Kubernetes Community, 1 Dec 2016

Previous Meeting Next Meeting

⏯

youtube image

►

From YouTube: Kubernetes Community Meeting 20161201

Description

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

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

Demo: Artifactory & Helm; Release Updates 1.4, 1.5 and 1.6; Github issue and notification management; 1.5 highlighted features.

A

Thank you all for joining us today. I will be representing how to use artifactory as a hemp. Depository moves to charge so little bit about myself, so I'm Janie's. They are working as a solution, engineer education.

A

Basically, it was more on and off providing solutions to partner and communities where I do both integration and pollution for know some integrations. So.

A

I'm, so talking about what we did so basically, we created artifactory user plug-in to use our.

B

A

B

A

Physically physically to proxy and guess all the charts wrong, looking storage, where you store all the your incubator and stable charts right so now, when we talk about, let me just do a little bit back and mention that there we are talking about using hem to install the charts or you know, sir, on services, the pods in two varieties. So this is the link to my plugin. So let me just.

A

Learning artifact free and use it as an empty positive going back now. Next question is why we do it right so definitely shapes of the lives in a universal approach for software development. That's why b2b support and manage all the different kind of packages be 10pm power, but docker images rubygems, the cocoa afford a pn composer, which is a new addition, and we are more excited to have common, which is C, C++ repository. So that's why you know we want to include hell repository in artifactory as well.

A

Now definitely have is a great packet manager for Cupid matey's, so providing support for hand Matt's. You know big sigh is priority for us. So when definitely you know, I will appreciate your feedback on this integration, and you know if you suggest anything doing we'll, definitely take it and improve it. So now, how does it work so, as you all.

B

Know what it might.

A

You will be familiar, you know how you can use that to install Jack's, which are basically a configuration dog, so you know starting a pods into Cuban. It is right so now, when we use artifactory as a hand repository basically, instead of downloading your all charts from do from google storage, and now you will be able to use a tea factory as a hen repository to download the charts and insulating cuban eighties. So let's see that in action now, so what I will do.

A

Basically, I will start artifactory as a talker container, and then I will install the plug-in. So I will just quickly show you how it works right. So starting artifactory is a really good thing, so why? Using this command extracted, artifactory pro as a docker container running on port 8080, equally rich will be able to access that.

A

81, so we have a pool, you know welcoming you Viper artifactory,.

A

It will take few seconds to get it out till then you, let me quickly copy my license, which I really need been stalling are defective.

A

Like you see large, and it looks like it's taking because the longer than usual to shout out if I click.

A

Just few more seconds: okay,.

A

Okay and for all, if I fades up and running, and how I will just install the license here,.

A

So our life and live install artifact is up and running, not.

A

So well now I will follow the steps, so basically, I have to buy exacting to artifactory container and install this plugin through to support artifactory as a handy for the tree. So I have my teacher grading. Basically, I will just copy paste the command going into the container moving to flooding spirit, 30 particles.

A

So now first I will download the plugins.

A

And I will be using 12 a month.

A

Ok, so my loving again now it has one dependency on the EML by ml pin. So I will just download that as well.

A

Now forgot exactly: you can reload the plug-in using ray sleepy eye, so I am using this command to reload the plugin and it's showing that plug in water. So I know loaded successfully. So now I am ready to use artifactory as a handy for the tree. So what I will do now? I will do it again bring more repository, artifact I will be of type Jamie. I will name it as a hand.

A

The remote and I need to point it to you, ah q1, basically a URL where, if you know you when it is turning all the end charge so I'm just putting this I have tasted it. So it's working fine successfully now going to article. I will just use set me up physically copy. The repository URL so for he'll hand to taste, and I have mini cooper running by instead of kubernetes cluster, I'm using mini good luster.

A

We can play my mini cooler than you.

C

All right, Chris.

A

D

E

Me guess is not muted hand.

A

Can I continue.

E

Yes, go ahead and continue.

A

So, as you can see my Cuban, it is gone every we need to close up and coming well. I have had mixed all the I one has me here: the tailor in a coupon. It is better I that I have that up and running. So now, next thing I'll, let me add, henrico and I will say, when or authority by three and pointing to my article, which will be URL my article to.

E

Finish: we've got about two more minutes: yeah.

