Kubernetes Community, 25 Aug 2016

Previous Meeting Next Meeting

⏯

youtube image

►

From YouTube: Kubernetes Community Meeting 20160825

Description

We have PUBLIC and RECORDED weekly video meetings every Thursday at 10am US Pacific Time.

https://docs.google.com/document/d/1VQDIAB0OqiSjIHI8AWMvSdceWhnz56jNpZrLs6o7NJY

Demo AppController, SIG-Docs report, 1.4 release update, 1.5 milestone update, K8s Dev Summit teaser, feature repo update.

A

Good morning or good local time to you today is August twenty fifth, and this is the Cooper Nettie's community meeting. We do this once a week. We have our videos of these events posted on YouTube and you can see whatever you miss in weeks upcoming or weeks prior there there's a link in the email I send out with the agenda to cooper nitty step each week. So this week we have a demo of an app controller by mon said ma, che, ok, ma che quick yeah.

B

Much easier, yeah.

A

Ok and we are going to have a couple of special interest group reports and an update on 1.4 from our 1.4 release are including what it means to be release are and then we're going to go through some 1.5 mile stone reminders, because just because we're finishing up 1.4 doesn't mean we're not already thinking about 1.5, so much age. You want to introduce yourself and get us started. Yeah.

B

Sure, thank you. Ok, so my name is magicjack. I'm a software engineer at my aunties and I'm going to present up controller, which is a tool that that was written by ilya ocean gets gamble. P&Amp;G nova can be mostly and it it aims to solve the problem of deploying complex, complex workflows, which have which have pieces which depend on each other. So we we handle cases in which some cabinets objects are heavily dependent on whether other planets objects are ready yeah.

B

So I guess I guess I can start and like I will answer all your questions or I will try to okay. So the first thing is this I my screen: this is the dependency graph. It's the first of the of two.

B

So this, as you can see here, are some jobs and Paul's. There's the replica set here the service, so the edges in this graph are dependencies. So we can see that what is really at the top of the graph is existing tests, job which isn't managed by up controller, but it can be yeah.

A

Make it the font size, bigger, I,.

B

A

Thank you, cool.

B

Okay, so, as you can see, this is an existing job that is not managed and not created, but by up controller. It is something that all that is meant to already be in the cluster. Before we run, we even run up controller and all the other objects are actually created. An object, orchestrated by up controller when the dependencies for for these objects are meant. Ok, so I go to the demo. So since we made demo I think we go to a go and see gaps, we added some stuff. uh You make.

A

This point sizes bigger again.

B

Ok, yeah sure sure. Is it good enough? Oh.

A

Let's go one more there, we go. Yes, ok,.

B

Cool, ok, so what we need to do is first I will create the job that I was talking about the diss already existing the cluster. Now I am creating a up controller port. You can see it being created in here. It has an init container which creates additional API endpoint vr4 Falcon resources right now, so we need to wait a sec for it too.

B

Probably it will it's possible that it's loading, the image, because I'm using the latest tag- and I think it could have been it- could have been yeah- it could have been updated in between I'm sorry about that will need to wait just for this. So in the meantime, when we're waiting for the image to download, we have this dependency file. We each basically it contains its iron Phi, which contains all the edges from the product, saw a minute fork.

B

As you can see, it can contain some. It contains normal communities, metadata kind, name, etc. It also contains our custom metadata. This is this is the stuff that we're currently working on. This is the success factor for replica sets, so it this particular dependency means that it will be counted as ready. Only eight forty percent of the replica supports are ready, let's see if it's tryna go and it's running cool okay. So now I'm going to create I'm gonna use the file that you saw. I mean for and create to create dependencies.

B

C

B

Will create third, I will create the definitions of the kubernetes objects, which are basically just they are basically just communities objects, raw definitions, wrapped in third-party resource, because we need to store them. Some hope for up controller to pick them up end and go okay. So I will just I will just run the script. There may be some errors, but fear not everything's fine, because we already have these objects here.

B

So the sleep sleep is because sometimes takes a bit. It's sometimes yeah.

