15:00:19 <mattmceuen> #startmeeting openstack-helm
15:00:20 <openstack> Meeting started Tue Jun  5 15:00:19 2018 UTC and is due to finish in 60 minutes.  The chair is mattmceuen. Information about MeetBot at http://wiki.debian.org/MeetBot.
15:00:21 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
15:00:23 <openstack> The meeting name has been set to 'openstack_helm'
15:00:25 <mattmceuen> #topic rollcall
15:00:29 <srwilkers> o/
15:00:34 <mattmceuen> GM srwilkers!
15:00:39 <gmmaha> o/
15:00:56 <srwilkers> hey mattmceuen :D
15:00:56 <mattmceuen> https://etherpad.openstack.org/p/openstack-helm-meeting-2018-06-05
15:01:04 <alanmeadows> o/
15:01:13 <roman_g> o/
15:01:14 <mattmceuen> Here's our agenda - please add all the burning talking points that have been keeping you up at night
15:01:27 <lamt> o/
15:01:34 <srwilkers> thats a long list
15:01:53 <mattmceuen> please prioritize them to talk about the OSH things first
15:02:06 <mattmceuen> if we move fast we can work our way down to spicy mexican food
15:02:50 <mattmceuen> First thing:
15:02:56 <mattmceuen> #topic Storyboard Friday
15:03:01 <mattmceuen> Storyboard Friday.
15:03:27 <mattmceuen> Not sure of the exact time  but we'll keep the migration communicated in the channel
15:03:45 <mattmceuen> I suggest to be safe, let's just stay out of Launchpad on Friday
15:03:57 <rwellum> Hopefully smooth sailing... o/
15:04:06 <srwilkers> hey rwellum o/
15:04:33 <rwellum> hi srwilkers
15:04:50 <mattmceuen> As far as our daily lives post-storyboard
15:05:10 <mattmceuen> 1) use Storyboard for all new scope and bugs and specs going forward
15:05:17 <mattmceuen> there are no more blueprints, no more bugs
15:05:22 <mattmceuen> Everything is a Story (hence the name)
15:05:40 <portdirect> specs?
15:05:55 <mattmceuen> Yep specs should be captured as stories
15:05:59 <portdirect> I need to read up - but would we no longer be doing them in gerrit?
15:06:02 <mattmceuen> We should find an example of a story spec
15:06:06 <mattmceuen> Specs in gerrit
15:06:21 <mattmceuen> Just no blueprint in launchpad any more - a story in storyboard instead
15:06:49 <alanmeadows> are folks generally creating a story linking to the gerrit spec for full storyboard awareness?
15:07:02 <alanmeadows> I am unaware of the sb customs
15:07:33 <mattmceuen> 2)  For linking PS to Stories (in general) here is a quick example from Kendall
15:07:33 <mattmceuen> As for instructions on git commit headers its pretty simple. You link both the task and the story in the commit. You should know this from Upstream Institute :wink: It ends up looking like this:
15:07:33 <mattmceuen> -----------------
15:07:33 <mattmceuen> Commit message summary line stuff
15:07:33 <mattmceuen> Commit message body info. All the details. Blah blah blah.
15:07:33 <mattmceuen> Story: 882882
15:07:33 <mattmceuen> Task: 2221
15:07:34 <mattmceuen> Change-Id: djk3kj24jk2kj21kj1kj25009sdfk
15:07:34 <mattmceuen> ----------------
15:08:05 <mattmceuen> rwellum - have you seen any good example / prior art for spect-storyboard linking?
15:09:14 <rwellum> Looking at ironic right now to pull one out
15:09:20 <mattmceuen> Awesome thanks
15:09:41 <pcarver_> I haven't been following storyboard development. Did they standardize on a way of showing release content?
15:09:44 <mattmceuen> Also thanks to rwellum for leading this effort
15:10:29 <jayahn> o/
15:10:31 <mattmceuen> Release content in what sense - like release notes?
15:10:35 <mattmceuen> hey jayahn!
15:10:39 <jayahn> hey
15:10:53 <pcarver> I mean the equivalent to Launchpad's "Series and milestones" pages
15:11:10 <roman_g> q: would we change URL in #openstack-helm topic from Launchpad to Storyboard?