A

Quickly, okay, yeah! So now we have added the repo will do Henry post it then artifactory you go back to artifactory, then you will see that you know Linehan cash I has such a stick in Des, okay, okay and once we do an install part, if I, through that, my sequel with liquid fetch the my sequel from artifact trees, and if you got in mind, if I click help.

A

A

Then firm, vividly he'll, say download and install the no help, all the cash open. It is from at exactly equal to see all starts rotting shop itself and joining, and it has patched the archive for the charts from are. Typically. This is basically simple how you can you know, use artifactory as a hand.

A

Repository now going back to my slides, so I in demo, I I could feel that is, you know: I display how to run artifactory the docker container, install artifactory user plugging to support, hence before type in artifact lee and I added hand, people and install chart in cuban. It is, you know, fetching it from artifactory. So that's what we did today and so find in future.

A

We are, we are planning to have you know specific repo as a hemp depository as a first class citizen in artifactory and as well I have you know what the submitted a pull request to head artifactory chart for q4 hem, so basically users will be able to install. You know: artifact three on hell are using a helmet and install command.

A

So that's that's pretty much. It worked. I had to visit this with anybody. Have any questions any comments? Yes,.

E

Thanks very much Senesh um does anyone have any questions.

E

Alright, well, we know how we can't get ahold of Janelle. Well, you can find me and I can point people to to you or Ganesh. If you want to add your contact info into the notes that that would be a helpful way that people get your.

A

Early aren't is not on this lights, so do not need anything from it again. Beats me on a quick that github and from undef initely I am there in my mokuba. Basically given it is select channel as well can be fed at, but all.

E

The way to contact people awesome all right thanks very much mm-hmm.

A

Thank you guys. Thank you for joining all.

E

Right so on next to what we have on our agenda. um One point four point: seven Jess I think I saw you you had. Are you available to shed if it was different, Jess or maybe just as micro.

B

E

Here all right, no, this one.

B

So 1.47 will be on december eight. So just keep pinging me on cherry picks and I'll. Try to get them in. There was like some willing two issues on that branch that are fixed now, so everything should start going through the queue again anything.

E

Else that you need help on or anything specific, that's happening. One point four point: seven. You want to talk about and now okay cool, alright. Well then we can go on to 1.5. Sud has a list of things all.

F

Right so 1.5 is scheduled for less than underway december. Ageless to date were expected to launch right now we have a total of ninety five issues in the 1.5. Milestone of those 12 issues are considered. Release blocking most of them are getting the attention that they deserve. um If you can, please go and take a look at those issues, see if there's anything assigned to you, um please give a status update daily for that.

F

If you're, not, if you don't have a release issued release blocking issue assigned to you, please see if you can help out with the release blocking issues beyond that for testing manual. Upgrade testing has been in progress. We discovered an issue without grade script, which turned out that upgrade script wasn't actually upgrading, so that testing has to be redone. Luckily, we found it in time and discovered the issue, so hopefully it will not cause the release to slip, but we're keeping an eye on that.

F

Automated upgrade testing has been in the red as well. There was a breaking change made to to cuddle that got merged the fixed weary gommage this week. So that should a lot and we'll take we'll keep an eye on that as well. So testing is completely red right now. It needs attention we're looking for volunteers. If anybody wants to jump in and take a look at this, that would be great. um The other testing suites for Cooper Nettie's 1.5 are looking good. Just a few flakes.

F

um Most of those issues have been triaged that are blocking or in the blocker list. Above the branch status at the moment, master is open for general development. Once again, this means that any 1.5 changes that are going into master need to be cherry picked. Please make sure to put the cherry pick candidate label on your 1.5 PRS that are going into master, otherwise I might miss them. I'm doing nightly batch cherry picks at the moment, but that's no guarantee that your PR will get picked up.

F

It is ultimately your responsibility to make sure that your PRS are getting cherry-pick back to 1.5. So please make sure that happens. I pushed beta 2 last week, 1.5 beta 2 and possibly later this week or early next week. I might do a beta 3 depending on, if there's a need for upgrade testing and then hopefully we're going to go out on thursday with the official release. um Finally, docs and release notes, so we set the deadline for docs and release notes. As this past Tuesday, most of the 1.5 feature is delivered on time.