A

B

A

B

Bit before the third part resources, api extensions are actually going ok, so here we can see the log of the actual app controller. It's already running yes, so we see that there are cycles detected. Okay. This is a new feature. It wasn't on the last demo, so we can see that some of these some of these some of these skewness objects, are in a cycle and tendencies. So if we see it, the graph I will spare you the I will spare you the details. The dependency that is the problem is actually here.

B

This makes the graph have a cycle right here, as you can see. So this is the graph that we really want to deploy. So what I can do is just you skip CTL.

B

To get dependencies.

D

B

So I'll, sorry, the the font is up. It's too big, ok, I hope this is fine yeah, so I I named this path dependency as bad, so it can be easy to remove it. I will delete it.

B

Yeah, it's rated, it's not here anymore, so we can actually run the app controller process again.

B

We do it by exacting the binary on the old right. Now, it's a bit cocky, but well, it's all really really quality. So, as you can see, the defaults are being created. The replica set is being created, the job is, it was already there it'll successful, so it didn't block the rest of the graph.

B

Yes I'll, basically here at the polls for the replica said there, because it is being created, not this one yeah. So right now we are at the stage when pod and what five are starting to be ready. We need several more seconds and then the rest of the graph will will go. As I said, we have replica assets as you can see it works. We have services we for now we check whether services are are created.

B

We are currently working on checking the state of all underlying surveys, objects, so the objects that are selected by the service later yeah. So these polls are ready, so the next pots are being created. We can actually we can actually stop it.

B

If we see the load, then we will see that nothing happens for a bit after these boats are ready. Next notes and grab should should be created, but they won't because up because of control and processes stopped. But it's not a problem for us because it will just be able to take take over when we when we are ready. If we, if we want to tweak some dependencies in further parts of the graph or something, then then it is possible to stop and just just start the control process again.

B

So, as you can see, nothing happening, I'm running it and it should it's creating here. The service was created six seconds ago when I started the process again, and the service was services actually right here. So eventually, like all seven should probably should we create it in the minute. So.

A

Much hey: we got a question about yeah.

B

A

So mike says upon what does something wait? What determines if a note is ready independence? Oh.

B

Yeah sure, okay, so four pots, for example, we use readiness probe. So you as as an operator as someone who uses up controller, you are defining how how does the port? How does the pot? Actually he becomes red here? It's like really. You know it's a placeholder. It doesn't deploy anything real, so it just slipped for a couple seconds echo something to TMP health and read his bro. As you can see just you know, checks the contents of this file, but you can do whatever unique in this pot or balls for jobs.

B

We check if, if it's, if it's completed for service as I said it's right now, it's it's. It's checking whether it's created and we are working on improving this for the four replica sets. We have these dependencies, which can which can where you can define how the percentage of the underlying pots that you want to be active. So you can say that okay, just replica set is fine for my for my deployment purposes when ninety percent of the of them top of the under rifles are ready, yeah, yeah, so I guess yeah. It's done.

B

We missed it. Yeah, so it's done whole deployment graph when I know it's kind of hard to to see that it's actually there, but it is actually there you can. You can actually run system on your own cluster by using our because you can clone the repository and and and run this if you, if you want to see that it really works but yeah. Basically, that's that's the end of life demo and if there are any questions, I will gladly answer them. We.

A

Have one more in the way? Actually we have two more in the chat. One is an easy one, which is: where is the link to the repo and the answer is we can put that in the notes? I believe it's in one of the marantis repos? Yes, yes,.

B

Yes, for now it lets in uh in right here mirantis gates, communities of controller. So I can send it to your via slack or sure.

A

B

A

That would be great and then I know that there is an open feature on this participating someone right now, it's in it's a container that exists on its own. That runs this correct.

B

I'm, sorry, can you can you repeat the question your.

A

App controller exists in its own container. Managing oh.

B

Yes, yes, yes exactly! This is it's basically right now, it's a pod! It's right here, which runs the run it process from it's based on Alpine. It's it's running indefinitely and we are exacting process on it. We plan to have it some kind of API which we can talk to to run and stop the app controller, but we had more urgent issues to fix an implement. So for now for proof of concept. We think it's it's ok, but we we want to go there eventually. So.

