15:00:04 <mattmceuen> #startmeeting openstack-helm
15:00:05 <openstack> Meeting started Tue Aug 28 15:00:04 2018 UTC and is due to finish in 60 minutes.  The chair is mattmceuen. Information about MeetBot at http://wiki.debian.org/MeetBot.
15:00:06 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
15:00:08 <mattmceuen> #topic Rollcall
15:00:10 <openstack> The meeting name has been set to 'openstack_helm'
15:00:13 <mattmceuen> GM/GE folks!
15:00:29 <tdoc> hi all
15:00:40 <srwilkers> o/
15:00:42 <srwilkers> tuesday again
15:00:44 <mattmceuen> Here's our agenda: https://etherpad.openstack.org/p/openstack-helm-meeting-2018-08-28
15:00:51 <mattmceuen> It's very Tuesday
15:01:00 <mattmceuen> please add anything you'd like to discuss, we'll get started in a min
15:02:20 <portdirect> o/
15:03:31 <mattmceuen> Alright let's get started.  We have several topics that got carried over from last week, so we'll need to stay on topic today
15:03:38 <mattmceuen> #topic ETA for 1.0
15:04:36 <portdirect> which ptl want to take this one?
15:04:38 <mattmceuen> https://storyboard.openstack.org/#!/worklist/341
15:04:42 <mattmceuen> ^^^^^^^^^^^^^^
15:04:58 <mattmceuen> These are the todos for us to get to 1.0
15:04:58 <tdoc> Yep, was just curious where we are...
15:05:19 <mattmceuen> A good chunk of the way through
15:05:22 <mattmceuen> with a good chunk to go
15:05:43 <mattmceuen> a lot of what are left on there are adding helm tests
15:05:58 <mattmceuen> which are great for new contributors -- any help with those would be welcome!
15:06:12 <portdirect> also - we really need the spec for values
15:06:16 <mattmceuen> yup
15:06:31 <portdirect> as one of the major commitments we have, is that a 1.0 is upgradeable
15:06:41 <srwilkers> yeah, ive dropped the ball on that one -- i was meant to take that and run with it, but have been dealing with other priorities
15:06:58 <portdirect> all good dude, thats been a hot potato to say the least ;)
15:07:00 <mattmceuen> you've had a lot on your plate srwilkers :)
15:07:40 <srwilkers> just need a bigger plate is all
15:08:08 <tdoc> ok, so i guess nobody dares putting a rough eta? ;)
15:08:11 <mattmceuen> Any other thoughts on the 1.0 target?  my gut says 2 months if we can focus on it as a team
15:08:19 <mattmceuen> but this is not the first time I've said 2 months ;-)
15:08:26 <portdirect> that seems about right
15:08:37 <portdirect> but we as you have pointed out, said that before!
15:08:39 <tdoc> ok, thanks
15:08:50 <portdirect> it very much dpends on how much time people have to put into it
15:09:03 <portdirect> theres about a month of work for the current active devs
15:09:25 <evrardjp> o/
15:09:26 <jayahn> especially, upgradable part would be so important since some of us is soon putting openstack-helm charts into production. :)
15:09:27 <portdirect> but being able to do that, without the whole "real job" thing getting in the way...
15:09:38 <mattmceuen> ++ jayahn
15:09:54 <portdirect> some of us, already have them in "semi" prod...
15:10:00 <jayahn> not only version to version upgrade,
15:10:28 <portdirect> exactly jayahn
15:10:41 <jayahn> really careful about things blocking upgrade, for example, chart name changes, pv stuff, ... etc
15:10:47 <portdirect> thats why we need to be careful as to when we rubber stamp "1.0"
15:11:18 <portdirect> im much happier with "unstable" than"stable, but lol we broke you"
15:11:29 <tdoc> yup, ok, i think that's pretty clear then...
15:11:45 <mattmceuen> Cool.  Any other thoughts on this topic before we move on?
15:12:00 <mattmceuen> #topic OSH user/dev survey
15:12:05 <mattmceuen> 'zis one you Pete?