F

There are a couple that are still outstanding that were pending I, think dad's, 2k and beeps had two outstanding issues. If you guys can take a look at that, that would be great we're hoping to have all the docs PRS code reviewed and merge by ended a Friday. So if you have any feedback on any of your dogs PRS, please address those and if you're aware of any, if you're aware of any action items for the 1.5 release, there is a issue that keeps track of that actually I'll paste that into the doc here on.

F

Please add your action items or known issues into that bug and then Aeryn Crichton, burger who's preparing the notes will add those issues into the release, notes. There's a draft of the release notes that's out from Erina. At the moment you can take a look at those other miscellaneous issues related to 1.5 released the infrastructure. Currently, the test infrastructure has been having issues. It's caused, the submit queue to be blocked.

F

Most of the issues with a continuous integration have been resolved, but there seem to be issues with the with a pull request tests that we've seen cases where the pull request tests will be all green, but the submit you won't pick them up or some tests just simply refuse to run. If you're having issues like that, the current recommendation is to rerun the individual test. That's currently blocked, don't rerun the whole suite, because that will overload the test infrastructure um and if you are really having issues paying the test in frome ain't.

F

A nurse team on on your PR just see siku granny, / testing for maintenance, and, if you want any more details, we're holding burn down meetings starting next week, it'll be daily. The length of the notes is in the dock as well. There's.

E

Also, a Google Group that you can join to get the calendar invitation if you want that and I believe it's in burn down notes. Yes,.

E

Anything else we're good all right, 1.5. Thank you, as always, sod and Caleb for doing the all the work that they've been doing to get the release out. This started as a one person job and is rapidly becoming a team effort. I think.

F

You know what foods to vote.

E

Is yes? Yes, this is very helpful as well. So thank you um which, because we continue on forward even as we are releasing things, the next topic is 1.6. So next week we're reviewing the sig self-assessments that we requested before coupon. Well, we've got four different things that have agreed to come chat about what their thoughts are on the sig self assessment questions we're trying to determine if 1.6 is focused as a stabilization release.

E

So far people have been very positive about it, so that is, that is what the whole idea behind this is so next week we will have the self-assessment under the self-assessment talks if there are other things that want to participate. Please just line up and put your put your sig name in the notes for next week.

E

All right. Anyone else have anything they want to say on stabilization other than stabilization means. Not many new features taking features that are already existent alpha to beta and beta 2 g.a would be fine, but not new, stuff or limited new stuff. All right, um then Brian, who is hiding behind the camera, wants to talk about github issue management and notifications.

E

G

So there is a thread on this on ur burnetii staff to renee's p.m. and communities contribute experience. Working group mailing lists catch up there for the full details, but after 1.5 we're going to make a bunch of changes to how issues in PRS are managed to make the project more manageable and get the issues and PRS under control.

G

One thing we're going to do is create labels for every sig, so a number of existing labels will be converted where appropriate and new labels will be created as necessary. Some of the team and area labels, for example, will be changed.

G

The priority labels are going to be made more a change from numbers to strings that people understand like a critical urgence for p0 things like that.

G

The we're going to try to make sure that all parts of the code map to some 60 tests, Doc's directories and so on us to make it easier to automatically assign sig labels for issues we'll probably to the template and automatically set labels from the issue templates and anybody who's in the communities. Org will probably be forced somehow, maybe by Auto, closing issues that don't fill it in to actually specify one or more appropriate cigs.

G

So we can treat triage as many issues up front or or what we call drought issues as much as possible up front when they are filed, because we have a lot of the proponents of issues are actually followed by us and we're just creating more work by not correctly robbing them when they're created they're a bunch of other things about curating the sets of labels, we have a bot actually that will copy the labels around to selected repos if we need the same labels and multiple repos that ba is currently disabled.

G

While we rearrange all the labels, in addition to fixing the the labels, we also need to get notifications that our control, because it's completely impossible to follow notifications on the project for most people now I think so. The idea is to create a more targeted teams, github teams for notifications, so that different kinds of notifications can be sent to different sub groups of people, for example, bugs features request proposals, PR reviews and API reviews can have distinct notifications and so that people who are not in the github org can follow.

G

The notifications will mirror these into google groups. I.

G

Think that's the that's. The bulk of the high level description like I, said for the for the details. Please see the mailing list if you are willing to help implement this stuff, please reach out respond to the to the thread or send me email.