A

We've got one more question here which is well one that I see at the moment. Luke doesn't understand fully the motivation. Why wouldn't services just retry if they can't connect to dependent services? What use cases might he be missing in this thought process with them, where it would be sufficient where, where that would be sufficient, and then you have to bring more than you can bring up services in any order? Ok,.

B

So 11 sorry one example is a multi slave nice glow Mexico, which basically needs you to run. Operations operation, run faults and what which we chilled in each you to run communities objects in in specific order, and you cannot really afford research there. Another example, maybe a bit later on OpenStack deployments in where you could you can you can specify your whole deployment graph in the controller in up controller format, so yeah I deserve to two examples. When you could actually meet this.

A

Excellent thanks and thank you so much for the damn. There may be some other lingering questions you want to answer in the chat or clarifications you may want to answer in the chat, but we're gonna go ahead and move on in the agenda. Thank you very much. Okay,.

B

Sure, thank you. Yeah.

A

Alright, so next up do we have I. Think up next in my agenda, was the update for cigs, so Jared? Are you on the call just talk about rocks yep.

E

A

Tell us we can tell us about cig docks and cuz you're Brent, well, not brand new to the project but introduced you as taking over the the cig ducks effort from the from the google documentation team, as John will house and has left so. Can you tell us what's going on with sig docks and update us on the plan ahead? I would.

E

Love to yes, I feel like I've, just been pulled out of my shrink, wrap so um I know exactly lots of crinkling, um so yeah I think I, I, guess I kind of have my agenda items in reverse order that so I think you know, based on your intro, the docs revamp plan is probably the best place to start and.

E

This should be open to everybody in the group, so feel free to take a look. um I.

A

Shared it with all of you burn, it is dead. If you can't see it, you're not participating in the communities done. Google group so excellent.

E

Is definitely open to the sig dot groups, so so this is just a quick overview of an assessment that Devin Donnelly did on the contents and I think it matches pretty closely with what engineers in the community have told us.

E

But overall, the goal is to make some immediate changes to urbanize docs findability in structure, so um we want to get together a solid base of our content and a solid set of processes. So there's some really low hanging fruit right off the bat which is fixing our autumn automatically generated reference content. There's some broken parts of that process which I'll highlight in the dead links report later in my chats.

E

But the core problem is you know, building a set structure and building some templates so that everybody knows what content goes where because right now we're at a spot where we really can't scale or grow, because nobody knows where to put the contents and developers coming in to Coober. Ladies, do don't know where to find the content so we're kind of at an impasse. So the assessments and Doc's plan lists some short-term fixes and some longer-term sorry q4 fixes that we need to put in place, but over the next quarter.

E

We definitely want to fix the release process with references content. Just to you know, create a baseline user experience.

E

Then we also want to roll out new templates and a new site structure and we'll be doing this in stages and I've been walking through those stages and granularity and deeper granularity at the sig Doc's meeting so feel. Free to this is my plug for the sig Doc's meeting so feel free to jump on the slack channel or hop in on that meeting its weekly on Tuesdays, but most of the contents in the documentation plan should not be a huge surprise.

E

We've already started working closely with Philip wittrock on fixing the the content, the content generation process for a reference content and Steve Keary and Devon Donnelly have already started working on creating new templates and we'll be sharing those the high-level overviews of those templates with the community next week and then, probably in two weeks, will be you know, giving those to the community to really bang away at and give us some hard feedback on. So um so.

A

Jared I'm not actually seeing anyone from external to Google in the doc so may need to share it. Okay,.

E

A

Burnett east ave would be helpful sure.

E

Let me do that.

A

A

If that note, or if that link was added after I sent it out to the community about a half hour ago, then it might not have automatically added it. Okay,.

E

I just added them to the doc, so please, let me know if you, if you can't get into it, so I'm not going to I'm not going to read the doc out loud to you. I don't think that's the best use of our time, but feel free to take a look at the documentation plan. I have broken this plan up into a spreadsheets and I'm using that for tracking progress and I'm.

