Layer5 Meshery Build & Release, 23 Jun 2022

Previous Meeting

Next Meeting

⏯

youtube image

►

From YouTube: Meshery Build & Release Meeting (June 26th, 2022)

Description

Meshery Build & Release Meeting (June 26th, 2022)

Agenda:
[Ashish] Update the compatibility matrix with multiple k8s version and other changes () ()
[Aaditya] Add workflow to push supported components onto docs.
[Hersh] Replace kubectl apply with meshery pattern apply/app onboard
[Ruturaj] Blog post on (see layer5.io/ #blog_kitchen channel in Slack, too). See
[Augustas] Open PR in istio adapter
Absent today. PR is under review.
[Mario] Cypress tests updates around Kubernetes multi-cluster suptosport.
Absent today. Rohith to engage here as well.
[Sekiranda] PR on Meshery Adapter for Consul: w
[?] other topics today?
[Lee] Minimizing the number of files in the root directory of Meshery #3210
Proposal to restructure the top-level folder structure of meshery/meshery by major architectural component

A

uh Hello, everyone welcome to the build and release call and, uh as people keep joining in, I will just start with the agenda item so that yeah yeah, so that we don't lose any more time. So are already five minutes past. So I'll I'll skip to the second item. First, which is uh about aditya's stuff, about adding workflow, to push supported components onto the docks. So I'll give a quick overview about what the ask is, and if anything you have any doubts, then you can ask about it and uh okay.

A

Okay, so, uh basically, like we have this compatibility matrix, which shows whether we are basically the test. The end-to-end tests are passing or not. Something similar to this table is needed and uh that the data I'll tell you uh where the, where you're supposed to fetch the data from and then how the table is going to look like and how the workflow. So there are two parts. The first thing is to first fetch the basically write a workflow to fetch the data from these repositories. So I'll give you one of the flows.

A

So, for example, we have so to get the components for istio adapter. What we would be doing is that we'll come here in this repository and in this repository you will find temp inside of so the path is same for all. The repositories templates, slash, ohm, slash workloads, so here we have all these versions right. So what we can do is that one of the one of the columns, uh the right most column, can be components and what what that can have is the basically sorted. Basically, it will show the latest component.

A

So, for example, here the latest component would be let's say 1.14.1, so the it would show one point, one four point one, and when someone clicks on that it would it would be similar to clicking on you know uh when you click on this particular uh folder. You get all these uh components, so you have to extract these names from here. So, although these names aren't um you know, these names aren't uh formatted properly so yeah.

A

If needed, we can format them as well, but yeah, but yeah they ask is that, as you can see, for each of the components we have here, definitions and schemas, so all you need to do is just take this uh uh in the dot separated uh the string. You have to take the first string like authorization, policy, destination rule and onward filter.

A

You can either use the the name of this file or the name of this file for reference, and you have to display all of these components under that particular version, and you have to do the same for all versions so for uh the default would be for the latest version. We would be showing like for right now that by sorting that latest version is 1.14.1, so when user clicks on that, it shows the number of components for this particular version, which in this case would come around.

A

I guess 14 or something and users can then switch between versions and they can scroll down, and it would be just similar to scrolling down uh this particular. uh You know this particular this particular uh directory and uh yeah. You will you'll have to do this for the uh that. So these uh things can be done for all the adapters, and then we also need to show the supported kubernetes components uh for supported kubernetes components.

A

We do not okay, so this was one of the things that was, uh that was on the to-do list to actually have statically stored, those some very common communities, the components like pods and deployments and services, and we don't uh have them right now anywhere, because, mostly when we connect to in right now. The case is that when we connect to a kubernetes cluster, we on the spot generate those components, so uh we don't store them statically.

A

But for the purposes of this particular uh this particular docs uh with this particular table, we can actually have those somewhere. So you can uh generate those components. I will help you generating those components, the community score components and uh where to store them. uh uh Okay, at least here um so he can have. I think he can give us more perspective on. Where do you think we should store those humanities components if we store them statically for purposes like these.