E

Awesome. Thank you much much progress. The labels thing something I know Garrett started on a while ago and that's awesome to see it come to fruition all right up next on our agenda unlisted. Does anyone have any discussion on this other than more go team contributor working group of contributor experience working group, no more discussion, okay, cool, thank you and thank you. Dempster dropping the timeline into the chat. That's awesome we'll put that into the notes as well.

E

All right then, evar and aparna are talking about 1.5 and what we are going to be putting putting forward.

H

As the feature the exciting features, when did what is 1.5 features? Sure, actually you are, I think, if he's online, it's gonna kick us off all right.

I

Yeah I'm on lamb, so I'm sure must cream is debrief, slides to get in and then this stuff it we have prepared. So I heard his brief report. Forgetting what's happened to his wanted 5 release. So is the feature your lot good, 5 release so I, just just a brief recap. So we have started our iteration on September 19. The feature phrase has happened on academic /. 10 future phrase means did before this date. Everybody was too loud to propose a new feature for this iteration. After this.

I

After this date, bony fishes except except a flag. Some exceptional one, wrist specific process so november. Seven is the deadline for development for active development features, and after this date there are boxes in a stabilization process happens. My maintain is the date when better releases and we expected a general availability of 15 on december eight. So this is the slide that has been copied from the initial session of planning one at five, so human notice, the main features that have been discussed during the session and being practically like krenek heads.

I

So most of these features being implemented in 125.

I

So are you might find older features, the targeted 105 on github in the features repo, you know, I simply saw them with 105 milestone and helo final yeah you're going to filter these features to get in the stage did they're targeting so alfabeto stable. You can aerosols labels. What else I pre created and all the features are labeled with their tonight also, as you remember so, I've been maintaining the future tracking spreadsheet during the press, release and I'll sedated during this release and I appreciated by the help of community to populated stuff.

I

So you have the feature striking spreadsheet. It includes all there, like all the features that tracks all the features that have to be put into this relates. We also have here's our list, no descriptions and add us cents for sigdoc scream for help. And yes, so in general we have 26 new features: direct well've, alpha, 9, beta and five stable features.

I

The most notable features that we have increased release that fateful stress targeting data. ah We also have the user implementation from signals in the user implementation kyrgyzstan alpha state. We have lots of new features from Federation. It's like reflects deletion state of this product, this project, as like, really solid in and being much more important in kinetis itself. We have also chase it up for mastery, commanded and too I. Don't know why there is no link, oh well, so we have also been your support, implemented by a parameter, I, prenda and sig windows.

I

G

We describe the CRI in terms of something that's relevant to users like does it enable anything in particular like upstream rock and support or something it's not a few.

I

Well, is there anybody from signaled who has been working on this feature and can describe days with his more details, yeah.

G

Okay, anyway, I think I think it's. You know it is very important, but it's worth thinking about how we can describe it in terms that will, at least at some point have impact users.

J

But yeah, that's that's, probably a bad point. I, don't know. From my perspective, I thought we always said that something needed to go into the features repo if it was something you were going to brag about a blog about, and here the idea is we're setting the stage for container runtime abstraction and we're just sort of slow rolling. This for doing this right now for docker, but it's going to enable more things later.

J

So it seems like it's something worth calling out to cluster operators that it's a notable change and how things are operating, even though it doesn't necessarily provide anything to an end user who is running their applications on turbidity straight right.

G

I'm not saying we shouldn't mention it that we just need to frame.

E

G

Just reframe it, okay.

I

Yeah, um so in coming back, so did it for, like the most suitable features that we defined your credit management profession that happened in the beginning of this week, so feel free to suggest you all. Your social gettin is also. We had subscribe, like the next whatwhat's going to next for the next iteration.

I

So we had our 2017 roadmap session at cuban actors, dev summit, where we have briefly described our goals for 2017 and your partner, and I are going to present this stuff on the next community meeting on the next week and also I am expecting to receive some feedback during one at fibrinogen. I hope you'll have it in the nearest time so I'd like to hear some feedback regarding this process, and how can we implement to implement it, make a perfect India in the future that most low from me.

