Red Hat OpenShift Developer Experience Office Hours, 21 Sep 2021

Previous Meeting Next Meeting

⏯

youtube image

►

From YouTube: Developer Experience Office Hours: Pipeline as Code with OpenShift Pipelines

Description

Join OpenShift's Developer Experience experts for our regularly scheduled program filled with cloud native, Kubernetes, and OpenShift tips and tricks for developers.

About OpenShift Pipelines:
Red Hat OpenShift Pipelines is a cloud-native, continuous integration and continuous delivery (CI/CD) solution based on Kubernetes resources. It uses Tekton building blocks to automate deployments across multiple platforms by abstracting away the underlying implementation details. Tekton introduces a number of standard custom resource definitions (CRDs) for defining CI/CD pipelines that are portable across Kubernetes distributions. Learn more: https://docs.openshift.com/container-platform/4.8/cicd/pipelines/understanding-openshift-pipelines.html

Twitch: https://red.ht/twitch

A

A

Good morning good afternoon good evening and welcome to another developer experience office hour, I am chris short host of the most of red hat live streaming. This is an office hour. Folks, don't hesitate to ask questions about anything related to developing on openshift, uh I am joined by two of my fellow red hatters chamel and dewan dewan. We are talking about what today.

B

So we are talking about the best places in canada to visit okay, but we might we. We might also talk about pipeline as code, just just as a coincidence. That's.

A

More on brand yeah yeah.

B

A

Yeah cool chamel, would you like to introduce yourself? I guess dawan you can introduce yourself too. Please.

B

Yes, yes, yes, so I'm a developer advocate at red hat, based out of moncton um beautiful, new brunswick eastern part of canada and uh shamoyel uh over to you.

C

Yes, uh hello, everyone.

D

My name is, as you said, I'm an engineer at reddit on the openshift pipeline, working on the concept on and different things around texans, I'm based in the in paris right, not the texan, one, uh firefighters and uh yep, and uh now I'd like to show you some of the things uh um what we do for pipelines that could.

B

Beautiful fantastic how's, the house, paris, this time of the year, shmoel.

D

uh I think we have the what we call indian summer.

D

It's not that warm! It's not that cold. So it's perfectly nice now, but we had like a very bad standard. So we had like a lot of rain compared to other places in the world. It was really hot for us. It was mostly running running so yeah. It's much nicer. Now.

B

Nice uh so chris, as I was mentioning that I haven't talked to you probably in ages, because I was on a two-week pto, I took a road trip from new brunswick to toronto. Ontario, which is a 1600 kilometers drive each way, and the the road from the new brunswick city is beautiful. It's like heli and all the leaves are turning from green to yellow and when it's coming back, it's like green to yellow to to red, and it's like mountains.

B

So when you, when you go down, you see like all the the swirly road matching in the horizon. So so that's what I wanted to mention that if you're in canada or if you're visiting canada, be sure to check out new brunswick and take a road trip and also, let me know like we can grab some atlantic salmon together there you go.

A

When the border reopens, I might take you up on that.

B

A

Sure um but yeah, no, it's it's that time of year here in the uh the northern hemisphere, so uh we are falling into fall as it were.

B

Yeah and she will when was the last time you I guess, went to any any in-person meetup or went to an office, or, I guess even got out of the house to to to do something.

C

Techie last time was.

D

uh Just around the I think there was like a dev desk in chicago just before the I think it was february so and uh and yeah that was the last time and- and I think I traveled just just like at the as the pandemic was starting. I was traveling to israel and uh and when I came back like I heard like so many things like you need to do tests and stuff and things you know it all started there, just so yeah.

D

That was last time because I took a fight and then I said I I stayed in france most of them.

D

What about you.

B

Same, um I guess uh it was march 9 2020, and that was the date uh my colleagues at ibm and I were told that we had to pack our stuff from the office. We grabbed all the stuff. The monitor came home and we never realized that it would be like months, uh yeah.

A

B

We would had head over and by then within that time period uh some of us came to red hat, and it's just it seems ever since I've been to an actual office or been at a tech conference which I'm really really hoping we can, we can go to some conference in person soon.