15:12:28 <portdirect> yeah  - this kinda leads into a thing i want to do at the ptg
15:12:45 <portdirect> but i think its time to get a user and dev survey out
15:12:54 <portdirect> as the users/devs of osh are growing
15:13:17 <jayahn> portdirect: siri had a happy time to deal with upgrade issues from changes in chart. :)
15:13:17 <portdirect> and i dont know if we (as a core team) are fully aware of how people are using osh now
15:13:43 <portdirect> and if we are to do the best we can for this community, we need to
15:13:57 <portdirect> how about a quick roundtable now, just to get a straw poll of whos coming to these meetings?
15:14:07 <portdirect> and then i can follow up in the ml
15:14:22 <mattmceuen> ah - great idea
15:14:38 <portdirect> i could go 1st?
15:15:02 <mattmceuen> the floor is yours
15:15:33 <jayahn> so, r u saying you would do user/dev survey on openstack-helm? r we referring a specific ml thread?  just try to catch up
15:17:13 <portdirect> hello everyone, im pete (portdirect) and I spend my days putting telco infrastructure into kubernetes pods, leading the effort for a large telco. As a result im very lucky to get to work with the airship team, who produces a k8s platform for me to use. I probabaly get to spend about 20% of my time working upstream - which is awesome!
15:18:15 <portdirect> srwilkers: want the stick?
15:19:51 <srwilkers> im steve, and i spend most of my days playing with services that provide an interface for operators to know whether or not the services we're building are meeting their needs through monitoring, alarming, and logging (the really fun stuff).  i'd say about 30% of my time is working upstream.
15:20:31 <srwilkers> tdoc: your turn
15:20:32 <srwilkers> :)
15:22:20 <evrardjp> others?
15:22:26 <tdoc> Hi, I'm Teun (tdoc), I work for as a dev for a company that makes infrastructure management so
15:22:27 <tdoc> ftware. We focus on the HPC space, but have also had our own OpenStack deployment tools for se
15:22:27 <tdoc> veral years now. We're looking to integrate OSH into our solutions. At the moment we have just
15:22:27 <tdoc> been evaluating but soon plan to start contributing. Our devs are base in Amsterdam.
15:22:52 <evrardjp> cool
15:22:55 <portdirect> evrardjp: would be great to get to know you a bit better ;)
15:23:05 <srwilkers> tdoc: awesome :)
15:23:13 <evrardjp> haha
15:23:59 <evrardjp> I am close of tdoc geographically, I am new to k8s world, so still learning. Evaluating the possibility to make a product out of OSH+Airship for my employer.
15:24:27 <jayahn> so, all of us is doing this now?
15:24:40 <evrardjp> you're it jayahn
15:24:41 <portdirect> oh - another one pops up ;)
15:24:52 <portdirect> jayahn: want to tell us a bit about yourself?
15:24:53 <evrardjp> portdirect: ;)
15:25:45 <mattmceuen> tdoc, what kind of workloads will you ned OSH to support - normal cloud workloads, NFVs, other things?
15:26:08 <mattmceuen> Same q for you evrardjp :)
15:26:31 <mattmceuen> I think this kind of discussion will help ensure OSH is meeting needs of its users
15:26:38 <portdirect> mattmceuen: you havent intoduced yourself yet ;)
15:26:45 <mattmceuen> I'm mattmceuen -- I do airship and openstack-helm development, on behalf of telco use cases such as 5G VNF workloads.  I probably spend  30% of my time on upstream dev proper as well.  Note: the rest of the time is not consumed by downstream dev (we just do upstream) but rather integration and unrelated stuff.
15:26:47 <jayahn> Hi, I'm Jaesuk Ahn (jayahn),  I work for as a team lead for sk telecom, spend most of my time checking issues, fighting with boss, communicating with potential osh users, etc... skt is currently deploying three production openstack based on osh, vdi case, big data case, and private cloud. soon, i have to do a deep discussion with our network operation team on putting osh into 5g se env.
15:26:52 <mattmceuen> Had it sitting in paste!
15:27:02 <portdirect> lol - that cover everyone here?
15:27:13 <mattmceuen> lol jayahn fighting with boss
15:27:31 <jayahn> yeah.... one catch is my boss does not know that
15:27:33 <portdirect> if so then - mattmceuen those are two really important questions
15:27:41 <jamesgu> i can go next...   Hello, I'm James, working for an open source Linux product company (coincidently same employer as evrardjp). I am still new to OSH, evaluating and prototyping to integrate OSH and Airship into our Openstack distribution. We are looking forward to contribute OSH very soon.