B

For now, probably within the mesh remeshery repo, maybe under there's a little bit of renaming that might be kind of nice anyway, I think there's there might be two folders or even three folders that could potentially be consolidated if my memory serves so because we've got things like om, okay, good. This seems like a natural landing spot for as kind of the answer to the question that you just you just asked so, okay, but but we're not strictly adhering to ohm, and so maybe we would want to rename that.

B

Okay, if you go back a couple to the root, folder, there's also other potential. I mean so like.

B

That's definitely an answer for your question, but if we look at consolidating a little bit uh and it's kind of nice to keep to consolidate to like help just organize and structure and organize the structure over time, so it's things like internal models uh and then those seem like candidates, there's other things that are kind of uh handlers and help like handler seems like those could go under wherever the rest of our golang is, if that's not where most of it is, but there's kind of two ui top level folders, there's no top level server folder.

B

There is a top level docs. There is a top level mesh free ctl. So if you, if you were, if your mental model was gosh, we've got the the way that we speak about. Our components is like mastery, ui mesh, redox, mastery, cto, srcli, actually server.

B

Then maybe we would have top level folders to capture those and for the these models. These components.

B

You know you there could be an argument made that they would. It would have its own top level folder, because it all over time it wouldn't just be kubernetes, but we would be generating components for spiffy and spire argo and jaeger and all of the ecosystem things because we're ultimately going to let people design cloud native infrastructure like any infrastructure, but for now you it could. That could also be consolidated under a top level like server folder.

B

So more of an answer than you'd wanted, because there's a little bit of work inside there but yeah. Maybe a proposal to put into a for the restructuring.

C

Maybe maybe a proposal.

A

B

Put forward for you guys to comment on, but immediately it feels like. What do you think does this feel like a pretty good fit.

A

um Yeah yeah so about the directory consolidation. I have taken a note on that I'll uh I'll see what are the caveats of it and I'll get back to you about that and we'll get rid of this om all together. I think we should be. uh I will create a pr by this weekend to get rid of this. What like? What should we just name it components or if we get rid of this thing, because this is just a directory name at this point, we're not using it internally anywhere so.

B

The other one that strikes me is like yeah components. It's a great suggestion that that would be hard to argue with there's this other and granted. These are two different things, but but there's another folder called models which also feels like a fit.

B

So, hey something that can be changed over time too yeah yeah.

A

So yeah so aditya for now, I think uh we'll rename this later, but for now I think we can uh have a directory name like lee said we might have started components for spiffy inspire and for other stuff as well. So I think uh this we can have those uh thing in separate, so basically the these app. This application is measuring score component, so we are the ones generating schema for it, so we are not pitting putting it up.

A

So one thing we can do is we can create another folder code and then put these two files, so just leave them for now and create another basically create another folder kubernetes and inside of kubernetes folder.

A

We can have some components like pods and deployment so uh and in that table you what you'll be doing is just like in other uh repositories, you're going to template some workloads here, you'll be going to home workloads, okay, it's here and here you'll be getting all of the kubernetes components and uh the one thing more that, instead of having uh is instead of uh statically doing this thing. This can be one good. We can have a workflow that does it every 24 hours, so it starts at humanities cluster.

A

It takes their components and pushes them here uh so I'll create an issue for this. For now we can manually have this so that you can get along with the table and later on, we can have a workflow which does that so that we don't have to update it manually.

A

So have you understood the ask here about the showing the number of components.

D

uh Okay, so fetch, like whatever the versions are there in the uh particular adapter depository, and when a user clicks on that version, we show all the uh click com, like everything that is inside, that folder in the deposit right.

A

Not everything so we have for each of the component, like authorization policy, we have a definition and a schema, so we don't want to show them everything twice, just take that authorization, policy and destination rule and so on and so forth.

D

All right and uh for the design of the table uh like how do we want to show them like uh we can uh like collapse? Everything like mention mesh measuring steel, measuring linker, shady kumar, everything collapse it and when we use the clicks on one of the component, it opens up and shows all the version, and when again, when a user clicks the version it shows the component, can we do like that? Or will it be too long of a table.

D

B

Oops yeah we might have, we might have lost him for a second, um uh if you like, maybe just a quick proposal in a spreadsheet, just because it's kind of a tabular thing that you're looking at um it's, it would be hard to have too much information.

A

Actually, sorry, I just broke off for a while. So what were you saying.

