14:00:20 <mattmceuen> #startmeeting airship
14:00:21 <openstack> Meeting started Tue Aug 27 14:00:20 2019 UTC and is due to finish in 60 minutes.  The chair is mattmceuen. Information about MeetBot at http://wiki.debian.org/MeetBot.
14:00:22 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
14:00:26 <openstack> The meeting name has been set to 'airship'
14:00:27 <mattmceuen> #topic Rollcall
14:00:32 <CobHead> o/
14:00:35 <alexanderhughes> o/
14:00:35 <nishantkr> o/
14:00:38 <kskels> |o/
14:00:39 <howell> o/
14:00:41 <dukov> o/
14:00:41 <mattmceuen> ty for sharing agenda alexanderhughes
14:00:45 <ian-pittwood> o/
14:00:50 <mattmceuen> <alexanderhughes> Alexander Hughes hello!  meeting kicking off in a few minutes, agenda here https://etherpad.openstack.org/p/airship-meeting-2019-08-27
14:00:51 <seaneagan> o/
14:00:55 <sthussey> here
14:00:58 <mattmceuen> GM/GE all!
14:01:12 <alexanderhughes> yw :)
14:01:19 <mattmceuen> we'll get going in just a minute
14:01:33 <mattmceuen> feel free to add any more topics to discuss to this week's agenda
14:01:33 <CobHead> What time is it over there?
14:01:45 <michael-beaver> o/
14:01:49 <mattmceuen> 9am here in Saint Louis! How about you CobHead?
14:01:59 <aaronsheffield> o/
14:02:35 <CobHead> 4pm in Norway
14:03:37 <mattmceuen> thanks all for joining, and away we go:
14:03:50 <mattmceuen> #topic Airship YouTube playlist
14:03:55 <mattmceuen> go for it alexanderhughes
14:04:12 <alexanderhughes> Just a quick announcement, Chris and the rest of the awesome folks at OSF have set us up a playlist on their youtube channel
14:04:19 <alexanderhughes> #link https://www.youtube.com/playlist?list=PLKqaoAnDyfgp8YjZbzjVrmZBJR9thV27y
14:04:39 <alexanderhughes> these are a collection of interviews/presentations etc. regarding Airship.  if you have more content you'd like to see added please shoot me a message
14:04:41 <hogepodge> Sunny and the rest of the marketing team deserve most of the credit.
14:04:56 <mattmceuen> that is awesome
14:05:02 <CobHead> Nice :)
14:05:08 <alexanderhughes> yes, huge thanks to Sunny for her hard work on the media channels
14:05:29 <mattmceuen> ty both for spearheading that (and Sunny for doing the hard work!! :)
14:05:30 <hogepodge> It became a valuable exercise for all of the projects so thank you for the suggestion.
14:05:38 <mattmceuen> nice
14:06:28 <mattmceuen> hopefully we get a few more on there after the Shanghai summit as well
14:06:48 <CobHead> Any presentations due for the summit?
14:07:04 <sthussey> May be worthwhile to have someone demo some of the projects there
14:07:12 <mattmceuen> ++++ sthussey
14:07:13 <sthussey> rather than having only presentations and interviews
14:07:30 <CobHead> Demos are always nice
14:07:57 <alexanderhughes> absolutely, there's a lot more we can do with this.  I'd love to see more content on the channel, making it a one stop shop for anyone curious about airship and use cases, how certain things work etc
14:07:58 <mattmceuen> alexanderhughes, were you already prepping a Shanghai presentation list or did I make that up?
14:08:32 <alexanderhughes> #link https://www.airshipit.org/blog/airship-at-the-open-infrastructure-summit-shanghai-in-november.html
14:08:39 <alexanderhughes> there's a list of Airship presentations on the blog
14:09:04 <roman_g> o/
14:09:17 <mattmceuen> ah, ty for sharing that
14:09:51 <mattmceuen> Alright, I think we can move on unless anything else on this topic?
