15:00:23 <portdirect> #startmeeting openstack-helm
15:00:25 <openstack> Meeting started Tue Oct 23 15:00:23 2018 UTC and is due to finish in 60 minutes.  The chair is portdirect. Information about MeetBot at http://wiki.debian.org/MeetBot.
15:00:26 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
15:00:28 <openstack> The meeting name has been set to 'openstack_helm'
15:00:45 <portdirect> hey peeps, we'll give it a few mins for people to turn up
15:01:05 <jayahn> o/
15:01:19 <srwilkers> hello
15:01:20 <mattmceuen> o/
15:01:40 <portdirect> looks like etherpad is down :(
15:02:09 <portdirect> lets use here instead for this meeting: https://beta.etherpad.org/p/yWMzLMvJTV
15:02:12 <jayahn> Yeap. Confirmed
15:02:36 <mattmceuen> now that's a fancy etherpad
15:02:57 <jayahn> Loading..
15:03:18 <jayahn> Oh, I always like beta
15:03:48 <portdirect> ok - I'm afraid people will have to copy topics over :(
15:05:05 <evrardjp> o/
15:05:08 <evrardjp> etherpad is up again
15:05:12 <portdirect> lol
15:05:42 <portdirect> so we also have this: https://etherpad.openstack.org/p/openstack-helm-meeting-2018-10-23
15:05:50 <portdirect> everyone run to the other side of the ship!
15:05:54 <portdirect> lets go
15:06:05 <portdirect> #topic summit-prep
15:06:18 <portdirect> so the berlin summit is fast approaching
15:06:35 <portdirect> I'm aware of three items that directly touch on OSH
15:06:49 <portdirect> are there any i missed?
15:06:54 <evrardjp> the indirect
15:07:00 <portdirect> ;)
15:07:06 <roman_g> o/
15:07:15 <john_W> o/
15:07:17 <jayahn> You have osh hands-on?
15:07:25 <portdirect> evrardjp: you know of any that should be on this - I know there are some cross cutting deployment tooling sessions
15:07:30 <portdirect> jayahn: yes
15:07:54 <evrardjp> Developing a Standard Deployment Tools Comparison and Deployment tools feedback (Thursday from 9:50 to 11:30) are the cross deployment tools sessions
15:08:47 <portdirect> evrardjp: awesome - I'll be sure to attend these
15:08:57 <portdirect> who else is planning on being in berlin?
15:09:05 <evrardjp> portdirect: disclaimer, some of those sessions have been attempted in the past and were not specially successful (to what I heard). I think it's still worth spending some cycles to at least hear/be present there.
15:09:14 <mattmceuen> I will be there - I 'll try to make that too evrardjp is possible
15:09:15 <evrardjp> I will be in Berlin
15:09:31 <jayahn> I will be .. with 3 other folks.
15:09:41 <evrardjp> jayahn: that's great news!
15:10:12 <portdirect> evrardjp: sounds like open sauce - full of tomatoes.
15:10:16 <jayahn> evrardjp: it is my honor to be good news.. :)
15:10:56 <evrardjp> :)
15:11:00 <portdirect> one thing I'd really appreciate help is on the: Project onboarding session
15:11:09 <evrardjp> That was my next question :p
15:11:13 <jayahn> Okay.
15:11:14 <srwilkers> i will not be :(
15:11:15 <portdirect> as this is somthing we really need to get better at
15:11:31 <mattmceuen> agree
15:11:50 <portdirect> i'll start https://etherpad.openstack.org/p/openstack-helm-berlin-onboarding
15:11:50 <mattmceuen> If we can craft material for that that could be directly reusable in a non-summit context, that would be neat
15:12:03 <evrardjp> I am not onboarded on this project, so I will attend to be onboarded :p I can give you a few tips but it's generally depending on the audience
15:12:16 <portdirect> but please feel free to add your pain points, highlights etc
15:12:28 <evrardjp> a few general 'project onboarding' tips*
15:12:39 <mattmceuen> (the best example of "reusable summit material" being of course portdirect's hands-on scripts that turned into our gates :) )
15:13:31 <portdirect> and the worst - i get both ends of the spectrum here i think ;)
15:13:35 <evrardjp> I think it's a very important point portdirect
15:13:49 <evrardjp> (the pain points)
15:14:09 <portdirect> which neatly brings us onto the next topic i think
15:14:14 <evrardjp> as these will be the friction elements and things harder to onboard.