D

uh So I was talking about the design of the table like how we wanted. I was thinking. Maybe we collapsed, all the measuring components like measure html machine, kuma, machi linker, and when you click, when the user clicks on any one of those, it opens up and shows all the version. And again when the user clicks one of the version, it shows all the components.

A

um Okay or we can have like, uh we can have the uh one column named component, which would have um just meshily, maybe that the core combo application thing that we have and then kubernetes and then linker, dst and all and one column can be supported, components and there you we only show for the latest component by default, so that we don't clutter it and then, when user can. Basically there is a some kind of a down button there.

A

So user can click on that and scroll down and select different version, and then we will show the number of components for that version and if, if they click on that particular thing, they'll get the actual component. Since the things that we were counting and showing them like the it would be, it would uh basically expand into authorization that list just like okay. So let me share my screen. First.

A

Yeah so just like we have here, you know when we click on something it expands into these test results, so it would be like when they click on so here we would have that version and when they click on the version, it would be like authorization policy, destination rule and so on and so forth, and they can change version from here and- and here we'll also show the number in bracket, maybe that that would be the number of supported components for this particular version and on top level we can or maybe uh somewhere around here we can show the total number of components.

A

Basically, we some aggregate over all of the versions.

D

And you user will be able to select those version like from yeah yeah. Okay, uh I think uh I'll make a design in excel then and show it in the ci. Then we can. If it's all right, then I can start implementing.

A

Yeah yeah, I think uh once we have the once, we have the workflow that fetches everything for reformatting the table is the easy part.

D

A

D

Right now, like uh it's, it's gonna be manual, work right, the previous one. We don't have any workflow for it from now onward.

A

The okay, so um the part where we fetch the stuff won't be a man should not be manual. So uh the part where we actually go and basically download all the I. I don't think we should download all the files there would be. We can use github api to just list the file names from templates, home workloads and there or there on. We can go forward so there would. There can be a small workflow that kicks in.

A

uh Maybe I think it should kick in and I guess in each of the adapters and whenever a release is made or whenever whenever emerges, whenever something is merged, and uh you can basically, you know the inside of the path ignore. You should only kick that when something changes inside the templates directory, so that every time uh because we have already have a pattern component generator which kicks in every 24 hours and generates new components for us.

A

So when that kicks in and when that uh changes that temp to make some changes in the templates directory, then, is when we want to kick this new workflow, which will actually go and extract those things and then push them into the machine core repository the names. So it would be like it can be a json file that can be pushed into, uh or maybe inside of the workflow like we are doing with the end-to-end test. We can on the fly, create that markdown and push it directly instead of pushing a json or anything.

A

Okay, if you need any help, it will be. If you need any help, help with the uh workflow. You can ping me.

D

Yeah sure yeah, I think I will need help but yeah.

A

Okay, so I think we can move on to the okay. So the second item was about this and the first item was about uh multiple gates version and okay, so about this thing is I don't think we have the folks who, okay so in the last call few people uh volunteered for this thing, but I don't think we have those people right now, so I would mark this thing is stale and I would believe that uh they are not on to this thing so I'll have to uh regurgitate.

A

If someone wants to pick this up so right now like uh we have, if you click on any of these test, results you're seeing this uh version with this community's version- and this is actually a stale version- we are not. Actually the tests are not running for this version. So uh to give you an example, if I go over to like uh instead of istio repository, if I look at any of the action from which this thing is generated, so it would be right around okay, okay, sorry, I'm not in sd.

A

A

I guess it was yep this one if I okay, so now, all these test results are actually running for each of the different communities version. So here, like uh one uh ran for this version, one run for this version and these versions are passed from the core workflow that we have inside of my stream messaging and instead of reflecting those versions, we are still getting uh we're still getting the the static version over here, which is, which is to say the 1.20.11.

A

So the slight changes would be needed that so what happens in this particular workflow, which runs over here? Is that it uh after it there is this update, docs job and what it does is that it send it. um Basically, this is the one responsible for creating the particular the markdown, which is then committed.

A

So here the way this markdown is created is we first take this json that was create and from this json, as you can see here, this thing is hardcoded, but but this, but this shouldn't be hardcoded, actually went over the whole thing and how this thing can be corrected in the last call. So if you need more references, you can look at the recording of the last call.