15:27:50 <mattmceuen> hey jamesgu!
15:28:02 <jamesgu> Hi Matt
15:28:55 <evrardjp> fighting with boss is indeed priceless jayahn :p
15:29:03 <portdirect> evrardjp/jamesgu: what workloads are you looking to run on osh/airship?
15:29:06 <tdoc> mattmceuen: "normal cloud", also HPC, our customer base is pretty spread out.
15:29:30 <jamesgu> what is "normal cloud" here?
15:29:30 <portdirect> so - everything ;)
15:29:49 <evrardjp> portdirect: you have a million dollar? Cause that's a million dollar question
15:29:58 <tdoc> tenant spinning up vm's that do god knows what :p
15:30:00 <mattmceuen> "normal cloud" = the proverbial wordpress I guess? :)
15:30:06 <mattmceuen> hahaha
15:30:15 <mattmceuen> that sounds normal to me tdoc
15:30:18 <jamesgu> @portdirect: from a a few nodes to hundreds
15:30:18 <portdirect> tdoc: of course http://images6.fanpop.com/image/photos/33200000/dr-evil-austin-powers-33220839-480-386.jpg
15:31:00 <evrardjp> I am still curious on how does it matter portdirect :)
15:31:02 <mattmceuen> So portdirect:  your thought is to do this a little more formally @ the PTG, and perhaps get into a little more detail?
15:31:17 <portdirect> yeah - this is a great start
15:31:24 <tdoc> I also operate our internal private cloud, about 20 nodes, some 500+ VMs for CI purposes.
15:31:40 <mattmceuen> nice
15:31:49 <mattmceuen> thanks all for sharing, this is very helpful
15:31:59 <evrardjp> mattmceuen: the PTG is a great place for that, as it is also face to face, so you create stronger bonds between ppl enduring the same pains :p
15:32:03 <portdirect> I'll get a proper tread started this week, but my initial thought is to create some personas for osh users
15:32:10 <mattmceuen> +1
15:32:13 <mattmceuen> thanks portdirect
15:32:16 <portdirect> so we can try to account for these while doing specs/dev
15:32:30 <evrardjp> portdirect: persona is very important as it can face your documentation
15:32:40 <jamesgu> +1
15:32:41 <evrardjp> portdirect: maybe start with the docs spec that spread the documentation by areas?
15:32:44 <mattmceuen> Alright, gotta keep movin' !
15:32:48 <portdirect> ++
15:32:49 <mattmceuen> here comes the topic change
15:32:53 <mattmceuen> #topic Plans to support nodes with different interface names
15:32:59 <mattmceuen> tdoc this is you sir
15:33:04 <tdoc> yup
15:33:23 <mattmceuen> I believe we support this today -- overriding interface names on a per-host basis.
15:33:56 <mattmceuen> portdirect do we have (or need) a list of what is or is not overridable on a per-host basis to clear up confusion?
15:33:59 <tdoc> hmm, i guess I don't know how then...
15:34:10 <portdirect> mattmceuen: id need to double check, im pretty sure we have some gaps
15:34:37 <portdirect> most of our tooling uses magic to ensure consistent names accross hardware.
15:34:49 <portdirect> (use airship yo!)
15:34:54 <tdoc> I was looking at auto bridge add values in ovs charts and looked like it wasn't there...
15:35:13 <portdirect> yeah - jayahn's team pulled that out some time ago
15:35:19 <mattmceuen> This would be the approach
15:35:19 <mattmceuen> https://docs.openstack.org/openstack-helm/latest/devref/node-and-label-specific-configurations.html
15:35:25 <portdirect> i think the same functionality should be in the neutron chart
15:35:34 <portdirect> but would need to check
15:35:47 <portdirect> mattmceuen: the issue here is that some of the config values we define are not under the 'conf' key
15:35:53 <mattmceuen> ahh
15:36:26 <mattmceuen> Can we take this one as a follow-up tdoc?