14:10:00 <alexanderhughes> nope :)
14:10:16 <mattmceuen> #topic Airship GUI Project Name
14:10:27 <mattmceuen> So last week we agreed to create a new project
14:10:38 <mattmceuen> But we didn't figure out anything awesome to call it
14:10:45 <mattmceuen> I sent out an email on the ML to remedy that
14:11:10 <mattmceuen> Details in that email - please read it and plan to share your opinions on the best names in next week's IRC meeting
14:11:19 <mattmceuen> That's all on that topic unless any questions
14:12:26 <mattmceuen> #topic New Project Proposal:  airship/isogen
14:12:56 <mattmceuen> Dmitry: the floor is yours, can you please give us an overview of the plan for this project?
14:13:07 <dukov> Thanks
14:13:25 <mattmceuen> I know discussion has been going on in the Bootstrap SIG, so folks who are not regulars may need to be caught up just a bit
14:14:11 <dukov> Basically as a part of airshipctl development we need Live CD iso to bootstrap ephemeral k8s cluster
14:14:42 <dukov> this cluster is going ot be used for new (Target) cluster deployment using metal3
14:15:01 <dukov> ephemeral cluster get destroyed then
14:15:38 <dukov> so to build an LiveCD/LiveUSB image it was decided to use container with set of scripts
14:16:04 <dukov> so there is a question: where to keep Dockerfile and scripts
14:16:50 <dukov> we also decided to create separate repo for all such image builders (e.g for Debian SuSe etc)
14:17:11 <mattmceuen> Makes sense to me.  So this new project will not contain the airshipctl module for generating the iso; that code would live in the airshipctl project, right?
14:17:12 <jemangs> o/
14:17:15 <dukov> it turned out that we may need more the one type of images
14:17:18 <mattmceuen> o/ jemangs
14:17:43 <roman_g> I support to have another repo, but for all images/Dockerfiles all together, which would be supplementary to the airshipctl.
14:17:45 <dukov> so may be isogen is not good name for that repo
14:18:11 <roman_g> airshipctl-images?
14:18:26 <dukov> +1 roman_g
14:18:35 <pramchan> youe ,ean not using quay.io?
14:18:38 <CobHead> Wouldn't it make more sense to keep it all time builders in the same repo? If the project ends up with builders for 6 distros, it might be difficult to keep track of them all?
14:18:57 <roman_g> similar to openstack-helm-images
14:19:46 <mattmceuen> CobHead: there's overhead involved in project creation/maintenance, so if it doesn't get out of hand I like starting with a single repo.  Agree roman_g openstack-helm-images is some good prior art there
14:20:07 <pramchan> can we not add these standard specified images to quay.io
14:20:12 <roman_g> CobHead: yes, that's what I would like to have - one repo for all images related to the airshipctl
14:20:26 <mattmceuen> pramchan:  yep we can store build images in quay.io, agree
14:20:32 <roman_g> pramchan: they would be published there, but we discuss where to host sources
14:20:34 <mattmceuen> we just need a project for the unbuilt images
14:20:52 <pramchan> OK that makes sense
14:21:22 <pramchan> for that we can create patches under tools and add those sources
14:21:22 <CobHead> A question for a later date is whether or not the Airship project team should be responsible for all the builders, or if some should be sourced out to the community?
14:22:34 <pramchan> we need to maintain at-least two builds to ensure variety
14:23:40 <mattmceuen> CobHead:  agree and I would frame it up straw-man style this way, an airship-images project would be only for image sources that need to be created or built by the airship team (potentially forked from a project) only when necessary, and would not take image build responsibility away from other projects that are a more natural fit (e.g. deckhand would continue to build its own image)
14:23:43 <mattmceuen> Agree?
14:24:12 <CobHead> Sounds good.
14:24:17 <jamesgu> +1
14:24:47 <roman_g> we could welcome other container images as long as they have high quality code, well covered by CI tests, and there would be someone willing to support them
14:24:52 <pramchan> +1
14:25:05 <roman_g> _integrated_ CI tests
14:25:11 <mattmceuen> Agree roman_g
14:25:46 <mattmceuen> do we like the airship-images convention?  Or do we want to come up with something more creative?
14:25:54 <mattmceuen> (in terms of project name I mean)
14:26:03 <CobHead> Keep it nice and simple
14:26:20 <howell> +1
14:26:24 <roman_g> airship-images or airshipctl-images?