A

So the ask here is to get rid of the sk8 version and instead of this thing have basically go when the exact kubernetes version under, for which we run the tests. We need to have that version, so um we need to have that version as the key and the rest of the results as the value and then uh so one can be. One point two two point: two uh I'll show you the core workflow, where we're actually setting it. So you have more clue what I'm talking about so.

A

Here you can see uh yeah, so you, as you can see here, we are actually setting these three communities. We are testing for two one, one point two, and uh we should we should be having these but uh three uh things over here, so it would be so when you right now, when you expand this thing, you only get. Basically you get these things expanded and you only see it for one version. It should be more like I I'm not sure of the exact flow, how that would look like.

A

But when you click on this, you should be getting these results for each of the communities version and right now it is for only one. So um after you are done with that, I can help you here with the uh with the markdown stuff. So I will we can talk about it on slack about how we want the markdown to look like and once I'll push the marker down. Can you update the compatibility compatibility matrix to reflect that.

D

Yeah, I I I yes, I think I can do that. We can have like uh three three cells in the open thing and when you click and use it on the cell, it opens, expands and closes other.

A

Yeah yeah yeah that kind of behavior, yeah, okay, so I'll uh since folks, who took this up last call, I I'm hoping no one is working on that, otherwise they would have shown up uh so I'll, be working on this thing and uh helping you at it when I, when I'll be making changes in the uh in the workflow, okay, okay, so moving on to the next item, uh harish is hershey: okay, yeah he's here so hersh. Would you like to talk about your item which we were working on.

D

Hello, can you help me.

A

uh Yeah barely.

E

Hello, hello, hello,.

A

Yeah, no, no, no you're, audible, yeah.

F

Right so I'll share.

E

Let's share my screen.

E

uh You can see my screen right.

E

So ultimately, the idea was that you know we change uh so in our in the measuring smp action. uh We were basically using cube, cut and apply instead of you know, using what comes with measuring itself measuring cdl itself.

E

So that was the goal. uh So, uh firstly, uh you know what what I was uh able to observe right was that uh it was. You know it's just very silently, throwing an error if you see here uh right uh for pattern, apply right. If I, if I do this, it's located in the it's located in the same directory in the same working directory as we are right. If I just run a meshley ctl pattern apply emoji, auto dot, then it just very silently gives you 400 and exits.

E

So even when we're you know using, let's say a github actions right so in the logs we just simply get this status code 400 somewhere.

B

E

Just silently fails right, but whereas where when we do a measuring serial app on board right, so this actually on boards, so ultimately the idea is that mesh. You know it's not a pattern. Fine, emoji water is not a pattern file. It's it's sort of a cube, kubernetes application file and an app onboard is the right way to do it and that's okay right, but actually I I sort of had a little bit of a break, and I uh it took me a little while to come back.

E

You know to understand what was actually going on. So actually this actually bugged me a lot. uh Do that. To that end, uh that you know, why is it? Why don't do not? Why don't we have a better response for this, so in measuring?

E

I think I remember in this uh system logs I do remember you know it's stating that you know it cannot find the name at the top of the pattern site so and in the pattern we have the name at the top and if we do not supply that particular name so so that was when I found out. You know that this is what was going wrong with uh with this right, so so first first was this uh then.

E

uh Similarly, you know this is when I found out- and I said that you know uh istio is a pattern like the bookstore like uh sorry. The book info that we have is actually a pattern and the rest both are sort of would be uh imported by app on board. So this is when I wrote this uh then like. Ultimately, I have progress on this and I have everything running, but it's just that stu currently in this particular action runs on patent apply, while the other do that we have a linkedin osm.

E

uh They basically run on app on board right. So that is how we have made uh this snp action to be right, and then there was this. uh uh One thing uh is that what I also found out that you know even app on board, wasn't always working the first time around with linker d and osm. So what was happening there is that you know I I had. I had run this so ultimately, what we are doing is, I can actually show you one odd run there.

E

uh I think it it will help you, you know, understand uh stuff, better right. So what what we basically do here is that we uh install the you know we install the service mesh. We do everything else uh we installed even that onboarding application that you see right. We have onboarded the app as well and then we open a teammate session. Just you know see. If anything you know wrong is going on right, so in in a few sessions you know what I also found out that when I did a cube cut will get pods.