B

I see I see tecton in the background and, and that kind of gives me some question. I've been doing. um I guess some git ops, related talks recently and- and first question I have in mind for sure well, is how does like pipeline as code relate to githubs uh like we, we hear like infrastructure as a code pipeline as code githubs and then all these ops, so I guess many folks and in the audience would would have the question that how does let's say pipeline has code and git ops relate or or doesn't relate.

D

Yeah, uh that's a good question, uh because a lot of people would say that uh what you can do with github you can do with python. Is the code because the python's a code offers you to develop your pipelines inside the inside your code? So usually you would have like this inside git and your pipeline.

D

Would you do acceleration inside the inside your code uh like to define micro only pattern and it would be defined as it goes like a iteration of your code, um you, uh the difference, is the main difference is here like the pattern of the code is like five of the code is really focused on pipeline on text on pipeline, not on other kubernetes objects, so kubernetes. Obviously, usually you get like some other products that are like github services, those uh those those that does uh like github and kubernetes.

D

Those are those products like the would they would be able like they would let you manage everything that you have on your secrets, your your wraps, your ingredients and everything like that. So uh so that's that's the main thing when the python is recorded. In this case, it's like you're, really writing like a pipeline text on pipeline or text on task, or things like that, and you really iterate over there.

B

Got it nice so would would would a generalization be uh if we can think of a subset versus superset? Could one of these could be thought of a super set of the other.

D

It's a super set of the order. Yeah in some way is that uh usually people would uh iterate like when they're starting it right uh from pipeline uh some from some pipeline. The really iterate like the iteration would be uh inside the code like over our pro request, and uh it would be a lot more like a full reference.

D

So, for example like it's a demo, I'm going to show like uh later on, hopefully, if it works yeah that that, if I, uh if you start, if you sound like, for example, you have your code, you have uh some code like in some languages like, for example, python, and uh you decide like who had an option and but uh like in inside that uh that ocean is like. You need some kind of other, um some, uh some uh some some and some new linter.

D

So, for example, you need to increase like a gamble, the entire input, so part of your course. You add that feature inside your code and part of the of the same comments. You can add uh the pipeline the like the task uh that you that's going to do like the yammer, and uh that would be that um that uh that would be. uh That would be the the diabetes.

D

The purpose you know on here is that your commit itself like whether there's a picture inside the code and what's the future inside the pipeline, like to add that thing and whenever you're going to submit that full request, I mean it's going to uh take into account that you exchange and it's going to uh use that pipeline. So that's so that's the main.

D

uh That's the main advantage whenever, when, when we talk about github in this case, like you will do this iteration, but uh you'll need a station like other more like a diplomat leader on an activity you can do on a ci level as well too, but uh in this case, like uh the the ci part, I mean would be handled by pecton and python's account like really liberates. Like the text. One feature of the director.

B

Makes sense, uh the other question I have is like in kubernetes, like the david server takes declaration as an input, so we do things declaratively rather than imperatively. So let's say if I have my pipeline as code or if we forget kubernetes for a second, let's just say we define our infrastructure or our pipeline as code.

B

How does the declaration uh matter here like? Why would someone consider uh tools, tools like techton, uh which which does thing in it, in a declarative way that we we define, how we want to see uh our pipelines.

D

uh Yeah, so that's that's more like a question for like what's the advantage of using pekton in this case is that is that, uh what's what you're doing with uh sections like declarative, like a pipeline uh like audio declare python inside section, so that data is like all those things like a very natural communities?

D

So everything is like, as a is a is a no sharing as a secure, as we said so, a lot of products are there or ci as a as a demon and uh a big fat demon, as we call it, and uh and uh and I'm trying like a lot of different things when here like we are talking about, everything is done in as a community crds as a for the viewer will know.

D

What are they and uh all those concepts like every concept of a pipeline is a crd, so you have like different things, uh different things in there in tecton like to build a pipeline.

D

uh You can, for example, you have uh yours like here's, a you have task, that's uh belonging to a pipeline, so a path line is going to be able to reference like attack and- uh and you have uh you have like. You said that you have different steps and tasks, pipelines and uh all other resources like they're all like crds, so you can manage them directly on the on the pressure.

D

So that's the advantage is that uh in this case, it's like you're going to declare your pipeline like we securing these resources and manage them as you maintest resources as well, and that's how you you're going to describe your infection.