15:37:06 <tdoc> hmm, yeah I might have to dig harder trying to understand all of this
15:37:19 <mattmceuen> I'll add a flup to next week to report out findings
15:37:33 <mattmceuen> if you need it sooner though let's chat in #openstack-helm on this topic
15:37:38 <mattmceuen> #topic Roadmap for Armada
15:38:10 <mattmceuen> Broadly speaking about Armada roadmap - it's a part of Airship and is part of the target state of that platform
15:38:28 <mattmceuen> Also a standalone deployer that will continue to deploy collections of helm charts in flexible ways
15:38:48 <tdoc> that was me too, though someone else asked if this is within scope of this meeting...
15:38:51 <mattmceuen> There will be a PTG session for AIrship, and it will be a great place to discuss Armada roadmap!
15:39:07 <jayahn> I did asked that. tdoc.
15:39:33 <mattmceuen> the #airshipit channel would be a good place to pose any specific questions (and the PTG)
15:39:49 <jayahn> i was wondering if osh community official recognize armada as a default deployment method? hens, deal armada manifest as first citizen.
15:40:34 <mattmceuen> Armada is the default way that a number of the OSH contributors plan to consume it
15:40:40 <jayahn> just want to check what voice osh community is delivering to users. :)
15:40:54 <mattmceuen> And it's what many of us would recommend (that's why it was built :) )
15:41:12 <mattmceuen> But I don't think we want to tie OSH to it in any way strictly speaking
15:41:14 <tdoc> Some on our team watched the presentation from the Austin and kinda got the impression it was presented as more of a poc back then, but it sounds to me like it's much more than that?
15:41:23 <mattmceuen> Oh yes
15:41:35 <jayahn> Oh yes
15:41:36 <mattmceuen> It's being used in prod by AT&T & SKT
15:42:05 <tdoc> We're also evaluating and probably will be using it.
15:42:16 <jayahn> for us, only Armada, not other parts of airship.
15:42:39 <tdoc> +1 here, most likely
15:42:53 <mattmceuen> cool tdoc - the armada devs hang out in #airshipit, and would be eager for any feedback and to help :)
15:43:13 <mattmceuen> I will move us on
15:43:18 <mattmceuen> #topic Helm 2.10 - finally TLS is coming to OSH properly
15:43:18 <tdoc> yup
15:43:21 <mattmceuen> yay!
15:43:33 <mattmceuen> https://review.openstack.org/#/c/588085/
15:43:47 <mattmceuen> This is a good one to share
15:44:15 <mattmceuen> Adding in Helm 2.10 opens the door
15:44:40 <mattmceuen> portdirect and others are working to get us through it
15:44:55 <mattmceuen> And then cluster TLS bliss
15:45:01 <mattmceuen> anything else to share or questions?
15:45:47 <mattmceuen> #topic Ceph - whats going on next?
15:46:11 <mattmceuen> This topic is a bit vague :)  Can anyone clarify the intent?  The owner got lost in my copy & paste from last week
15:46:26 <srwilkers> im not sure what it was meant for, but i hijacked it
15:46:34 <mattmceuen> lol
15:46:43 <mattmceuen> Moving ceph to osh-infra
15:46:44 <mattmceuen> https://review.openstack.org/#/c/596880/
15:46:44 <mattmceuen> https://review.openstack.org/#/c/596881/
15:46:44 <mattmceuen> https://review.openstack.org/#/c/596891/
15:47:03 <mattmceuen> good call srwilkers
15:47:11 <srwilkers> this is likely to be contentious, but i think it's time to move ceph to openstack-helm-infra
15:47:27 <srwilkers> bonus: it's one of the items on the 1.0 roadmap
15:48:12 <mattmceuen> In terms of contentiousness and timing - is the only concern that there's a lot of active work going on in the ceph charts now?  Or additional reasons?
15:48:27 <srwilkers> yeah, largely because there's a lot of work hitting the ceph charts right now
15:48:32 <mattmceuen> yup
15:48:39 <srwilkers> along with the elasticsearch snapshot work ive been doing
15:49:28 <srwilkers> once ceph hits osh-infra though, we can deploy the Elasticsearch s3 snapshot repo backed by rgw's s3 api by default, and expand the elasticsearch helm tests to verify it works
15:49:40 <mattmceuen> Alright - let's keep moving for now, we can add more specific topics to following week agendas
15:49:43 <mattmceuen> #topic Team meeting time - any other times that would work well for most team members?