14:26:24 <pramchan> +1
14:26:25 <alexanderhughes> descriptive is nice.  there's no question what airship-images is
14:26:49 <sthussey> why wouldn't this go into airshipctl if this is only to support code being added there?
14:26:49 <pramchan> airship-images as ui will also use it
14:26:52 <nishantkr> airship-images +1
14:27:06 <jamesgu> +1 to airship-images
14:27:12 <mattmceuen> I'd go for airship-images roman_g, I think it makes sense for other non-cli images we may need in the future
14:27:43 <pramchan> looks majority is airship-imaged go for it
14:27:58 <pramchan> airship-images typo
14:28:05 <mattmceuen> I think you are right pramchan, I will sip coffee for 30s in case anyone comes up with a counterpoint
14:28:21 <michael-beaver> +1 sounds good to me
14:28:26 <pramchan> welcome to colambian cofee!!!
14:28:36 <roman_g> sthussey: we were discussing on bootstrap call, that there are high chances that we would have more airshipctl modules implemented as containers.
14:28:47 <sthussey> right
14:29:10 <roman_g> sthussey: so I see reasons to keep it with airshipctl
14:29:11 <CobHead> modules as containers - I like it :)
14:29:13 <sthussey> Is the code in the airshipctl module going to be coupled with the container contents?
14:29:13 <dukov> and not only airshipctl i guess ...
14:29:33 <mattmceuen> nope the airshipctl module code will live in airshipctl I believe
14:29:40 <mattmceuen> for isogen, etc
14:29:47 <sthussey> coupled as in co-dependent
14:29:52 <mattmceuen> where etc = built-in modules
14:29:54 <pramchan> I think given the trends it's a no brainer and best practice by definition will be use Docker conatainer or some container that fits your needs
14:30:21 <sthussey> a change in module code requires a change in the container build - or vice versa
14:30:44 <roman_g> sthussey: line 65 and onwards https://etherpad.openstack.org/p/Airship_bootstrap
14:30:48 <mattmceuen> what do you think dukov in terms of coupling?  Do you expect a stable interface?
14:32:01 <sthussey> really just an ecosystem difference. OSF projects seem to shy away from a repo having more than a single artifact produced from it
14:32:14 <roman_g> probably not super stable. but intention was to use Depends-On: xxx for cross-repo CI checks
14:32:27 <roman_g> and to use cross-CI Zuul jobs
14:32:34 <dukov> mattmceuen: well the idea is to define a contract of communication between container and airshipctl which means any image that implements interface can be used
14:32:39 <roman_g> *cross-repo CI Zuul jobs
14:33:11 <dukov> that being said we still need a place where to keep couple default ISO builder container
14:33:30 <dukov> *containers
14:33:30 <mattmceuen> makes sense dukov, it would have to be pretty stable or else it would wreak some havoc
14:34:15 <sthussey> just seems odd to me that images built for this and the airshipui are considered coupled enough to be in the same repo, but putting this in the same repo as airshipctl isn't. anyway, carry-on.
14:34:24 <dukov> technically  we can use airshipctl for that but eventually it will create a mess
14:35:03 <mattmceuen> airshipui - I think container build code would live in-tree with application code, but I haven't been tied into those convos and aren't an authority
14:35:53 <mattmceuen> Ok, I think we can move on unless anything else on this topic:
14:36:14 <roman_g> sthussey: are you for separating supplementary images into one additional repo, or are you for keeping supplementary images (e.g. isogen) inside airshipctl?
14:36:47 <dukov> so what is the conclusion? are we going to create separate repo ?
14:37:19 <sthussey> I was just bringing up a viewpoint. It sounds like the folks working on it are for this generic images repo
14:37:39 <sthussey> I tend to avoid repo sprawl
14:37:45 <sthussey> but airship is probably long past that point
14:38:23 <mattmceuen> Yep I think we still have consensus for airship-images (unless anyone de-consensizes).  Appreciate probing the pros/cons of this sthussey
14:38:31 <mattmceuen> Agree on repo sprawl 100%
14:38:56 <mattmceuen> I was leaning toward bundling this into airshipctl till learning more from dukov
14:39:12 <mattmceuen> thanks for bringing this up and filling us in on progress dukov!