15:14:32 <portdirect> ++
15:14:41 <portdirect> #topic Gating questions
15:14:50 <portdirect> evrardjp: the floor is yours
15:15:04 <evrardjp> I thought the summit discussion was not over!
15:15:15 <portdirect> oh - sorry!
15:15:23 * evrardjp quickly takes existing notes :)
15:15:32 <portdirect> #topic back-to-the-summit
15:15:45 <portdirect> anyone have anything else to add/raise?
15:15:49 <evrardjp> I had a question for summit
15:16:02 <jayahn> Oh..
15:16:11 <evrardjp> who will be presenting the project update session?
15:16:18 <evrardjp> portdirect I guess?
15:16:22 <mattmceuen> portdirect! :p
15:16:22 <jayahn> ptl?
15:16:29 <mattmceuen> I will be in the background with tomatos
15:16:31 <portdirect> I will be, along with mattmceuen
15:16:35 <evrardjp> well I think it would be nice mattmceuen is there
15:16:40 <evrardjp> great
15:16:54 <portdirect> I cant think of a better person to describe where we came from than mattmceuen
15:17:08 <evrardjp> indeed ;)
15:17:39 <jayahn> Then, I will be in the background with tomato!
15:17:42 <evrardjp> other question -- is there any OSH have forum sessions in Berlin?
15:17:42 <portdirect> its why hes coming to berlin :D
15:18:06 <portdirect> not directly
15:18:11 <srwilkers> ill throw tomatoes at his desk while hes gone
15:18:32 <portdirect> I need to add them, but there are some that i think we have to attend
15:18:39 <portdirect> there is one on container security
15:19:08 <evrardjp> srwilkers: take pictures for sharing on the channel :)
15:19:47 <evrardjp> portdirect: sounds nice. I think we should eventually find time to discuss together, maybe organise meal. But that can wait I guess :)
15:20:01 <roman_g> 🍅🍅🍅
15:20:02 <mattmceuen> There is a forum session for cross-container-team security, to share lessons-learned
15:20:04 <evrardjp> together == with the team
15:20:30 <portdirect> https://www.openstack.org/summit/berlin-2018/summit-schedule/events/22797/cross-project-forum-securing-containerized-infrastructure
15:20:34 <mattmceuen> so all container-related openstack teams are invited / encouraged to participate in that!
15:20:56 <evrardjp> #link https://www.openstack.org/summit/berlin-2018/summit-schedule/events/22797/cross-project-forum-securing-containerized-infrastructure
15:21:15 <evrardjp> thanks portdirect mattmceuen for the information!
15:21:52 <mattmceuen> evrardjp let's definitely grab some grub as a team, good idea :)
15:22:07 <portdirect> evrardjp: may need to be lunch - but we will find time
15:23:00 <evrardjp> portdirect: as long as you're bringing the picture of srwilkers I guess we'll have enough 'team' presence? ;p
15:23:26 <srwilkers> i have just the picture to send
15:23:29 <portdirect> evrardjp: i wish that were the case
15:23:37 <portdirect> a picture of srwilkers is one thing
15:23:44 <evrardjp> I have nothing more to add on the summit topic
15:23:47 <portdirect> but the reality is so much more
15:23:50 <srwilkers> can also facetime me in -- i can adjust my sleep schedule as needed
15:24:03 <evrardjp> haha
15:24:04 <portdirect> it would have been great to get anticw and lamt there as well
15:24:21 <evrardjp> indeed!
15:24:45 <portdirect> #topic Gating questions
15:25:06 <portdirect> your up evrardjp ;) sorry for pulling the trigger early last time
15:25:13 <evrardjp> thanks for giving me time on said questions
15:25:19 <evrardjp> no worries portdirect :)
15:25:43 <evrardjp> so two questions for you, to clarify something more that reporting the status, sorry for that.