15:11:18 <srwilkers> roman_g: i would imagine so
15:11:24 <pcarver> Where you can see what bugs and blueprints are planned for any future milestone or were contained in any past milestone
15:12:17 <rwellum> BTW airship is already using storyboard: https://storyboard.openstack.org/#!/project_group/85
15:12:46 <mattmceuen> Lucky them, they started there so don't have to do a migration :)
15:13:10 <mattmceuen> Note I'm planning to shamelessly model after the links on the ironic wiki
15:13:10 <mattmceuen> https://wiki.openstack.org/wiki/Ironic
15:13:28 <mattmceuen> re: bug tracker / feature tracker
15:13:43 <alanmeadows> airship is already hoving over storyboard
15:13:44 <mattmceuen> specifying the release is not jumping out to me pcarver
15:13:58 <mattmceuen> alanmeadows is here all day, folks
15:14:13 <pcarver> Hmm, looking at OSH's Launchpad it doesn't look like you've been tracking any releases yet anyway
15:14:18 <portdirect> nope
15:14:25 <portdirect> as we dont yet have one
15:14:37 <portdirect> though probably should have bee tracking toward it
15:14:42 <portdirect> *been
15:15:34 <mattmceuen> Yeah, we just have a spec for 1.0, and I'm going to translate remaining work items into stories.  I'll release teg them in the right way if we can figure out what that looks like.
15:15:42 <pcarver> Neutron is what I usually look at as the ideal example https://launchpad.net/neutron/+series
15:16:12 <portdirect> pcarver: you know that launchpad is going away?
15:16:25 <pcarver> You can see every milestone all the way back to the Diablo series and click on any one to see what was in it
15:16:38 <rwellum> Good example of ironic, plus tags: https://storyboard.openstack.org/#!/story/2002064
15:16:58 <pcarver> portdirect: I'm aware of it, but when I asked the Storyboard folks about this functionality  8-12 months ago it wasn't even on their radar
15:17:10 <portdirect> roger
15:17:16 <mattmceuen> aha
15:17:50 <mattmceuen> does subject-prefixing stories-for-specs with "[RFE]" sound good to y'all?
15:17:57 <pcarver> I'm not sure if it's because they didn't/don't care about recording history or if they just weren't familiar with Launchpad's interface
15:18:09 <rwellum> This amongst other reasons is why the transition hasn't happened for all. But we reasoned that being fairly new, for OSH it's better now than later.
15:18:19 <pcarver> They kind of suggested that every project would just invent its own practices
15:18:32 <rwellum> pcarver: yeah same message I got at the summit.
15:18:33 <pcarver> Which sounded like an awful idea to me
15:18:45 <portdirect> we work with what we have :)
15:18:50 <mattmceuen> +1 :)
15:19:04 <mattmceuen> aka "do what Ironic does till we see something we don't like"
15:19:30 <mattmceuen> Alright - any other Storyboard items before we tie this one up and breathlessly wait till friday
15:19:30 <portdirect> shouldn't be to physically painful
15:19:30 <pcarver> But if everybody is on their own to establish and follow practices, then it would be at least a good idea to figure out what those are
15:19:54 <rwellum> pcarver: think it's worthwhile firing off an email to openstack-dev?
15:20:53 <pcarver> rwellum: Possibly. I haven't been following storyboard migrations, so maybe some teams have already figured out standard practices.
15:20:53 <mattmceuen> would be helpful for teams to have all best practices summarized in one place, even if it has the caveat to "do whatever you want"
15:22:08 <mattmceuen> Alright we'd better keep rolling.  Let's stay in sync on this in the #openstack-helm channel
15:22:14 <mattmceuen> #topic Review OSH troubleshooting brainstorming
15:22:21 <pcarver> The main question for any give team to answer (e.g. OSH), is whether the team sees value in maintaining a history of what the series and milestones were and what was in each
15:22:52 <mattmceuen> https://etherpad.openstack.org/p/openstack-helm-troubleshooting
15:23:00 <jayahn> i do think there is value in maintinaing history
15:23:17 <portdirect> I really see value in release notes
15:23:18 <mattmceuen> Agree - we should start off on the right foot
15:23:29 <portdirect> as meeting part of that requirements
15:23:32 <mattmceuen> +1
15:24:12 <mattmceuen> We have a reno PS ready to merge in, but we're waiting on having our own release notes best practices documented before pushing that out
15:24:17 <rwellum> For the trouble shooting, I filled out of a lot of this, which is 'ironic' because I have little osh experience to date - most of this was from my notes on kolla-k8s
15:24:28 <portdirect> how much is osh specific?