E

I believe I have a PM on the on the google side, who's going to be helping me out track some of that progress and doing some read out reports for for folks at these meetings. So I will start inviting her and her name is amanda lewis, she's gonna be helping us out so, um which is great because I am, I am one person and I'm running into to sustainability problems with with juggling all these plates. So um so on this. So the high-level bullets of this plan are there any questions right off the bat you're.

A

Not sure if we have everyone able to get in yet so we may have to take those asynchronously. As far as questions on this plus, we do need to keep moving in the agenda, so I think maybe feedback in coover, nidhi's, sig docs on slack might be a good way to do. This sounds.

E

Great yeah I am I'm always on slacks and feel free to to ping away at me. um The other sort of high-level bullet points the cabriolets boot camp draft, is out feel free to take a look um to. If we just had this feedback right now, I believe Steve Perry is taking a look through that content and doing a quick fix of some of the contents of lakhs, but I would love to have feedback from folks of like a positive negative I've already got a lot.

E

It's been great, also the dead links reports, an engineer ran a link checker against all of our contents on communities on Io, and it revealed a lot of broken links in our reference content, which means we need to fix that process, but there's also a lot of broken links outside of that that we're working to fix. So if people want to take a look at that reports at one of those fixit's I will be most grateful. So thanks a lot. That's all I've got excellent.

A

Thanks so much YC fill up on here are 1.4 release. Bizarre hello, yes, I, see, Phillip, hear your message: Cooper net e cig docks.

D

Yeah yeah, oh yeah,.

F

Sorry I was a trouble with my camera and clicking the buttons. Oh, but.

A

They're so complicated gerber.

F

A

About what it means to be a release are now that you've been handed this scepter and you know, are in charge of releases re sure.

F

Yeah, so the role, I think, is a it changes a bit from released released to adapt to the needs of what's going on, but from a high level. It's too mmm Sheppard the process of getting their lease cut and coordinate, identify what the items that need to be completed our and either coordinating with other people to make sure the responsibility for those is taken up, or in some cases doing it myself.

F

Things like cutting the release. Alpha I've been working with david mcmahon and manually doing, but he's helping me as well as setting up projects for release branch testing and this sort of manual work. I do a lot of. It also means I. Do a lot of police work and end up being kind of bad guy. A lot of time.

F

So so some things going to talk about in a second that would probably be at the top of the release. Our lists are making sure like that features that are going into one point for that: the folks shepherding them and owning them have communicated their status and that we can clearly figure out what made it, what didn't make it and what might still be coming.

F

Another thing is a test stabilization, so Daniel did a ton of awesome work last release around just so many improvements for how we handle flaky tests, but unfortunately, there's still a lot of manual. Human intervention required to fight them back all the way down, and so I'm going to be probably uh reminding folks and encouraging them to own any flexisign of it.

F

I'll get a little bit more into that later. So I'll talk about. First about the alpha release status, which I was asked about us community meeting and I've been working hard on getting that alpha cut. There's just been a number of issues, one after another that have actually kept us from getting it done and I'm hoping today's the day but I'll.

F

Let you know when it's cut some of those issues in case you're curious our difficulty getting tests passing at a single commit hash, where this variety of test suites that run, that we depend on and uh different ones, have been failing for periods of time. So getting all this fixed another and.

A

That relates back to the test flakes that we were talking.

F

It does go back to the text flakes and in the actual test suites required for cutting the branch is a superset of what's available. What's on merge to blocker, just because the merge kids running doesn't mean we're actually good enough to be in a spot where we cut a release. Branch specifically the scalability are the serial tests have been quite painful and they're. Both they both run slowly, which means it's hard to iterate and get them fixed, and then they also have been failing.

F

Quite a bit, I've been working with a good kept on that too. So Mike Denise, if you're here. Thank you for your help there.

F

So moving on from the alpha release, feature free status. So in case you don't know, I think TJ may talk about this more later, but a future freeze was on Monday of this week, and so what that means is we do not want to be merging any changes that are not either bug fixes tests or docks. So if you have a change, that's either for a feature going in that you want to be in 14 or a baker factoring or basically any changes you're making to the code that could introduce problems.

F

We we really don't want to put those into 14 so that we can cut it really spans soon here and open up 15 for everyone.