E

Excellent, a partner did you want to add anything nope we're good, look fantastic, um so this leaves me I, don't know how you teed this up, quite so perfectly for me- or this leads me to the question of when do we want to do or 1.5 retrospective retrospective? We generally have done them two weeks after launch on a Friday. How are we at an eighty percent confidence rate on the 8th side, seventy-five percent seventy-five percent confidence rate for the egg?

E

Okay, because two weeks after the eighth on a Friday is actually the 23rd which is not cool, so I'm, thinking that we might want to pull that in and either do it on the twenty-second. Oh wait, I'm not! Actually even oi will be on the tray. Second, um on the 20 seconds at be converted, East community. Being on that Thursday, we can do it that day, I think waiting until after the break loses us a lot. What does anyone else think on this 22nd?

E

Anyone care I'm.

F

Thumbs up for the 22nd I'm happy to facilitate again as the community one side thanks.

E

Jason, your facilitation has been great fun.

E

J

Well, just a question: where does the 22nd please relative to the start of 16 or 16, released I felt like the problem we had with the last release was that the retrospective in the planning process overlap so much that is unclear to me how much feedback from the retrospective actually made it into you. The next release well,.

E

We are at some level starting the 1.6 planning already and we can't do the retrospective until after we release. um So we could pull it into Friday the Friday, the 16th.

F

Might be too soon might be to 2-under on thurs yeah.

E

With a seventy-five percent confidence rate on the eighth I, don't know that we could make a good case for the 16th about.

J

Twenty second is fair.

E

We could also do earlier that week, but we already have this as a meeting and my suspicion is. We will run low on topics, as everyone is rather more interested in heading to the holidays than thinking about what the Cooper Nettie's community meeting topics will be on the twenty-second. It's just a guess.

K

Can we take that slot at ten am on this 22nd yeah.

E

That's exactly that's exactly what I'm proposing buggy.

A

E

All right, then I will. I will make that the 22nd, the retrospective for 1.5. That gets us a little bit more time, but not a lot and hopefully we won't slip and then we'll be good to go aaron. Can you speak a little bit more about the feedback that you think hasn't necessarily made it back from the retrospective and he's when we, when we did the retrospective of the retrospective and Jason, went through the 1.3 retrospective notes and talk through the 1.4 release?

E

We actually had hidden, many of them, so I'm wondering 1.4 1.5 will do that as well, but I'm.

J

I'm gonna have to get back to you on that. Okay I can sort of start off the retrospective with that, but like yeah.

B

J

For me is the features repo as well as fantastic as e horse presentation was, if you click through almost any of the links that he's got there or to any of the features that are in the 15 release, they have nothing checked off and we difficult to tell like what was actually going on and I thought we'd just discussed how, like maybe I, wasn't quite working, but we've limp along with the spreadsheet again, and the spreadsheet has been wonderful yet again, but I really think we need to take it step back and figure out how we're gonna continue to organize this stuff, because I don't think it's working at all, but the broader list of what didn't actually get addressed.

J

Let me get back together. Yeah.

E

That's fine dance.

G

So some of the issues are the threads for contributor experience touched on this as well. um We definitely need a much more concise.

B

G

And we actually need to require that stuff to get build it and just auto close things that are not filled in right. Like I, think there are three or four fields that are absolutely necessary to fill in, and we should just trim it back to that and provide a document which explains the process in more detail, rather than putting all that in the templates.

E

Ok, so to number one features: Rico and making sure, as we did last time, that we start the retrospective with a retrospective retrospective.

E

Awesome. Alright, we are short on agenda items today. So does anyone have any topics or do we get 25 minutes I.

J

Ask something that came up with 6 scale around the submit q there's a concern that the backlog in these of the queue is so large that basically, the backlog represents the lost the rest of 2016's work, I'm curious. If there's anybody else on this, call that feels that way and sort of I guess I ask is like. Is there a concise description of what the list of priorities are like how this would make you besides?

J

What's emerged in front of something else, because I heard during like the release down we're now meeting and saying six scale, the proposition of life? What if we just started bumping priority labels- and there was discussion at the release- burn down meeting that is bumping of priority labels is getting in the way of like effectively cutting the 15 release. So it is like a list rules. We could go look at just to understand how this operates.

G

K

Question, yes, is the answer to the question the.

B

K

B

K

And go to the help tab, there's a tab that says ordering rules yeah. When that it will tell you the rules that the body uses to determine order.

J

Step. Thank you. That's excitement, question.