D

uh Yep awesome.

B

So chris uh question for you, so you have been in the industry a long time so so, if like from from let's say if you, if you think as an external point of view when you first thought of like what pecton or came to know, kubernetes native tools like these, what were your primary questions or concerns.

A

um I mean my primary concerns: were around I've been in the industry, this long stability, but after that, right, like these products have matured now. um So it's it's more about like management and creating you know good practices and processes to put in place for these things. So that's kind of my biggest concern now is that we have all this great tooling.

A

Finally, right like we can do all these a mad all these things that we couldn't imagine we could do before um now, it's time. Finally here, but we weren't, you know sure of what tooling was going to come in place. So, let's figure out the process which could lay down- and you know, protect on yaml um that exact process for how we want to do things in the future.

A

So, like the intersection of the two I think, is interesting because we've we've come to that point where it's now like all right, I can do all the things.

A

How do I do all the things and so like instituting that process as kind of code? That makes me feel very good about things, because declarative configuration is very hard sometimes because it's not just one configuration it's always we've got dev environments or you know staging environments all this other stuff. So you got to have like good configurations for everything.

B

You hit the nail on the head and there's a similar question. I got from from some customers as well uh uh like the first question they said like, is it being adopted like uh and it was like, maybe two years ago, when techno was still younger, that the house adoption are folks using it in in production and in the similar line.

B

The other question I got is: can I just use tecton and then just do everything on my own, like using the open source uh version and how hard it is for me to uh just focus.

A

B

Adopt on on on my own uh social as the expert, uh so how how would he capital capital d uh we're not letting him escape with an a article, the expert? How do you address those questions like the the adoption of uh tech done and how easy or difficult would it be for an engineering team to do everything on their own uh for tech time.

D

So so you can definitely do use things on doing most the things on their own um on on openshift. First, I mean it's going to be a little bit harder, though I mean when you're going to start like, so you can install minitube or kind of kind of like that, and then start like this to try to do the upstream installation and it's probably going to work fine.

D

But there are one things where, like you need to move it like with position to open shift and kind of thing like that, uh you're going to see that there is like things that aren't going to work that that well anymore, with the upstream one, because you're going to see that there is a like because of the security constraints that uh that we have on openshift.

D

uh It's uh it's like things like are not going to install properly. That is going to be missing some tasks like kind of thing like that. So that's! So that's uh that's! What we offer inside the opencv pipeline is an upper record. I've been to install uh most of the things properly uh so so, like everything, would be uh next integrated, like between all the different services, so there's different services inside uh inside openshift pipeline, which is a trigger uh it's uh it's.

D

A project triggered like like, usually you would install pipelines and triggers together because trigger is something it's the thing. That's going to reply to answer to uh events, that's coming from our site, so, for example, when you have like a github web hooks or different web books, you want to try a pipeline. Usually you would install trigger.

D

So that's openshift pipelines like I do the integration, because they're cool and and for other another example that uh the integration is done is that uh for when you're on like to build like a uh a container image on the authentic, uh you need like some kind of privilege, for that.

D

So the way we we're working with this, like, if you, if you're going to do to try directly I'm going to if I'm going to work, to build an image, but we do set up like a service account like in a way that's basically or we like, and we make sure that uh people can uh just use builder directly. Whenever they're going to use the task builder like to build containers, then uh they will be able to do so and they will be able to reuse the openshift registry as well uh directly.

D

So it's all those integration with the is a great uh openshift ecosystem. That's the advantage of using the openshift pipeline uh that uh um and uh yep. So so you get uh so you so you get all those features and uh and uh and yeah those uh those are mostly uh mostly like a new. We make sure that everything is uh supported.

B

Awesome thanks- and I guess that that also answers the question of adoption, because products like openshift pipelines and- and I know many other products uh are using techdon as the underlying engine, so so that that speaks for the adoption point. um I know I asked a lot of questions so I'll pause now and let shmuel uh go ahead uh if he uh has any any uh demo to show or any any overview uh for pipeline as code yep.

D

uh Thank you and uh if, if you have any questions or anything or some people, the viewer, please feel free to introduce me. I mean uh no problem with that. uh um I'm going to uh to show what uh websites on the code and it's uh it's a death uh feature that we just introduced with the last uh release of uh openshift pipeline. uh One point: five, I think uh it's on march 22nd and uh which is the land.