F

And so, if you, if you're working on something that is not a doc's change, a test, a change or.

F

Should I forgot what the other one less anyway, if you work on something else, then send an example: send an email to me or TJ to try and get that figure out what you should do there I really know I'll be I kind of ever been monitoring the queue because I know some people, don't always all get the message side ones here so I do apply.

F

Do not merge labels to PRS where I wonder if, if I don't understand why this is in 1.4 milestone, so I'll try and follow up with you on that, and that's just why that's happening, and once I understand what's going on there and it it makes sense to get in 1.4 I'll, remove that label.

F

Test flakes its identity, well to finish. Dev I think yesterday, basically with subjects saying important, basically saying: please take a look at all the tests flakes assigned to you. It doesn't mean you need the p0 and p1 specifically doesn't mean you need to fix them. That day it doesn't mean you need to drop everything and do them, although that would be great, but the first step we do need to do is figure out.

F

What's just been dropped on the floor and doesn't have an owner because the person is assigned to is on vacation or is doing something else right, and so that way I can look at the issues that I know I need to find, no turf or so just remove, apply to label it's an email, but those labels saying that he requires my attention.

F

So, if you're the correct owner and you're going to own it remove that label, if, if you're not going to work on it in the next, probably three days on your test, flake on a particular test lake then probably leave the label on there. So I could try and find a new owner for it. A code / estimate, length and branch cutting I believe TJ can can correct me on this if I'm wrong, but I think the original estimate for the cutting the branch would be September.

F

Second I really like to have that cut down as much as possible, because there's a lot of people that really want to get stuff into the 14 merged into the 14 milestone rumors right away that actually needs to go in 15 is more important for 15 and they just don't want it sitting around forever, which is totally reasonable to feel that way, and that's why I want to cut the branch early and, lastly, down the road. Just something to be prepared for is upgrade. Testing is still a pretty manual process.

F

I took a lot of time last release I'd like to get in front of it as early as possible. This release boat we're going to need help there and.

A

So the first alpha would be cut after the branch know. You already have off us. You.

F

Don't have enough yes, uh the office should, because we should had an alpha office should be cut before the branch. Okay, we're just there's technical difficulties. Getting that alpha cut after the branches cut will start cutting betas, okay,.

A

All right so does anyone have any questions about 1.4? What.

C

A

Your ID on slack, please ah rock yeah,.

F

G

F

Had a real quick question.

G

First, the piece of feedback: this is fantastic, so um this is sort of exactly the sort of release status. Stuff I've been hoping for prior to every release of pure benetti's, and we finally got it. um So that's super awesome. Thank you very much. Oh Mike will.

A

Be here every week.

G

Noticing that we've had complications cutting an alpha for the past week, since we went into code slush I'm wondering if it makes sense to try and cut the beta slightly before the second just because the second is the friday before Labor Day weekend and I. Think a good chunk of people are probably not going to be super crescent, and it would just be the ideal that we can ensure the beta is cut for a long weekend or a lot of people are going to lose context. Okay,.

F

That's a really good feedback. Thank you. I'm, going to work with I'm gonna, try and make that happen like it seems reasonable to me, but I want to make sure that there's nothing. I'm not aware of that. That might be a bad idea for so.

A

There's also, we would level as much help from the whole community on testing the betas, as we can. One of the challenges we had with 1.3 was not getting a lot of upload in my test cluster at my local company and drun, weird stuff against it kind of testing. Once we cut betas, so whatever effort we can get from you, the community around that would be super helpful issues. We evade us out.

A

And we have something from Daniel: he'd argue for cutting the branch first and first beta ASAP, yes, Daniel I think we're stuck waiting to try and get a test. That is a test suite that is consistent across the single hash. A commit hash, yeah.

F

We actually finally got that broken and now we're having a separate issue with birch conflicts for some reason, I'm going to dig into it more okay,.

A

Awesome but yes, I, think the goal Daniel is that we are going as quickly as we can to that all right. Other questions for Phillip.

A

It sounds like a big fat, no or you're on mute.

A

Right: well, thanks we'll try to get a release status from Philip every week. We can his time until we get this thing out. The door yeah.