15:24:35 <portdirect> and how much is general k8s?
15:24:59 <rwellum> So no hurt feelings if we cut stuff.
15:25:05 <rwellum> 99% kubernetes
15:25:25 <rwellum> But  added a section: what to do if something goes wrong, that I would expect to be more osh related of course.
15:25:38 <mattmceuen> +1000
15:26:01 <mattmceuen> I really like that section at the end
15:26:06 <alanmeadows> I would take this opportunity to say long-term whats needed is a crowsnest OSH chart
15:26:19 <alanmeadows> something to look at data from k8s, cni, prometheus, OS sevices, and so on
15:26:22 <mattmceuen> Because it will save us a lot of grief :D even if that is k8s specific, it's a quick list of what info to gather
15:26:25 <jayahn> crownset?
15:26:27 <alanmeadows> and provide simple checklists for end users
15:26:51 <alanmeadows> most of this checking can be mechanized
15:27:06 <alanmeadows> "You have the following pods in a CrashLoop: ..."
15:27:32 <alanmeadows> @jayahn: having fun with nautical words.. a simple dashboard
15:27:32 <mattmceuen> https://goo.gl/images/fZFwuG
15:27:46 <jayahn> ah..
15:27:46 <rwellum> +1 alanmeadows - that's what I was getting at. because you can filter out so many questions this way.
15:27:53 <tdoc> Being a new user, I think most of the stuff in there is useful, by now most of it I had figured out, but it would surely have helped to have from the start.
15:28:24 <alanmeadows> this would be the first non-helm development effort of OSH, but it would be tremendous
15:28:33 <mattmceuen> Since we're in brainstorming mode, it's better to err toward putting stuff in that etherpad now
15:28:55 <mattmceuen> The trick will be curating that down to OSH documentation without getting too far into the business of hosting k8s-general docs
15:29:11 <mattmceuen> Maybe some really key concepts, and then link out to k8s docs for more
15:29:24 <rwellum> +1
15:30:00 <mattmceuen> I think we should let the troubleshooting guide material keep maturing for another week or two - just wanted to remind people it's out there and to please review / add to it
15:30:16 <jayahn> +1
15:30:19 <tdoc> +1
15:30:34 <mattmceuen> If alanmeadows wants to turn it into a design doc for a new crowsnest, well that's one of those four opens :)
15:30:34 <mattmceuen> https://etherpad.openstack.org/p/openstack-helm-troubleshooting
15:31:09 <mattmceuen> thanks to rwellum, roman_g, and anyone else who's contributed so far (can't see everyone's names)
15:31:22 <alanmeadows> We threw around the idea of simply extending cockpit, I can look into it
15:31:28 <alanmeadows> full 360 view
15:31:40 <mattmceuen> pete has an item on the agenda, but he's in a side convo so we'll come back to him
15:31:47 <mattmceuen> #topic Elasticsearch S3 snapshot repos via radosgw s3 api
15:31:54 <mattmceuen> @srwilkers take it away
15:32:10 <rwellum> portdirect: can you take a look (or recomend another experienced user) at the 'what to do when something goes wrong' section, with a strong OSH bias please?
15:32:26 <srwilkers> so this is something we talked about quite some time ago, when i first introduced curator to the elasticsearch chart
15:32:57 <srwilkers> originally there was a PVC in the chart that relied on a readwritemany provisioner, that could be used to create filesystem snapshot repositories for elasticsearch
15:33:38 <srwilkers> i think we all agreed it wasnt the best mechanism for providing this functionality, and wanted to use the s3 api for radosgw instead, since elasticsearch includes s3 snapshot repository functionality
15:33:39 <srwilkers> https://review.openstack.org/559417
15:34:08 <srwilkers> there's a few things im working on adding to bring this to the finish line, but i've got it to a point that it's functional
15:34:42 <srwilkers> im hoping to get the loose ends tied up before the end of the week
15:35:02 <mattmceuen> that's awesome
15:35:22 * portdirect is back
15:35:22 <srwilkers> the last bit im working on adding is the ability to create an arbitrary number of snapshot repositories, which can be useful for when curator needs to manage multiple indexes
15:35:24 <mattmceuen> Once we have the Swift chart in, that will be a secondary s3 mechanism for backups
15:35:31 <mattmceuen> *snapshots
15:35:44 <portdirect> does swift support s3?