D

It's uh it's really been uh been in development right now and the goal of partnering the code is that uh to make it much easier to uh to develop your pipeline, it's uh so the goal of uh of spectrum by default. Like it's really like a um I see a tool like gives you a lot of a very flexible way to do a lot of things together and, and people can do.

D

Can people can do anything like anything with it, but it's very flexible, but the point is that it doesn't have like a flow that uh people would expect. So, for example, like a lot of people that come from texan are coming to us and asking us, but how do I do? How does it work uh technology? uh How can I get the same experience that I have on the on other three ci tools like github actions or traffic ci's or something, and uh how can I have the same experience because currently like what?

D

uh How does it work with uh tecton is that uh I need to start making my pipeline? uh Maybe it's going to be energy repo, and uh I need to deploy it uh locally on a on my cluster and after that, the pipeline is going to end like the event. But if I need to modify the pipelines like, I need to modify it or maybe like.

D

If I'm, uh if I have good habits like I'm, going to use github uh way, uh if I'm going to, if I have a or it's not like I'm going to just you like your ctl edit, it would be like uh bad because you're going to lose your your your configurations like whenever, like you go into guitar or camera anyway, uh with uh yep.

A

Shameless start interrupt, but could you increase your font size a little bit please? Oh.

D

I'm sorry for that.

A

No worries that one of the audience members asked for it.

D

B

If you could do an echo on the terminal as well, so that the audience could see if the font is.

D

B

uh Let us know.

D

Let us know if the channel is good, yeah.

C

Thank you. Thank you. Yeah thanks.

D

uh Okay, so uh yeah so pipelines because gives you like an opportunity like to uh to get like all that flow that make a new. um You know to develop your pipelines like inside your git people and uh and make it uh and make it very easy with it. uh So the feature that we have is that uh that we have on python, as I called like the flow.

D

The way the flow would look like is that it's the same thing that if you have some expensive traffic ti or youtube actions and uh and other things uh is that you have your people, uh you send your your pr to the git repo.

D

If you start to have like a that takes on the that tecton directory inside it, then it's going to detect uh the pipeline that you have inside I'm going to show the pathfinder and, and then it's going to pick it up and automatically like whenever you do like a full request, uh you're going to have like a status, uh the same things on github with the normal status like here. It's like some code coverage and uh yes like I have some details. Anything like that.

D

So so that's that's something that gives you for free, which is something that you don't need to set up because usually like, if you have like, if you get that experience like you'll need to set up a lot of things uh for a pipeline uh inside your pipeline to uh to make it work. But here is like my record like lets.

D

You focus on writing your pipeline itself, like the textual pipeline and all that strum being behind it uh like stuff like doing like uh updating the the the checks api, the github checks right here, which makes you uh tell you like if, for example, the the like that has checked api like it doesn't disconnect this, uh this sha is validated anything like that. um That's uh that's all using uh you.

D

You can do a quick request and push so it's actually, you can add some full requests and on push it supports like different actions, so the action that you you would see, it's actually something that we use on the on pro and uh that's important impact on the code.

D

So, for example, if I do, if I go here and and I'll start to do, a slash retest like I would get uh automatically like at the restart of the of the of the pipeline, uh it's going to come out in two secs, but uh let me go over the overview first and uh yeah. So that's uh that's the feature that the super like there is a different uh there's, more feature which I'm going to show like during the demo, uh something that uh we make sure that uh to make it uh to make from.

C

D

Start is that uh is that to not make to make it as not secure, but uh not open to anyone by default, because a lot of ci's out there when you open them by default, like there's some security issues with full requests from someone who's, not part of the organization. For example, we would be able like to uh exploit some of the secrets there. uh We, by default, like everything, is, uh is limited like to the people that owner or the of the repo or or part of the organization, for example.

D

Here it's an organization open his pipeline, uh so everybody in python's record part of the openshift link can open, or we have like uh some way to add a file inside the repo to give you the owner and uh and uh to to give a acl from other people.

D

So the point is that that we try like to make sure we don't make uh to make the mistakes like to open to everyone, because uh we we've seen in the news uh lately, especially about ci's, like of exploits on those things and- uh and uh we try like to uh to avoid that and uh there's another features like we have a cli tool which I'm going to show as well.