15:25:56 <evrardjp> first, I'd like to talk about pre-run and expectations
15:25:59 <evrardjp> Are there any actions that needs to be taken in Zuul CI that contributors are expected to do before deploying OSH? Install docker? Update their systems? Configure specific networking on each host? Configure specific package source mirror?
15:26:01 <evrardjp> I think clarity would help on the pre-run steps, aasnd I am not sure the documentation states the expectations properly (or I misread it, which is possible...)
15:26:20 <evrardjp> as I am not sure*
15:27:09 <portdirect> evrardjp: it does not spell this out fully :(
15:27:15 <evrardjp> I'd also like to know the opinion of the community there: What is happening on their environment? Do you integrate more with your existing practices, or do you completely use from a clean system (ubuntu/centos/whatever)?
15:27:23 <portdirect> there was an effor to change this - though it dies on the vine
15:28:01 <evrardjp> portdirect: the idea behind those questions is reduce unnecessary code and clarify scope
15:28:21 <portdirect> evrardjp: i think most people are integrating with existing systems for dev work
15:28:53 <portdirect> and occasionly spinning up fully 'clean' envs when requried
15:29:11 <portdirect> i know this is how mattmceuen, srwilkers, lamt and myself are working at least
15:29:13 <jayahn> What is exaxtly integrating witj existing system? Something like running it on your laptop?
15:29:25 <portdirect> ^ this is how i read it
15:29:31 <portdirect> but we should clarify :)
15:30:34 <evrardjp> jayahn: like integrating with any existing infrastructure: ranges from a clean machine to  a machine with kubernetes installed , and a machine with docker installed is inbetweenn
15:31:01 <evrardjp> storage is another topic on "integration with any existing infrastructure"
15:31:38 <evrardjp> if people are assumed to have a ceph deployed or a nfs deployed it's not the same.
15:32:04 <jayahn> Okay got it. Will be back in 3min with answer. I am currently entering elevator at my apartment building
15:32:08 <evrardjp> Keep in mind I don't mean removing any current developer "convenience" scripts. I am just wondering what is convenience and what is scope.
15:32:43 <evrardjp> separating said domains will help on simplifying gating down the line.
15:33:11 <evrardjp> Anyway, I have my first set of questions asked to the community, I will wait for feedback next week, if you don't mind
15:33:13 <portdirect> i think its in scope to have a 'devstack' like exp for developers
15:33:26 <evrardjp> that sounds understandable
15:34:06 <evrardjp> portdirect: should we draft this goals/mission/delimiters somewhere in the docs?
15:34:33 <portdirect> evrardjp: i think the etherpad you started would be a great place to start this
15:34:47 <mattmceuen> agree
15:34:50 <evrardjp> (complementing https://docs.openstack.org/openstack-helm/latest/readme.html#mission basically)
15:35:00 <portdirect> I'll also reach out to rich wellum, we did a load of work there to make this clearer though i must confess we lost it :(
15:35:42 <evrardjp> thanks. Keep in mind the idea is just for clarifying gate scope. I want to avoid "repeating ourselves" and putting barriers/limits helps at my work.
15:36:12 <evrardjp> let's go for the second question there?
15:36:36 <portdirect> go for it
15:37:13 * portdirect notices rwellum is here - you still about?
15:37:23 <evrardjp> let's talk about post-run. It's long and I am not sure how to parse it, how useful it is. I'd be willing to simplify it, but I am a noob. So I am proposing something: Is it okay to have multiple levels of debug, and by default having minimum output?
15:37:48 <portdirect> evrardjp: its very-very useful
15:38:04 <portdirect> though it may capture too much in some places, and not enough in others
15:38:13 <mattmceuen> is the idea that in gating, it would have lots of debug, and in a dev env it would not evrardjp?
15:38:17 <evrardjp> I am thinking that people having failure could temporarily enable extra debugging on the post_run assuming we have defined that
15:38:18 <portdirect> a refactor would certainly be good
15:38:52 <mattmceuen> or just to leave big-debug for special occasions?