E

The answer is yes, so.

G

Where's that earthy I think it used to be on an info tab. Yeah. Oh.

K

Crap, my back on you Sam the.

G

C

G

C

Hi this is dims. There is a batch thing that went into submit q couple of days ago. Can somebody speak to that as well?.

J

Nope, that's what brought to the discussion in scale is I.

D

Somebody's somebody.

B

J

Like ask the pot to testing 17 times in order to get something to actually go in yeah unclear like what the deal is. Yeah I.

C

Did find a UI but I, don't know whether this is going to be the final resting place for this batch results so from what I can see in that UI is like three or four PRS that get tested together at the same time, and then they get merged. So as yesterday there was something saying: don't treat us things that are in the batch queue and went hunting for it.

G

So back on to the topic of priorities, please do not abuse the priorities. There is a PR I, think it is open or an issue at least or something about a favoring milestone over priority. I think it's already the case that milestone is favored over things without a milestone, but.

K

It goes priority. First, then, milestone with a date than everything else. Yeah.

G

So you know things that absolutely need to be in 15 should be tagged with the 15 milestone. Please do not add priorities to anything else unless it actually needs a priority in order to get merged ever right. So traditionally, we've used priorities for doing things like breaking cycles of read a cell for pr's. If we use priority sparingly and the right thing will happen right, things with 15 will get merged, preferentially things that need to get a bump to the head of the queues. They don't need to be rebased again will get merged.

G

Otherwise, and everything will be fine eventually, you know we have a backlog right now, just because of the freeze once backlog clears. We should be good again.

K

Please detecting the parties yeah also when, as Brian talked about earlier as priority numbers changing to name. That means that those names will have no impact on the submit you. So the issue that I linked was the one he just mentioned talking a little bit about. What should the ordering be? Has a tiny bit of discussion about as we move into a name world? What mains names make sense for PRS first issues, because Brian's discussion has been largely around issues: yeah.

G

But I think priorities for names for PRS would also be useful. Like you know, if he would this, this PR has been rebased too many times or something like that. Then we can control the policy. You know people grow up, provide the information and we control the policy yeah.

K

I submitted it on that issue that I just think over and check for people who care some names that made sense to me, but likely those aren't the names that actually makes sense.

G

E

Okay, I want to jump back to joe's comment in a chat if we're done with pr's and queues and submits and priorities and actually in fun. Okay, so Joe asked, on top of topic of future releases versus stable bugfix releases. Have we really done any long discussion on this video JS communities onto a model of even number releases being stabilization? We have sort of tried to with some success to do that, and that is the plan and the idea around the 1.6 release being stabilization release as well.

E

It's moving to that sort of model so that we are making sure that we are moving features through alpha beta 2 g.a, making sure that we're we're doing we're doing work that pays on technical debt and that we are listening to our users that they want things that are that are long-term, stable and maintainable. So we have had some discussions about that at varying points. We've not been identica bout, it or restrictive about it. It's been it's an open source community.

E

You can try to whatever you want and in feature, but the target is that even-numbered releases are stabilization releases. So that's why we set out these questions to the states. This thing's then had to a self assessment of where they feel they are. As an example, 6 storage did a lot of work off. Cyclone put big changes in 1.3 and yes, I did that right.

E

Okay, one boy, three put a lot of big changes in 1.3 and took 1.4 and 1.5, and did a bunch of stabilization work in in that because they needed the work done. So you know the things are where the control for these these feature choices are happening. So that's why we've asked them to do a self assessment of where they are and if they think that this needs to be a stabilization release for the different six.

E

That is our your question enough Joe, yeah.

L

Thanks Eric that ain't my question: I am I, heard a comment about 1 dot, 6, potentially being a stable release, and it's you know if we as we look ahead and we should discuss whether that's something we want to formalize, you know more concretely or not our experience with enterprise customers is you know the stability is really becoming a more important than some of the features as they're moving production clusters and in cases a little production cluster, so yeah. So thanks, you answer my question. I'll.

M

Say I'm really concerned about those every other release being a stabilization release, just even as a concept, every release has to be stable. What we're setting ourselves up for is, if we start saying every other releases, the stabilization release we're just cutting out huge chunks of people who are going to say. Oh well, odd-numbered releases I'm, not even going to consider using it, I think, is this concept needs to be pretty carefully discussed. I yeah.