D

um That's uh that makes it easy like for bootstrap uh automatically and uh to manage the pipeline and see the status and see my lap so enough with the blah blah blah. Let's, uh let's try to do it like a demo. So here's my notes here, you're not going to show like very well the phone, but that's all right, there's something for myself for my uh from my uh from my business.

D

So, first of all, like I'm going to go over here and I'm going to create a repo, that's using the github to create like a open uh repo, that's something that people would do uh would do uh would do normally like they would have like a like a repo on github and they're just going to uh I'm just going to create it. So, let's do that, so that's just another thing like I don't need to explain, I'm going to say it's going to be a python python repo, and I'm not around that.

D

I can't if I really want- and it's going to create it and it's going to check it all right, so it's created like for that. There's no magic here. Like that's other things, I just created a repo inside github and uh and that's fine so so now that people have been created uh before I do before I do something like I need to change. That's the way yeah, because uh my I'm not using http on my machine, it's the only shp, so it wouldn't grow on the push.

D

But anyway, like you, can uh it's not bad, so I just set up my repo, that's the first thing I have to do uh now. It's like, uh let's say that uh this repo is going to be like a python app like, as I said before, so I'm going to create like a small pattern, so that python app would be would be a buggy icon app and I'm going to create like a file with the ai, and here we go so I have like a small file.

D

uh So that's pythons like compile compounds, so you can run it. I think it's uh yeah. It doesn't copy that, but uh there is like some something here there, but it's not like photo of python code. If you know pythons like you need to have uh styling is going to tell you that uh those variables are ready.

D

You need to define it in the other way or you need to do those conflicts, but in another way as well, but at five, like I write like I wrote basically like just to show like when I'm going to write a pipeline and I'm going to add the piling like it's going to pick it up and I can attract my pipeline over it as well.

D

So so here's what I've created my uh my pipeline, uh my my app. That's a python app for that things, like I just created my uh so now, it's like I'm going to use the I'm going to use the tkn pack repo to say that I want to start like to handle that uh repo uh inside uh pipeline as a code.

D

So I'm just using this so this tool, like uh it's a pro that we did, uh it's called pkn pack, it's a plug-in to the pkn cli. uh The tk ncli is the cli that uh to manage the text on the on the command line, interface and uh and this year extended.

D

So when I get let me let me let me do that, so we so we started to. We start like uh we start doing that, like it's. Just everything to ask you for a namespace.

D

So the idea here is that uh every pipeline that you develop, like the user, who developed that, will develop that inside their own namespace project inside openshift, and- uh and that's where, like you would, you would have like uh all your uh all, your all your results and everything like that. And if you have a team of different people working together, they will give access to that uh shared repo and they will be able to develop their uh their own uh their uh the pipeline.

D

B

Question uh would uh tkn pse h show uh what the what uh the command is expecting as inputs or what the command does in general, you.

D

B

To yeah yeah sorry, it didn't mean to yeah.

D

That's all right, that's all I can. I can do so. Tkn pack free fall is just like it gives you like a so yeah. uh That's a good question because I should have introduced that before um so repositories is, uh is a trd that we have introduced part of pipeline at the club. It's something that uh that you have uh that. That is something that uh that that's not uh inside the inside text on itself, but it's actually something that we work upstream to uh make it more tender.

D

So we have uh on python as a code. You have a way to uh specify repository, so you say, like that's my own repo in there and it's going to handle that event. So it's a push or pull request going to that um to that uh to that uh to that branch, so it could be main or, for example, release 1.0 really 2.0.

D

So that's that's a crd! That's actually the configuration uh that says that tells patterns the code uh on which namespace this repo is going to be tested and which user like uh since, like the user who created that that, before you store, is at the right. So we're going to install only on that on that namespace as well, where you will be able like to see it and uh and thing, and uh it as well gives you like the status. So all the status of uh of a repo like to say like that.

D

This had been delegated like he has like a green, green and a kind of thing which I would show up there. That gives you, like all the searches there. So, okay.