15:36:01 <srwilkers> so we can have a repository for standard log snapshots, a repository for things we may want to keep for extended periods of time, or whatever else your heart desires
15:36:14 <mattmceuen> I haven't tried it, but read that it does
15:37:14 <srwilkers> anyway, that's all ive got here.  can expect some updated docs highlighting the current state of the lma union along with this when its ready
15:37:16 <mattmceuen> my heart desires lots of logs
15:37:34 <mattmceuen> Excellent - thanks for the update srwilkers
15:37:52 <mattmceuen> portdirect:  you're up
15:37:55 <rwellum> I think you mean 'your heart will go on'..
15:37:56 <mattmceuen> #topic Helm-Toolkit (portdirect)
15:38:06 <mattmceuen> don't make me start singing rwellum, I'll do it
15:38:27 * srwilkers grabs the pennywhistle
15:38:31 * rwellum covers ears
15:39:08 * jayahn think he needs to do something as well
15:40:03 <mattmceuen> We will come back to portdirect again :)
15:40:07 <mattmceuen> #topic Production vs Bare Bones values.yaml
15:40:10 <mattmceuen> alanmeadows!
15:41:19 <jayahn> don't understand exactly what it means
15:41:25 <alanmeadows> sorry dragged into a call
15:41:34 <mattmceuen> want to hold off on this topic?
15:41:51 <alanmeadows> sure, its been discussed before but I want to bring it to a wider audience
15:42:12 <alanmeadows> we can revisit next chat
15:42:12 <mattmceuen> ok - we will table for now.  jayahn a teaser of the topic (bullet points) is in the agenda
15:42:28 <mattmceuen> srwilkers:
15:42:30 <mattmceuen> #topic OpenStack Exporter
15:43:39 <srwilkers> well, this one couldve used some input from alanmeadows and portdirect but i'll throw it out regardless.  it was pretty clear to me in vancouver that there's a real need for a solution for monitoring an openstack control plane via prometheus
15:44:09 <srwilkers> monasca even goes as far as to mimic prometheus's functionality by providing mechanisms for scraping exporter endpoints
15:44:15 <portdirect> and AT&T has a nice one in att-comdev?
15:44:23 <srwilkers> bingo bango
15:44:35 <portdirect> so - lets move it into -infra
15:44:44 <srwilkers> portdirect: i think that's a great idea
15:44:49 <mattmceuen> +1
15:45:03 <portdirect> srwilkers: its only great as it was yours
15:45:14 <portdirect> i just stole your punchline
15:45:26 <srwilkers> portdirect: youre making me blush
15:45:55 <mattmceuen> srwilkers can you please collaborate w/ the author to get it into OS infra?
15:46:02 <srwilkers> mattmceuen: sure can
15:46:15 <mattmceuen> great
15:46:45 <mattmceuen> alright we have portdirect --
15:46:48 <mattmceuen> #topic Moving of charts to their correct locations
15:47:18 <portdirect> so - we really need to get the charts in the right place
15:48:15 <portdirect> it would be nice to use this an a 1st try at storybarding things?
15:48:33 <portdirect> and also for newer devs to get used to the gates etc?
15:49:17 <mattmceuen> Agree.  The work is mentioned in the 1.0 spec and should be a good way to get involved
15:49:23 <mattmceuen> in the code proper
15:49:33 <portdirect> it would be fairly simple for me to do a mega ps that moved them all - but think this would be a great opportinity to get some more people working on the gates :D
15:49:57 <portdirect> rwellum: could you help me with the storyboard side of things
15:50:09 <portdirect> I'll have a stab at it, but expect I'll need some help
15:50:23 <rwellum> Sure portdirect
15:50:31 <portdirect> then once we have that people could grab to story for each chart?
15:50:36 <mattmceuen> Yeah - let's create per-chart stories for this that account for gating, and then divide & conquer
15:50:44 <mattmceuen> +1
15:50:47 <portdirect> nice
15:51:04 <pcarver> portdirect: storybarding sounds like a nice variation. Matt can sing along with the stories.