F

And it's going to give it a quick shout-out to uh Eric tune. He's helped me a lot with this and- and I think this releases a lot smoother because of the experience ah he's helped dealer he's.

A

The ribbon releases are emeritus, he was the 1.3 release. Are okay, so questions you can find Phillip on slack at P, went rock and yes we're seeing many thanks in the chat which is important to say and continue to convey all right. ah Next up is 1.5 miles, don't schedule. This should be pretty quick, I think TJ said so. um Tj, yes,.

H

I, just unmuted and tried to share is that.

G

H

That looking good yes.

D

H

So here is the pr that is out for the 1.5 schedule. If you guys are seeing that I'll go into the views, we can actually talk through this being proposed at today's community meeting, we went over.

H

With which, since 1.4 is on track for middle of timber, that puts us firmly in December, obviously the last two weeks or so of december are a bad time to try to release things. A lot of people are on holiday, and so basically, I sort of set the in date at towards the beginning of december and built at work back on that there's also, the the u.s. thanksgiving holiday is in there, which is two days off for a lot of people working in the US, and some people take more also I think cube.

H

Con is in early November, which is a two day conference. I'm sure lots of people will be at so.

A

Let me stop you for one second there, which is count yourself for three days. Oh consider three days, because we're also going to do a developer summit after khoob, con and I will be getting more information out soon about how Helen who gets registered for that, because we're going to try and keep it to a small group while still making it possible to have new fresh flood in. So there will be more conversation about that coming up, but consider that it might be the eighth, the ninth and the tenth of November.

H

Okay and then otherwise, I took one other thing into account. I got feedback from a few people that for one point for a lot of our sort of due date or end of the sub milestones landed on Fridays. A few people said that was that was hard going into the weekend. They'd, rather have it land on a Thursday on Monday, something like that, so I tweaked by a day or two a few things. Otherwise, this looks a lot like 1.4, it's the same amount of coding and almost the same amount of bug fixed time.

H

So, let's move into the overview, so coding would start just the day after our predicted release date for for 1.4, so be September, 19th, I've booked, seven weeks of coding, which is identical to 1.4, and that would bring us to feature complete on Monday November seventh. So this is an instance of moving it from a Friday to a monday, I think for 1.4. We had good success with sort of letting the queue drain through the weekend as as needed.

H

Hopefully we don't need that, but I won't bank on that and then getting the remainder of things in by monday sort of midnight. So we have that one extra day and then from there do what we're doing right now, which is moving into a bug fixed period. I I put this one at five weeks, instead of the four weeks for 1.4 partially, because cube con is in there partially because Thanksgiving is in there.

H

So that looks a lot like a week for for a lot of people and then I put our release for good target for December eighth, which is a thursday if we manage to slip by a few days the week after this I believe, is the 9th 12th through the 16th, it's probably feasible to get some work done and maybe get a release out. If we go to the next week, the 19th, through the 23rd, we're going to see a lot of people on vacation.

H

So I didn't want to push this too much further back just in case. We do end up slipping for a few critical bugs by a couple of days, and then it can scroll down I put in a little more details. This is sort of in the same format as as this, this page was for, for one point full or in a couple more details.

H

Looking for feedback or suggestions or any recommendations, although with the kind of hard deadline of middle of December, there's not a whole ton of options to move things around.

A

Thank You GJ does anyone have anything they want to talk about her? Can we just send your feedback offline, or do we want to try and just say yes, we're good to go? Let's merge this or prove it doesn't sound like there's a lot of commentary, so I'm good with. Let's put this out there and we adjust as we there we go. We've got a bunch of plus ones coming through.

A

Let's go with, it looks good and we'll adjust it if we have to, as we have done with so many things in this community. Yes,.

H

Absolutely all right, thank you. Yeah.

A

Thanks TJ awesome all right, so we have a little bit of time, and so there were. We have the possibility of another special interest group report. Is there anyone who wants to give one an update? I have didn't ping anybody because I'm a container gone this week, so I missed that as far as pulling up another special interest group volunteers.