B

So that's it before so before uh the pipeline has caught. uh So my experience, maybe a few years back where uh we had the pipeline run where we defined the we dynamically inputted, the repo which was feedback to the pipeline. uh If it's a private repo, you have to add the service account or the secret for that, uh and I guess the repo as a crd was introduced more recently. What is one of the advantages of having this repo crd over just entering my repo as an input like like previously.

D

Yeah, so the advantage is like those are two things do different things uh uh at first like at the the the reporter. The cd of the repo would just like, be able like to uh to tell python as a code that this user wants to handle that event, so because the difference is that uh is that, usually, if you want to do that kind of thing, you need to write like some event listener inside triggers, which is like the way uh you do that, but you're like.

B

D

Near the end, kind of thing it is like is that people things like you, don't have to do that like it's a it's an easy way, because it's an opinionated way, you don't need all the flexibility that you, the event listener and the and the triggers and the trigger templates and all these features like from the trigger like here, you usually need, like you, don't need that you just need uh to specify the repo itself and uh and the type of event, and I would handle it, and after that I mean the pipeline run and all that things like everything that gets set up like I'll show you like in two sec.

D

I just need to boot fast and I'll show you it works. Maybe it would make more sense. Let me know if it doesn't, but uh hopefully it does so I'm just going to spread the repo with that and you'll see after the bootstrapping. Now it works, I'm going to say like I want to create it inside the uh well. Where do I want to do it open it uh yeah, okay, more test? Okay, so that's the repo! uh So it does, it did automatically detect uh like which people you are on.

D

So I'm going to use that I'm going it's going to say like whenever. There is an event coming from that url uh and going to target that point, which is men and going like on a pull request to a push. So in this case, like I'm, going to say, pull request and, and then uh it's going to handle like those events and it's going to check out the the code of uh of the of the repo and going to see to discover if there is any pipeline pipeline inside the text on directory automatically testing.

D

So let me let me the first like is going to ask me if I want to create it. Yes, please, and now it's going to create like a text on file so which this is the thing I was talking about, so the text on file. uh So it is like it's created like a simple pipeline, so I'm going to edit it and that's where, like things like, are going to make more sense, so it automatically created like a python one. So maybe you've seen that before.

D

Like a long time ago, usually, you could have like a pipeline and a pipeline run that close the pipeline and task, but you can have another way as well. You can have like one single pipeline run which are like which have everything embedded, uh so you can have like a as the first restarting the pipeline rest directly. You can have a python stack it.

D

It just makes it to make it easier for people to uh to understand that uh you don't need to have to have like a pipeline separated from the backhand bone, but they can do. They can adjust that, like embed the space directly inside the pattern.

B

It's very convenient.

D

It is much confident, and uh there is there- is a pipeline run, so the data cleaner has been created, obviously, and he it adds like a few annotations, so the notation is specific to python records. So there is one thing that I need to say: is that a lot of uh a lot of those concepts that you have here, uh everything that uh everything that the everything that that pipeline is like a standard pipeline from uh from tecton? So we try like to not diverge.

D

We try like to do standard things, not introduce like a new, a new dsl like a new new like configuration language. Everything like that step-by-step. The only thing that you found is those annotations, which is uh other things which are going to explain and there's a little bit of difference. It's going to answer your question before is those viable, so those variables like they are not like pure tecton. They are like coming from a section of the code and those are going to be replaced automatically by the event.

D

So if they're coming from github.com well uh slash democracy pattern is going to be replaced automatically and going to be replaced on that position. So so that's that's! That's the advantage in that case of, like you, don't need to specify, like all that input and all that thing and stuff everything is going to be like a a viable. That's going to be dynamic and uh you don't usually you just stay like that, and even if like, if you move your rifle somewhere else, those are going to be replaced properly uh in some way somewhere else.

B

And can I use a single pipeline or an eml for for multiple repo bills uh or it's a one-on-one relation every repo? I need one pipeline running ml.

D

It's currently like a one-on-one ripple, but uh you can potentially I didn't try to be honest, but financially you could. uh You could install that same repo to uh to another uh to another so that same uh that same soft four code. I would say uh another repo and uh and all it does there definitely uh but uh currently like you, you like, like we like. We took the design decision like make it like just 101 before story like to one but like.

D

Maybe if we start like for plenty of future requests for that, then I will try to to get it.

C

D

Much uh it's actually much harder, though,.