15:51:19 <mattmceuen> They will be ballads
15:51:21 <portdirect> its the typo that keeps on giving
15:51:36 <mattmceuen> we are a heroic team are we not
15:51:55 <portdirect> my ambition is for us to become the plumbers of openstack
15:52:13 <mattmceuen> We have 9mins and three items on agenda to go -- portdirect want to tackle helm-tk?
15:52:17 <portdirect> ok
15:52:23 <mattmceuen> #Topic Helm Toolkit
15:52:44 <portdirect> so I'm gonna try and get most of the functions in helm toolkit using maps(dicts) as inputs
15:52:51 <portdirect> rather than the tuples we have today
15:53:04 <portdirect> as this will help with a couple of things:
15:53:09 <portdirect> 1) documentation
15:53:21 <portdirect> 2) allowing greater flexibilitiy
15:53:26 <portdirect> 3) wait for it
15:53:34 <portdirect> ... proper unit tests
15:54:04 <portdirect> so - as we do this i think its essential we start unit testing the helm-toolkit lib
15:54:05 <mattmceuen> That would be a much nicer interface to htk and also UNIT TESTS!  Are you thinking we should refactoring the dependent chart code as we go, or adapt the old interface htk interface around the new?
15:54:18 <rwellum> for item in dict.iteritems(): .....
15:54:24 <portdirect> refactor as we go
15:54:32 <portdirect> as we now have the gates for it :D
15:54:49 <rwellum> There did it for you portdirect ;) jk obviously.
15:55:15 <portdirect> I'll also work on airship and and other projects (ceph/tungsten fabric) to help them with the shift
15:55:32 <portdirect> but overall this will put us in a much stronger place
15:55:44 <srwilkers> +1
15:55:49 <portdirect> thats all i got there
15:56:01 <mattmceuen> Awesome.   Let's look for ways to spread this work out as well, since there will be a lot of it, and with some proper examples it shouldn't be too bad
15:56:17 <mattmceuen> #topic     multi-node deployer feedback
15:56:22 <mattmceuen> #rwellum go for it!
15:56:37 <rwellum> At the summit I spoke to 2 people about OSH, who like me, they had all successfully deployed a single OSH node, but got stuck with multiple-nodes. They were following the guide - but with some confusion over the fact - is it a guide or is it a bunch of gate script(s). I recall this was a somewhat fatal error we made in kolla-k8s - whenever someone claimed is didn't work we'd point to a green Gate run and say 'it
15:56:37 <rwellum> must be your fault'. Problem was the gate scripts were so utterly tweaked for that specific environment it wasn't always obvious how to apply it to your own. So my question is: should this work in other environments to the gate environment? And has someone experienced tried running through this from scratch lately? (And hopefully - the debugging guide will help too.)
15:57:09 <rwellum> Hopefully answer a lot shorter than my qn.
15:58:29 <mattmceuen> rwellum can we chew on this one a bit and discuss next week?
15:58:40 <rwellum> For sure...
15:58:52 <srwilkers> or we can take it to the channel after the meeting too
15:59:03 <mattmceuen> I think the answer is probably in the middle.  And I also want to go thru the install guide fresh with the latest myself.
15:59:13 <mattmceuen> It's been a couple months
15:59:23 <mattmceuen> #topic Vitrage Chart
15:59:30 <mattmceuen> 1 min left :)
15:59:34 <rwellum> I know we're tight on time...
15:59:41 <mattmceuen> probably can't do this justice but trying...
16:00:06 <rwellum> Basically my Mothership did Vitrage and it would be a good way of getting some developers into OSH by producign the chart.
16:00:21 <rwellum> Was going to have a crack at it
16:00:28 <mattmceuen> That will be really awesome.
16:00:36 <portdirect> nice
16:00:40 <rwellum> So it's a good candidate then?
16:00:42 <srwilkers> rwellum: thats awesome. we'd be happy to provide any feedback along the way
16:00:48 <rwellum> Cool beans then
16:00:53 <portdirect> rwellum: if a dev wants to work on it
16:00:58 <mattmceuen> Eagerly looking fw to helping with reviews on that - let us know if there is anything we can help with getting that going
16:00:58 <portdirect> its a good candidate :D
16:01:11 <mattmceuen> alright we're over time, great meeting all!
16:01:13 <mattmceuen> #endmeeting