N

If I can just weigh in there nope, nobody has committed to anything like that. Nobody is saying that every other release is unstable, absolutely no position whatsoever.

N

This is just so happens that in 16- and let me use this as a moment to plea to everyone- you take the time over the next week to talking to the cigs that you care about, and things like that, to make sure that everyone is comfortable with with, what's coming in 16, that there aren't any major features missing, but our goal is to really minimize the number of new features. So to your point, Bob Owen is saying that that every other release is going to be unstable.

N

All we're saying is that in 16 we are taking some time to to do some stabilization to move some things from alpha beta and forward, but we are a community if we decide that we need to do this again in the future, we're more than open to that. No one, you know. No one is saying anything here. Basically, I have.

G

A couple of comments, one is: we need some mechanism, and this may not be the best one and I'm open to talking about other possible some mechanism to force everybody to spend more time on project health issues that have been neglected for a long time, and if you know the easiest way is to say we're not taking any new route, new features that are released and that's what we'll do if there's a better way we'll do that we get feedback from a lot of engineers on the project from multiple companies that have a lot of pressure to deliver features, and they don't have mechanisms to push back on their management about that.

G

So we need to provide some kind of stick that they can say. Well, look there's nothing I can do. I have to do this in order to get our features in eventually. The other thing is, we need to create some breathing room for all the the leads to not have to just review feature proposals all the time.

G

So we either need to rate limit the number of features that can go into every single release or we need alternation or something like that. Otherwise, we can easily spend a hundred percent of our time just entertaining proposals.

M

Just for clarity and I know: I'm I know that those of you have heard me make comments on this before now. I'm being consistent on this, which is I, think every release needs fewer features and more stability. I agree.

G

With that, though, we just need to figure out a mechanism to enforce that.

N

Agreed and again, no one is saying that this is the pattern going forward. All we're saying is that we're taking 16 look, let's just be very pragmatic. Let's take 16, let's focus on that in this case, we're going to do a lot of the things that Brian just described, to kind of find our footing again and then and then move forward. I like the idea cause.

F

Taxonomy, we just say heart instead of stabilization, because he like there's an inherent mount of stabilization, that's necessary for a team and pass the test necessary. You can promote it into the branch, so hardening sort of is more we're trying to make it in a more robust feature, as opposed to a non-functional or a functional feature. I.

E

Think the language, maybe what we're stumbling on here more than so maybe it's a feature: release versus a non feature, release where feature is user facing quit anyway, um and.

N

I think we ought, I think we all agree that naming is hard. Let's try not to do it here and use up 50 people's time. One.

E

Hundred percent agree all right uh anything else that we need to cover. Let's check a chat quickly, just.

J

To put a voice to what we were talking about in chat, I think that Clayton suggested like if we could focus on what you can focus on what core is and then say that core is the thing that gets hard. This allows us to enable more people to develop their features outside of core on their own, and we've always already kind of done this in the large, with the Cooper Nettie's incubator organization and projects that go in there right. So it's unclear.

J

What's the difference between a project in the future and if we more clearly define what core is, maybe that helps us sort of enable the same models in the small and then with their slight by define it by tightly defining cord? That's the thing: that's got to be rock solid and stable. That doesn't necessarily mean your feature has to be rock solid, stable anyway. Just to put that out, there.

E

A bit of summary, thank you and Josh reminds me very graciously that and asked to anyone who took notes at the Gubru Nettie summit. Please contribute them to the community repo and we'll start linking them all together, because Joshua correct, I have not gotten around to pulling them all out of mail. I think most of them got sent an email to one mailing list or another, but I haven't had time to pull them all.

E

So if you took notice that the developer summit, please please put them into the community repo and we will organize them up there. We.

H

Had a huge session on, we had two sessions on roadmap planning and your has the notes. We're planning to present members rating.

B

Next week,.

H

And then ours all.

I

Right now, where she discussed, it was just being slack so sure person like crawl, our own, oh so I'm going to share them with scheme, and we also shared up like the polished and ready to use for the people.

D

Version on the next commitment, perfect.

E

All right, anyone else have topics or notices.

E

All Right see you all next week where we get to hear the 2017 um planning work that happened at the dev summit and hear more about 1.5. All right talk to you soon.