14:39:58 <dukov> mattmceuen: you are welcome! I'll update CR accordingly then
14:40:18 <pramchan> if we run into unanticipated issues we can always move it out later
14:40:27 <mattmceuen> Perfect ty - then feel free to add that on to the agenda review list as well so we can put +1s on it
14:40:42 <mattmceuen> #topic Has anyone been able to roll AIAB in the recent weeks
14:40:52 <mattmceuen> CobHead I take it you're hitting some issues :(
14:40:54 <CobHead> Yes, I can't for the life of me get past shipyard trying to build Armada. It's stuck there until it finally fails. Then it moves on to try submitting Horizon, but that inevitably fails too. Anyone else in here experienced this?
14:41:26 <CobHead> I initially suspected a timeout, so I tried bumping it, but to no avail.
14:41:36 <mattmceuen> And just to clarify this is Treasuremap single-node AIAB, right?
14:41:41 <CobHead> Yes.
14:42:03 <mattmceuen> That's what I thought - yeah, I tried a couple times a week and a half or so ago with two successes
14:42:15 <mattmceuen> We need to figure out your RC
14:42:36 <mattmceuen> Remind me, are you running in a local VM or in a public cloud or?
14:42:48 <CobHead> Local VM - however I've tried public cloud aswell.
14:42:56 <mattmceuen> same behavior?
14:43:13 <CobHead> Actually, it's different. Can't get past PostgreSQL in the cloud.
14:43:21 <CobHead> Locally it stops at Armada build.
14:43:53 <mattmceuen> Well it oughta be working fine in both places.  Which one do you want to fix first
14:44:02 <CobHead> There feels like there are some caveats here, and I'd like to resolve them on the top, so people who are unfamiliar with the project don't hit a dead end when they want to try AIAB.
14:44:23 <CobHead> I'd like to fix local first.
14:44:25 <mattmceuen> Agree, may be some hidden assumption we need to ferret out
14:44:35 <mattmceuen> Ok - would you be free to power through this in the IRC channel after the meeting?
14:44:37 <pramchan> does AIAB stand for Airship in a Box ? or is it Airskiff my ignorance please clarify?
14:44:47 <CobHead> Airship in a Box ;)
14:44:48 <mattmceuen> Close pramchan :)  Bottle
14:44:57 <pramchan> ok
14:45:00 <CobHead> mattmceuen: Absolutely. Can start deployment now.
14:45:01 <mattmceuen> again with the nautical theme
14:45:54 <CobHead> After we have gone through the changes for review, there is another thing I'd like to take up.
14:46:08 <pramchan> Is it then etcd dependency that is fainling in local VM similar to PostgresSQL in cloud for AIAB?
14:46:29 <CobHead> At this point, I have not a single clue.
14:46:42 <pramchan> Can we not track the failure or error code?
14:46:49 <mattmceuen> Since we only have a few mins left, let's save the troubleshooting for post-call -- but please stay tuned pramchan I want to get this fixed
14:46:57 <mattmceuen> #topic Requests for Review
14:47:04 <mattmceuen> These are copied from last week's; confirm whether more review is needed before meeting:
14:47:04 <mattmceuen> https://review.opendev.org/#/c/671575/ - Shipyard (one more +2 is needed)rt dependencies)
14:47:04 <mattmceuen> https://review.opendev.org/#/c/675440 (add airskiff suse cite and ci gate)
14:47:04 <mattmceuen> https://review.opendev.org/#/c/672688/ (Tungsten Fabric support in Treasuremap)
14:47:04 <mattmceuen> https://review.opendev.org/#/c/678311/ Airskiff dep fix in Treasuremap
14:47:05 <mattmceuen> https://review.opendev.org/#/c/671575/ Shipyard
14:47:05 <mattmceuen> https://review.opendev.org/#/c/677980/ Bump ucp keystone timeour
14:47:06 <mattmceuen> https://review.opendev.org/#/c/674409/ - shellcheck linting (this one - for divingbell) - need to evaluate if it's usefull to be implemented to other charts we host under airship/*.