A

All right, then, let me tell you a little bit about what I'm planning for the developer summit after coupon for November 10, so Cameron, Tahiti and I are working on pulling together. First Cameron has I believe secured space at this point, so we will have it on the tenth of November, which is the one the Thursday following coupe khan, and the idea at this point is we want to keep it a small group.

A

So we have really meaningful discussions about the future of Coober Nettie's, and all of this is going to come out in formal docs. And how did you how to engage with this in the upcoming week or weeks? But it's good to give you guys a heads up. There are going to be multiple rooms which will have topics assigned to them and we're going to ask people to submit topics, so this isn't submitting talks or proposals for talks.

A

This is submitting more like AB off topic and we'll do some curation for some of the rooms to give us to give us space and intent with what we want to make sure gets talked about, and then we're going to leave some rooms as more of an unconference, empties or without topics so that people can have ad hoc topics come up as necessary. There, it's going to be one day in keeping it to 150 people, we're going to do this in an I'm, proposing that we do this in an interesting way.

A

We do this about half or a little bit more by invite. So we know that we get the people that need to participate for sure we get make sure we have at least one person, one developer from each of the core contributing companies making sure that we're we do have the right group of people participating.

A

They can help move things forward, make decisions and potentially take information back as a to their own to their companies and then we'll have some small portion based on who submitted topics that need to be talked about, and so people will have a topic submission process, and so sometimes some people will be invited based on submitting topics and being willing to guide those discussions or even be you know, administrative help on those discussions, and then there will be another pool of seats which will be lottery, so it will be, please sign up, and then we will fill from that list through a random lottery for people that that so that we can get fresh blood in so that we can get people engaged, because we're going to have too many people, I, suspect, I, hope we have too many people that want to participate too.

A

For this planning and community discussions- and you know a face to face- is like the sig storage had a big face-to-face week and a half ago two weeks ago. That kind of thing so does anybody have any commentary on that as a model?

A

At this point as to how we can make sure we have a small enough group to have meaningful discussions and also make it make sure that the right people are participating, that there's broad distribution and that were not excluding people who might be newer to the community and eric says honestly 150 seems pretty large. Yeah I was struggling with that number two now think of that as 55 tracks, so five different rooms. So it might be five separate conversations of more like 30 to 50. There will also be hallway tracks, I'm, sure, and so on.

A

That still seem really bigger.

D

It is essentially so far just sent you.

D

A

Where I got ended up with.

A

Your connection is really loud and really bad, so I'm gonna mute you I apologized for you. You did already excellent. Thank you. It is robo tune exactly Jason, um so I agree that that it's big but I think we're gonna have even more people than that wanting to participate. So that was why I set the number that way and tried to make it partly invite and partly partly lottery.

I

I think that one concern, which is the decision-making model around this just to make sure it's open and in it's not sort of the behind closed doors, everybody making decisions and then there's some big thing that comes out of it, so I think just making sure that we have transparency in terms of what decisions around the table to be made and what is the? What is the accountability?

I

So if the council of elders does present or whoever is sort of the decision-making body at that point, I think that you're really good just to make sure that's happened. Okay,.

A

um And yes I, my I think my whole role here is to try to improve transparency generically so I'm totally on board Jason with that and the topics that would be proposed. I want to do as some sort of vote, which leads to the next I just said vote, which then suddenly makes some people try to run away from the community.

A

We're trying to do a bunch of this by you know general broad consensus as much as we can, without falling into risks of you, know, block loading, and you know the the dumb pieces of politic. So it's a cig, Taylor ship, okay, there we go so Clayton's. Actually, clay makes a good point: we're building a great big distributive system which doesn't have which, which doesn't have power either hierarchically hierarchically, managed or completely distributed to the to everyone who has an interested opinion in it.

A

You know informed or not and that so we're trying to distribute the power into the special interest groups as much as possible with exception processing handle do differently, so we will definitely be pulling in the special interest groups to make sure that topics for their areas are covered well and getting people with general interests.

A

So this again, isn't it's more like a republic and less like a straight democracy in in the land of voting, though, into your point about elders, I had kicked to the can of how we would vote on elders down down the path of it and I still need to close out the proposal on elders and get that signed up, but there were, there will be further plan about voting after I. Get that done all right.