E

I mean this is just as a references, so I I wasn't getting any resources like I wasn't getting any ports at all, but when I did the entire steps again right, you know that I deployed the adapter again and did everything else, because these are the steps that are there in the action itself right, but when I redid them and then I did cue petal get pods, I I was able to get these.

E

So that is another thing that I raised, that you know it wasn't working reliably. It wasn't working the first time around uh right so that that was another. Another thing that I have pointed out uh and then after that, after that is mainly, you know, a chat regarding the issue itself and what what other things just a couple of review comments. So this was mainly about you know uh that particular issue. I I hope I hope it's so right now right now. What I would say is pattern applies.

E

What we are doing for sdo and uh app on board is what we're doing to the other two, but even that isn't coming out to be exactly reliable uh yeah. So this is probably how I sum that up.

A

Okay, so with the last thing were you talking about that so with app onboard? Is the thing inconsistent, so sometimes it works. Sometimes it doesn't or exactly.

E

Exactly I'll what I'll do is I'll? Actually, uh let me let me show you this right. uh So.

A

Basically, the same steps that are taken place in that particular action. When you do that manually, then the thing works, but here it.

E

Is right right right right, uh so I think, where exactly is that? uh Let me see so this is where we install mesh day. This is where we'll install the service mesh and right. huh So this is where we install the service mesh and everything, and uh you know we- we do a perf reply, so this is this is where we do uh uh everything and- and as I showed you here right in the earlier sorry I'll, just go back a couple of these things right so.

E

So, as I showed you right so ultimately, at this step we were able to uh what we do is we ultimately have a service mesh running and then we have an onboarded application on top of it as well right, and then we initiate a teammate session. So this happens every time right so, but when I go inside when I go inside that node and I try and you know see what's going on there- I do not get parts at all, so this you know reflects that.

E

You know that is the behavior is inconsistent and when I do the same steps, these are the same steps that I copied from the you know the same files. When I do these steps again and then I run this, then I then I get the points up and running.

A

E

Right, so uh that's that right, super you had. I think uh you and I guess you- you said that you know uh that we can. uh You can help, convert those into package, so that was one yeah.

A

E

A

This inconsistency is not with pattern, apply right, no.

E

No is that yourself yeah, it's only with that. One.

A

Actually, the other message with already I'm saying it's: it's still like a problem to address right if app on board is not yeah.

B

A

Of course, of course, of course, of course, so, okay, so I'll create an issue for that or uh if you mind you can, after the call you can create an issue for this thing in the mastery score repository, although this is related to the smp action.

A

But um there might be this problem with the app on board, so so just to clarify you're sure that the same two commands that you run these same two commands are being run in the workflow itself or is the workflow trying to do the same thing with different commands so that.

E

I'm pretty sure.

A

These are the same ones, because I copied and pasted them: okay, okay, okay, okay, I get it! I get it so uh so until this app onboard inconsistency remains you're stuck on the spr right.

E

uh Yes, but it does work some. I I mean it's it's it's, it's sort of a double export right I mean I am, but it does sometimes work and I have successful tests on it, but those really don't matter until we fix the app on good issue on the other side. If we can move to pattern apply, then that makes sense for me as well, so both so any of the any of the two fixes make sense.

A

Okay, so uh have you run okay, so have you opened design configurator up until now? uh No, I never like so yeah. Do you have a local instance of messy running right now.

E

Right now, no, I don't have it.

A

E

I'm using my name, I'm using another laptop right so that that's what was it.

A

Okay, okay, so yeah, it's just three steps. So there when you'll go when you'll, open, measuring and you'll go into app, there is an applications page. So when you'll go there, you can import an application so yeah! Okay, thank you. Thank you lee for sharing the screen, so there you can see inside of application. So you can. Your application can be either your docker compose or your man, your general communities manifest.

A

So if you click on import application, you can basically import a local file or a remote file, whatever it is and uh when you upload it uh you'll see that communities manifest over here. uh So you know, or you can so, let's say that yeah. So can you open one of these uh files yeah? So this is your general communities manifest? Can you can you put the networks tab.