15:49:46 <srwilkers> could also do some cool stuff, like snapshot the logs gathered during the gate runs and provide them as part of the gate post-run jobs
15:49:53 <srwilkers> ok
15:50:07 <mattmceuen> So the OSH team meeting has been at this time for a while
15:50:15 <mattmceuen> It doesn't work well for our Korean teammates
15:50:35 <mattmceuen> But we also have a lot of north american and european team members who come frequently
15:50:49 <jayahn> true, but finding a good time for everyone would be very difficult, since we now have people from europe.
15:50:50 <mattmceuen> This is a general poll, not for specific times (that'll take more time) but just for ranges
15:51:20 <mattmceuen> Are there any general time ranges that would be reasonable for everyone?
15:51:41 <mattmceuen> jayahn would end of day or beginning of day be better?
15:52:07 <jayahn> depending on which exact time.
15:52:11 <mattmceuen> same for tdoc evrardjp jamesgu ?
15:52:44 <jamesgu> I am in NA, west coast
15:53:03 <mattmceuen> Cool - we have several other folks from that time zone as well
15:53:05 <evrardjp> tdoc and I am Europe, so ...
15:53:14 <jayahn> I am in Korea...
15:53:41 <evrardjp> so it won't be possible to please everyone, so maybe start rotating?
15:53:52 <evrardjp> if that matters
15:54:09 <mattmceuen> Does anyone have any experience with rotating meetings on other teams?  Best practices?
15:54:19 <mattmceuen> We may need to put the final stamp on this at the PTG
15:54:39 <jayahn> let's get survey on this.
15:54:49 <mattmceuen> I think a survey is a great idea
15:54:54 <mattmceuen> I did one once before
15:54:55 <jamesgu> +1
15:54:56 <tdoc> this is a tough one indeed, current time works for us, though it's pretty late in the day...
15:55:08 <jayahn> without some sort of automatic timezone translator, it is difficult to get an idea no others' time
15:55:36 <mattmceuen> Ok - I am determined to get thru our agenda!
15:55:40 <mattmceuen> #topic "OpenStack-Helm  Authors" -> "OpenStack Foundation" copyright update
15:55:59 <mattmceuen> we discussed this a few months ago in the team meeting, and I finally got un-lazy enough to put in a PS
15:56:07 <mattmceuen> https://review.openstack.org/#/q/status:open+branch:master+topic:copyright/osf
15:56:20 <mattmceuen> This is just an FYI since it's wide ranging.... ~1400 one line comment changes :)
15:56:51 <mattmceuen> "OpenStack-Helm Authors" is not a legal entity, and these PS update *just those* copyrights to the OpenStack Foundation.
15:57:00 <mattmceuen> Finally
15:57:07 <mattmceuen> #topic PTG Etherpad
15:57:21 <mattmceuen> Reminder to please add content to the etherpad to discuss -- https://etherpad.openstack.org/p/openstack-helm-ptg-stein
15:57:35 <mattmceuen> #topic Roundtable
15:57:41 <mattmceuen> jayahn shared some PS needing review
15:57:45 <mattmceuen> https://review.openstack.org/#/c/590184/
15:57:45 <mattmceuen> Prometheus-alertmanager: modify wrong variables.
15:57:45 <mattmceuen> https://review.openstack.org/#/c/587812/
15:57:45 <mattmceuen> Node-Exporter: allows to set collectors enable/disable
15:57:45 <mattmceuen> https://review.openstack.org/#/c/549824/
15:57:45 <mattmceuen> Cinder: add pass message for enabling application
15:57:45 <mattmceuen> https://review.openstack.org/#/c/593991/
15:57:46 <mattmceuen> Tempest: add option to execute cleanup always
15:57:55 <mattmceuen> I will add them to the #openstack-helm channel as well since we are out of time
15:58:22 <mattmceuen> Also FYI -- next week portdirect will be leading these! :)
15:58:49 <mattmceuen> Any other closing thoughts team?
15:59:10 <evrardjp> none
15:59:12 <evrardjp> thanks!
15:59:18 <mattmceuen> In that case -- thanks everyone, and have a great week
15:59:27 <mattmceuen> #endmeeting