A

Questions for me about the suggestions I'm operating on the no surprises model of community management or trying to anyway so I float ideas early here before they're, fully formed, get some feedback and try to tune them up.

A

Okay, so we have a cig, tater ship and a new t-shirt thanks Aaron. Alright, anyone else have topics that want to talk about before we break for all of august. Our next meeting is September. First.

A

J

Hi I'm going to say one quick thing: I wanted to thank everyone for all the work that you've done on the future. Repo I can't stress that enough how valuable people have found that and I know. Tj already talked about the a timeline for 1.5. Please start thinking about your features. Now talk to your sig leaders, get consensus in your cigs and and bring the proposals, there's also a bunch of stuff that moved from 14 to theoretically the 15 timeline. Please go back, let's try and avoid things going stale. So just do your best. So.

G

It's just kind of quickly on that comment. I think there are a couple features that were marked as staple and ready to go at 14 that still don't actually have any of the check boxes checked off. Nor do they have any links to any issues within the crew benetti's projects that detail how is finally implemented or what its current state is. um I'm, thinking specifically of the node spec Peng and dr.

G

validation, I trending up those features in a second and I know if Donn is very busy- and maybe this is just a conversation to have with her but just sort of a broader thing. It seemed like when we were going through the future turn down. A lot of people said yep totally we'll take care of this later volatile check will like put the label on now and we'll check the boxes later, and it's been a couple days and I still really hard seeing a whole bunch.

G

So I agree it's a huge step forward, but just while I sort of understand the spreadsheet that you are presented to us last meeting and what's considered like ready to go for alpha, it's still kind of difficult to get a model of like what's actually implemented.

J

Are there any others, besides that one that you're concerned about I mean I, here's my default, my default? Is it gets punted? So if it's not done, I will say Don sorry, this doesn't make it, and that will be that that's my vote. If we're going to be following model, then we should follow a model, so I will have a very frank conversation with that. Are there other people that you'd like me to have a conversation with well.

H

It looks like there's a decent number of them where the right work is happening, but the feature bug updates are lagging sorry. The feature repo is out of date. Yes, basically our checkbox list people are making steady progress, but they're not going back and checking every thing, and we tried to sort of prod people to do that earlier this week, as we hit feature complete, but some still are not in that updated state. Yeah.

J

All right, well, let's I, will be very polite, but firm with as many people as you call to my attention. So I will have a conversation in 10 minutes with Don, hey.

C

Dave David much appreciated you have our support and sticking to your guns on that. One I think it's important. Okay,.

A

Awesome, thank you, and we are, of course, trying to write up the summaries of what is in 1.4 and getting it ready for all of the press releases as well, because that's coming soon with release.

A

So Ryan also mentioned that there was the CNCs swag store was announced today and for those of you who are contributors, there are patches showing on there right now, which I'm going to go scramble and send an email to remove because they shouldn't be there, because they should be only for contributors, and so, if you get there really fast before I do, then yes, the price right now actually is contribute. So it's a pretty high price Justin. Thank you.

A

Yes, so in any case, if you, if you need or want any computer nity swag, we can now, you can now purchase it.

A

And if you send me a code, we don't want you to necessarily have to purchase or if you send me an email, and I knew the deluge of emails now, I'm not looking for everyone to have to purchase it, but if there's somebody who hasn't gotten a t-shirt etc and really wants it send me a note, and I can send you a code because we want to start being able to get these out more evenly distributed and two more people, of course, and have you also you also as the people who are leaders in the community be able to offer this to people to get them engaged in the community and get them?

A

You know rewarded for work that they do etc. So keep me in mind and tunes. Getting a tattoo I see fantastic, it's the Cooper Nettie's logo. All right! Thank you all and I hope you have a great week. I will see you to David's point. Everybody did wonderful work against 1.4 and it really.

C

A

A delight to work with this community we seem to have a very high standard of code, quality, interpersonal communications, quality, fairness for our own value of fairness and- and it is a delight to work with you all. So, thank you so I'll work on temporary tattoos all right have a great week. I will see you all next week, exhale.