15:38:55 <evrardjp> mattmceuen: the only impact would be in openstack ci
15:39:32 <evrardjp> have a minimal debug like this (to be discussed/improved together): see comment in https://gist.github.com/evrardjp/a0f58e1304aaa78d3a8cacf7e5465492
15:40:10 <portdirect> evrardjp: i started a thing like that
15:40:12 <portdirect> 2 sec
15:40:46 <portdirect> https://review.openstack.org/#/c/550467/
15:41:00 <evrardjp> then the idea would be, should somewant want to increase logging, he can change the job definition in the repo to temporarily "increase logging"
15:41:02 <portdirect> may be worth updating?
15:41:22 <portdirect> is there a reason you are wanting to decrease logging in the gate?
15:42:22 <evrardjp> portdirect: yes, while we can increase post jobs timeouts, I am not so sure the resulting artifacts are useful for most of the runs
15:42:49 <evrardjp> so we are taking infra disk space, inodes, and cpu resources for nothing
15:43:03 <evrardjp> (because a lot of the things we gather are quite IO intensive)
15:43:19 <evrardjp> or it looks like it, as sometimes it takes minutes to get responses from k8s...
15:43:35 <portdirect> evrardjp: ok - just wanted to check you and i were thinking alike
15:43:50 <mattmceuen> if something fails in gate, and we don't have the logs to troubleshoot it, would there be a way to trigger the job again with extra debug logging?
15:44:11 <mattmceuen> or would it require a developer to ad-hoc run the gate locally?
15:44:12 <portdirect> i think we can probably cut about 50% of some of the things at a sweep - eg we get yaml and txt output for many objects
15:44:24 <portdirect> so we could prob just get yaml without loosing anything
15:44:28 <evrardjp> mattmceuen: I'd say not without a code change, but it could be possible with a code change
15:44:35 <mattmceuen> gotcha
15:44:54 <portdirect> it would just be a case of adding a conditional round a few things
15:45:11 <evrardjp> precisely
15:45:13 <portdirect> can we leverage zuul to increase longging on a retry automagicly?
15:45:25 <portdirect> *logging
15:45:39 <evrardjp> I doubt that, but I will check.
15:46:09 <jayahn> automagically...... Good word. :)
15:46:13 <evrardjp> portdirect: ok so I'd like we discuss this together: talking about things we can sweep.
15:46:58 <srwilkers> would it be sane to have experimental jobs that can be triggered that include the difference between what's offered by default and the more "verbose" option?  not sure if that makes sense or not, but if i were to be messing with something that really required me to dig deep enough to need everything, it seems it'd be helpful to just fire off an experimental check instead of having to wait for a potential failure
15:47:02 <srwilkers> but maybe that's just me
15:47:08 <evrardjp> Can we prepare this conversation for next week? (Goal: Explicitly listing what needs to be 'the minimum a developer expects from zuul' 'the maximum a developer could need from zuul'
15:47:48 <portdirect> evrardjp: sure - i think that would help here
15:47:51 <evrardjp> srwilkers: I thought of that too. I might have a mix of both in my ideas.
15:48:32 <evrardjp> but I'd rather speak about what needs to be done now, rather than how.
15:49:17 <evrardjp> thanks for listening to me, sorry if I took a long time in the meeting
15:49:56 * evrardjp giving the mic back to portdirect
15:49:57 <portdirect> evrardjp: you always bring the hard topics
15:50:06 <portdirect> and thats somthing you should never be sorry for
15:50:40 <portdirect> #topic roundtable & reviews
15:50:41 <evrardjp> I agree with you there, but it's still a community meeting, there are other people than me that deserve speak time
15:51:10 <portdirect> we are out of items on the Agenda
15:51:22 <portdirect> are there any other topics people would like to bring up
15:51:25 <evrardjp> woot open discussion!
15:52:06 <evrardjp> thanks for the reviews portdirect srwilkers and others!
15:52:13 <evrardjp> that's all I had to say.
15:52:15 <srwilkers> evrardjp: of course :)
15:52:25 <srwilkers> the early morning chats are always a pleasure
15:53:05 <portdirect> srwilkers: before 9am is too early for humans
15:53:13 <jayahn> ㅜ.ㅜ......
15:53:30 * srwilkers suddenly questions what he is
15:53:39 <portdirect> we all do srwilkers
15:53:50 <portdirect> and with that cheap insult
15:54:00 <portdirect> lets wrap up for today
15:54:08 <portdirect> #endmeeting