mordred> cody-somerville: why?
15:06:07 <mordred> cody-somerville: I expect our production instance to remain CD for its entire life
15:06:18 <SergeyLukjanov> mordred, ++
15:06:22 <ttx> yeah, not convinced we need a separate domain for guinea pigs
15:06:33 <ruhe> mordred, so the first step is to configure backups :)
15:06:45 <mordred> ruhe: luckily, this is something we have in infra already :)
15:06:48 <SergeyLukjanov> but we can always move our CD
15:07:10 <ttx> we do backups now ? awesome
15:07:22 <mordred> ttx: yes! we also use trove-based mysql
15:07:29 <mordred> it's like we dogfood this cloud stuff
15:07:31 <ttx> now I'm scared
15:07:34 <cody-somerville> mordred: I suspect that for production we'll want it to be delayed a bit - little bit more conservative - not always running trunk.
15:07:38 <mordred> hub_cap tells me it works
15:07:47 <mordred> cody-somerville: I do not share that suspicion
15:07:48 <SergeyLukjanov> heh, H/2 * * * *  backup the world
15:08:00 <mordred> cody-somerville: we run most of the rest of infra directly off of git master branches
15:08:14 <mordred> zuul and nodepool both run that way
15:08:18 <cody-somerville> mordred: ie. around release time, storyboard dev shouldn't have to halt but we wouldn't want to risk disrupting openstack engineers.
15:08:49 <ttx> cody-somerville: for the rest of infra we just soft freeze
15:09:03 <ttx> hold on disruptive changes
15:09:04 <mordred> yah. no dangerous changes - which means no patches written by monty
15:09:13 <ttx> that's a good rule of thumb yes
15:09:17 <SergeyLukjanov> cody-somerville, zuul/nodepool outage is the same risk
15:09:40 <mordred> now - that said - we DO have the capability of spinning up a dev server and a feature branch
15:09:44 <mordred> if we need to at that time
15:09:50 <mordred> but I'd like to shoot for not needing that
15:09:57 <mordred> because if we do need it, it means we've done something else wrong
15:10:26 <cody-somerville> the thing is, we're bound to do things wrong :)
15:10:29 <cody-somerville> #endmeeting