14:47:06 <mattmceuen> https://review.opendev.org/#/c/678618/ - airship/armada - Fix: Armada Exceptions docs rendering on RTD
14:47:19 <mattmceuen> Let's please get some eyeballs on these today, team
14:47:34 <CobHead> Another thing, mattmceuen - which everyone should hear.
14:47:38 <CobHead> Or read, for that matter..
14:47:46 <mattmceuen> Yes please CobHead, go for it
14:48:07 <mattmceuen> #topic Roundtable
14:48:19 <CobHead> https://review.opendev.org/#/c/672540/ Take a look at this commit. This commit removed the package dependencies for Airskiff, since they are being handled in OSH - but the OSH package install script is never handled by the airskiff script(s)
14:48:40 <mattmceuen> ahh
14:48:48 <CobHead> So essentially, if you deploy Airskiff using the steps described in the docs - you will halt when Armada is built
14:48:55 <CobHead> Since that requires make.
14:48:57 <CobHead> Make is never installed.
14:49:04 <CobHead> Likewise with the Python client for Openstack
14:49:10 <CobHead> Since python-pip is never installed.
14:49:14 <CobHead> https://review.opendev.org/#/c/678302/
14:49:20 <CobHead> I've made some amends in that commit.
14:49:24 <mattmceuen> Yep - I hit the Python / pip issue yesterday
14:49:59 <CobHead> But please - if there are dependencies removed - please also remove them at the gate. Zuul has no issues with this as those packages are listed in the gate.
14:50:19 <CobHead> But not in the script - so the CI succeeds, while the developer fails.
14:50:26 <mattmceuen> Yeah
14:50:47 <mattmceuen> Good catch, thanks for fixing that
14:50:50 <evgenyl> https://review.opendev.org/#/c/678311/ I think this should help with this.
14:51:18 <evgenyl> Gate Ubuntu image has a lot of stuff pre-installed, this is the reason it does not fail.
14:51:30 <CobHead> Including make, I presume.
14:52:05 <CobHead> I think evgenyl's patch should be merged instead of mine, as it addresses both pip and the other dependencies.
14:52:07 <CobHead> I
14:52:11 <CobHead> I will abandon mine ;)
14:52:35 <mattmceuen> ok :)  great minds are thinking alike
14:53:22 <CobHead> That was what I had to bring up :)
14:53:58 <mattmceuen> I think we hit everything in the agenda, anything esle you'd like to discuss today team?
14:54:26 <mattmceuen> I will be on vacation tomorrow through next Thursday, so don't expect much out of me :D
14:54:51 <pramchan> Can i ask a question on baremetal?
14:54:56 <mattmceuen> sure pramchan
14:55:25 <roman_g> sthussey: you wanted to review https://review.opendev.org/#/c/674409/ - shellcheck linting. I would appreciate a lot your opinion.
14:55:34 <roman_g> Thank you.
14:56:03 <pramchan> I am trying to find a directory where I can submit the code, shall I call it baremetal as a new folder to start coding for Redfish boot for testing vBMS?
14:56:18 <pramchan> vBMC
14:56:47 <roman_g> start anywhere, and we will find a right place :)
14:57:08 <mattmceuen> pramchan: this code will be the Redfish equivalent of vBMC, right?
14:57:17 <pramchan> cool thx will do
14:57:54 <roman_g> airship/airshipctl/tools/ <- probably somewhere here
14:57:56 <pramchan> partly yes will push to cmd later once I can test it
14:58:14 <roman_g> as it would be part of gates, I think
14:58:16 <pramchan> yes tools is where I will put the vBMC
14:58:21 <mattmceuen> Agree with roman_g, get it out somewhere public and we can figure it out.  You can always make changes in a new github repo and then we can import it later when we know where it should go - we have done that many times
14:58:38 <mattmceuen> that's awesome
14:58:53 <pramchan> thanks and I will try get it out the base by week end
14:59:11 <mattmceuen> perfect - drop us a note in the IRC channel when its ready for a look!
14:59:25 <pramchan> OK
14:59:29 <mattmceuen> ok I think we're out of time - great meeting all, appreciate your time & discussion
14:59:34 <mattmceuen> have a good week!
14:59:43 <mattmceuen> #endmeeting