A

Okay, so if you just open the network, tab, yep and okay, so now, if you click on deploy, so let's say you, uh you have already uploaded an application and you click on deploy, so can you it might fail, maybe because an adapter is not running or something but just deploy it. It doesn't matter uh yeah.

B

Yeah bug I just have to catch this bug real fast, because that's uh yeah all right.

B

I may or may not like I I don't know. If I I don't know, if my machines, I think it's connected, I don't know if this is a valid.

B

um A valid um istio gateway, like I don't know if I have a sd. This would require that I have istio installed.

A

Yeah yeah, so can you just click on the second request? There were two requests: the first one actually converted it to patent file. Okay, and can you look at the uh okay, so uh this whole thing is actually a patent file uh over here? Okay, so can you can you click on the first request? Actually, actually this is yeah. uh Can you click on the payload here? Look at the payload of this deploy.

A

A

You click the request. I think it would be in request if you uh here inside of a response, you know we're seeing response at the right side if we just yeah yeah so yeah. uh This is the pattern file, so you can just copy paste this thing so for a particular app. This is the corresponding pattern file and you can just use this.

E

Yeah make sense.

A

E

uh Can we can we also write these steps somewhere if we need them, because we might, you know, for other meshes as well, find such manifests right and and not patterns of course. So I think even that helps.

A

Yeah yeah yeah, so okay, so where should I note it so that it is persistent enough for future use?

A

I think it makes sense to me to, although this is more like a uh developer thing, so actually I was going to request some kind of a button which converts because we can convert a manifesto into a pattern file, but I think that's not an end user specific thing that yeah that's more of a okay. So um so I I'll write. I might write this on the dock, but I think that won't be consistent. The slack chat is also not consistent enough. So should we have this in the documentation itself.

C

A

Or yeah, maybe or maybe I can write yeah, this can be a blog post as well, so that we can just refer users back to that uh contributors back to this thing, yeah so uh for now have you you've got it right so for I'll create a blog post for uh other people, but I hope you've got it for now and if you run into anything you can ping me on slack.

D

By the way, if, if you don't mind, I can write the blog post, I'm kind of trying to get it.

A

Okay, yeah yeah, of course, of course. Of course uh it look yeah. Thank you ruth raj. It would be nice to have a blog post from you. So can you uh so can you create? I don't think instead of creating an issue you can. You know how to create a pr for a blog post right. Do you know so we have a layer, 5, io, slash, layer, 5 repository and uh so lee uh will you?

A

Will you open that via report yeah.

B

Yeah, actually, actually just this is. I think this has just happened in the last couple of days. There is there's a community handbook and then the last couple of days I believe, there's a new page out there about well there's about a writing program and about um how to do this.

B

So deborah priya um batacharji had just published this, and so she certainly would she certainly wanted to engage with as well but um ashish. You were saying also, if you just skip right to it, um uh yeah that you would pr the layer, five blair, five io, slash layer, five, repo one thing that you'll you'll wanna try to ensure happens is that you might want to.

B

We should probably make sure that this is at the very top is to just get sort of get it in queue or get the idea in queue in queue for being promoted as being up for review and kind of being promoted across the social channels, and that way you can make some noise about it, so we might want to yeah, but anyway, this is a good starting point. Don't let that stop you from like either writing?

B

If you do write it in the google doc at first, you might try to just like write it in markdown anyway, because you'll end up taking it from google doc to markdown. So.

A

Okay, okay, so I guess we can move on to them uh so um good to go right. Yes,.

A

I think he's in the call yep he's not so we'll skip that thing. He has an open pr and it's been open for a while now, so if he doesn't respond by this week so either I'll take this up or someone else can take this up. Okay, so going on to the next item, which is for mario and I think he's absent as well.

A

A

Yeah, so, okay, so I was relying on mario to talk about end-to-end tests for uh cypress end-to-end test for multi-cluster, because uh due to multi-cluster a lot of hell, a lot has changed in ui and that is failing cyprus tests. So we need new cyprus tests that takes into account the end-to-end tests, and so I think um I think we can take the take this conversation on slack and ask him about what his views are on this thing or what it.

A

What is his approach for this thing so that, if others want to, if someone else wants to angry, basically, you can uh raise your hand and engage with mario for this thing so that he has some comrades to work with. So, if you're, so, I think yeah yeah a good road. So can you just uh on slack you can? Can you just coordinate with mario to to figure out.

B

A

Since yeah, thank you yeah, so if everything goes well by next, uh no in next quarter I'll be expecting some here, a pr from you here in the cyprus test for after the multicluster thing yeah, and that brings us to the end of the call. Does anyone have any other issues or uh okay? So we have hamza here so so are the tests passing now in your console adapter now the the pr that you had there are the test passing.

C

No not yet, and also I don't know when I try to generate the hashicorps pattern file on my machine. I think my machine is weak or something, but it feels it kind of does not generate it. So I don't know if I need if it's embargo, it's a problem.

A

A

Okay, so your tests are failing because.

A

Your test fields are the very last step, so.

D

That's probably.

E

D

A

Yeah, I guess it's the so.

A

Yeah, it says no resources, so basically the tests are failing because uh it can't find any resources here, that's peculiar the tests are not okay, so um are you? Is this uh uh basically? Is this uh rebased with the upstream? So is this in sync, or is this way out of sync than behind brs, because we had a lot of uh changes?

A

uh We had to push some changes that that were making this thing. These tests fail, and now the tests are not failing on the upstream, so the test shouldn't fail here as well. Only so this the on this step on the pattern apply step. The this error basically means that the pattern wasn't applied successfully. Although you had console mesh mixture, which shouldn't be the case, so yeah uh say ahamsa, are you rebased with the master or are you behind.

C

um I updated it to use the resource based co-op patent, the research based score workflow, but I haven't three bests because I saw lee pushed some some comments I think, like yesterday or the day before I started.

C

I pushed some comments to this branch, but I have not yet replaced it. I've been struggling with this pattern. I think, but and so I've I've sent through to write the the kubernetes manifest. So I hope it works on his environment. Then I will use it in and put it in. Sync then also put the pattern: five in the end, to enter.

A

Okay, you can remove this k8 version from here as well, because, as uh pointed out, this thing is now redundant as we were talking as about about it, though, as the first item of the call, so you can get rid of this.

A

This is of no use now other than the fact that, okay, so other than the fact that we are still hardcoding it to the docks. So if you will remove that, although that is redundant, we in the ui, it won't show anything. So I think we should let it make his change first and then we can okay. So um what uh so? Are you sure that these two parts are the uh parts that that we expect when you uh uh basically locally deploy that particular pattern file on your system, console client and console server.

C

Control server there are two but they're, both named console server, but then console client is one. So I don't know if I should add another console server here and.

A

Yeah, okay, so there are two parts, so um basically two replicas of console server that come up and one console client, but inside of uh yeah, but inside of your action, uh they're not coming up. No nothing is coming up right inside of the action yeah.

A

Okay, so I'll try to debug it and I'll actually take this over uh from you. If uh I'd need your help, I'll tell you and I'll push changes on your pr here, so that and I'll keep pushing it until the I'll push two or three comments where I think the issue might be and after the tests are passing, I think we should merge it because it's a long standing pr. uh Is it fine with you.

C

A

If I find, if I find a particular problem that might fix it, I'll first ping, you for that, that you should go and make that change, and if you don't apply within like 48 hours and if you're busy with something else, then I can go ahead and do it because you've already worked on it enough. I mean this is from 14th april, so I think we should get it immersed. Okay,.

A

Okay yeah, so uh that was all of the all the items that we had for today. If someone has any doubts or anything now would be the time we have few minutes left before we end the call.

A

Oh okay, so we have a newcomer on the call so yeah would you like to introduce yourself? I think you joined in late, I or I might have missed you.

F

E

Myself arya I'm a computer science student.

F

In second year, I'm new to the community- and we are just trying to learn different things.

A

Okay, nice! So what are the things that you work on tests? What.

F

Like currently.

A

B

F

Web development, like full stack development- I know django and some frameworks like react other than that, like I am familiar with python and proficient in python.

A

Okay, okay, so any plans of picking up go because you know I've seen a lot of people from python coming to go because of these similarities. Just in the syntax.

F

Yeah yeah, actually recently I was trying to participate in the gohax hackathon. So over there I've learned a little bit of book, so yeah.

A

Okay, nice, so yeah. We have a lot of react issues opened, so that would be a great starting point. So if you want to contribute, were you there in yesterday's call? Because yesterday we have a debt call and there are a lot of issues were being discussed.

F

No, this is my first call.

A

Okay, nice, so we have a dev call every week so make sure to join next week, uh and by that time you can cruise around the repo with the mystery mystery depot and all the other repos and you might find one or two good. First issues. There are always one laying around so that before your next dev call, you have something some goodies on your hands.

A

Okay, thank you adam.

F

Thank you very much.

A

Okay, so I think we have covered everyone.

B

Yeah, um hey just a bit I've been kind of since we have a couple of minutes. I just want to give an update to all of you, because I think it is a reflection of your work um this week in austin, which is where I'm based there's a open source. Open source summit is the name of it. It's a combination of a bunch of different conferences.

B

I think there's about 1200 people here, so it's relatively small, but that kind of makes it nice because you get to catch up with with people. Actually there's a community member shivai lamba who uh had been one of them.

B

I don't know how to put it at a one of the first or just a very early community member um he's from the same university. I think um that ashish is, if I recall, is she do you know yeah.

B

Yeah, okay, a couple of you know you know, there's various degrees of smart people that go to that university. So I'm not sure, like you know anyway, he's here he's here in austin he traveled over he's giving a talk about lair five mesh mates, which is which is pretty fantastic. um So he's been been here a couple of years and um can't get enough like he just.

B

Yeah anyway, anyways it's great what a great story, what a great story: I'd love for that to be the same story or much more for a lot of you. um Sometimes some of you will be kind enough to join me or or others on stage and and talk about the projects and the works you're doing um actually a different mesh mate mesh mate, aditya krishna just got done presenting at I think dr bangalore meet up there.

B

There's a docker community leader here, uh adina rav, like there's a lot going on anyway. The reason I kind of interjected to mention something is to say that the the engineering team behind docker hub well, they the way that I don't know if you how many folks know this but docker hub. Like you know it's not the world's only container registry. There are any number of container registries now, but it's kind of the de facto the default. If you will tons you know thousands, you have no.

B

I have no idea the numbers that they deal with in terms of images pushed and pulled. um If you recall it was a year a couple of years ago, maybe a year and a half or so there's now rate limiting being done on the pulls that happen for repositories, and so um we had given uh the community. Actually it was kaneshgar jay.

B

I think um he's he's off doing a different full-time job now, but he had joined me at dockercon on stage and talked about webassembly and how you can intelligently rate limit, pull, requests and stuff, and so anyway, advanced to today that engineering team is saying that um they're they're smitten with uh mesheri and with mesh map, and they want to as they look to get off of their ha proxy infrastructure, which runs in aws.

B

They are well they'd like to can hold hands so to speak with the project and move toward a service mesh and use the the tools that have been built here as they go forth. And so um that's great. You know like what a great set of use cases. I expect them to stub their toe on a on a bug or 12, or you know, on a bug or three um but but kudos to all of you. It's uh uh it really says something about the work that you're doing like we're. Just just earlier.

B

You guys are talking about applying patterns, um onboarding apps, just the mechan, all the tests that go on the mechanics, the compatibility matrix that that all that all rolls up into someone being able to use the tools being valuable and yeah. So if you think about it, the engineering team responsible for the world's largest container hub being up and available they're coming to these tools for help yeah.

B

If you, if you have your resume handy I'd, go write that down somewhere. Like you know, it's a pretty cool thing. So so, if you see bugs opened up like with respect to um the docker engineering team and then what they're trying to get done, help them out or you know, be be kind.

A

Okay, thank you for motivation.

A

There yeah it's nice to hear kind words, so that we know that our work has been uh basically is out there and people are noticing it and from what it looks like a lot of people are noticing it and the number should grow in future.

A

Okay, so we are still four minutes short, but I think we are at the end of the call now- and this call happens uh bi-weekly. So this will happen next next week again and uh we would be expecting prs from a few folks by that time. So, up until next next week see you all later bye.

D

B

C