C

D

Like you know, apparently, but anyway, it's like yeah, let's go back to the the things like those annotations notification. That's inside I found the code with the reference like what do we want to enter, which even we encounter we want to even like we want to enable a cool request. We want to uh under knock on men and but you can have like different syntax, like with a greedy operator, like we say like, I want to enter like every tag like 1.0 2.0, and I want to make a pipeline for that.

D

We put that as well, and there is another magical things in here. uh If we are saying like, I want to be able like to install the task called gitloan, uh then what it's going to do is like it's going to uh go over text on the hub and it's going to fetch that uh the latest one from deep clone and automatically you'll be able to you'll be able like to use it.

D

So, just before running the pipeline is going to uh get the key to class and get and uh and python record will put it inside. The advantage is that you don't have like whether you can we use the the the text on hub uh class and you just need to reference like eclone, and you always get the latest one and there is like a different thing for syntax there's.

D

Another feature that we had like that has been requested a lot and that you can't get from this with the normal section that you can get with background records is to be able like to only get the last five past time. So one of the problems we have with section is that uh is that we don't clean up like those pipeline run, but uh for different technical reasons, which I'm not going to go over right now.

D

But I guess like uh since we are it's easy to do, that in spotlight, like you'll, be able to say, like I want to keep like the last python run and uh and uh engage with it, so it generates like that pipeline. As I said before, like he has like a few variables in there, which is like those variables like that specifically pipeline of the code, but those are viable.

D

uh What's coming from the triggers from the from the trigger event and uh and yeah so so I just put like some standard effects the task in there like when it happens. They could like just generate them like it's going to use like the git on task so that they contact like it's a task right.

D

It's not like a task space, but uh this thing is that, uh since we said before, like I want to uh integrate it so automatically this game, you can reuse it like automatically and uh and and after that, like I have a simple type called no task which is going to uh uh years like it's, not a password, that's a so where you can embed a directory and it's just going to uh execu.

D

So that's uh so! That's uh that! That's uh that's! uh That's the thing! uh So, if I'm going to first of all, uh something that I I need to I didn't tell uh about is that um five miles across currently is very tight to github. But we are working on the vcr support and it's very tight to uh github apps github apps, like applications that you can automatically like install inside your repo, so which are being uh something that I'm going to do right now.

D

It's all done, I'm going to copy that and I'm going to configure my github app, which I've created before and I'm going to configure like on the repo that I created.

D

So basically I'm going to say that the app it's going to handle the uh my my repo that just gets created and- uh and that would be- and that's the only thing that the user will need to do to add section as a python of the code to uh their own uh to their own uh uh opencv twitter to uh to their own repo, so that they just need to add the app.

D

So that's uh so nice like it's, been installed uh now, I'm going to add it commit it hello from that pack and I'm going to push it to the origin, uh and I just I forgot something. I'm sorry. I need to do a branch I'm going to reset it because I don't want to mess it up. So now it's like. I have a branch called demo openshift pipeline and I'm going to push it to github and to generate like a full request.

D

Here we go now. I can just uh and let's see if the person is a code.

D

D

Okay, you didn't push it, you didn't push it. Oh yeah, I don't know- maybe maybe this anyway, uh okay, so it's like that's my request. I'm going to do my pro request. I'm going to create it and uh you'll see. Python's record has already started. So it's already started. So you go. If you go to the ui, it's like! That's the that's just the start.

D

uh If you go, if you use the pipeline like if you go to the back end, if you're on admin, you can see that uh it generated like a python burn in here and it it, it just started the python one.

D

So all the execution is generated like in a five point, one which has all the logs the admin which is really like an admin thing to do so now it's like it says, like I've, just started the the background one I did match the so using the report I was saying before, like I match it to the opencv mode test repo and you can watch it from the command line like that or you can just use the openshift console directly to watch the execution.

D

So here's like, I did fetch the repo so, as I said before, like it's like it's using like the fireball and you need to extend it and it's going to do exile. So that's uh that's! That's another thing, and now, when you go back here like you, have like a mexican report uh with everything and with the committee that has been validated so now, but I'm going to uh do like the pull request and I'm going because it's a python code, I want to add the file link.

D

So let's uh skylink is like a linker for python and I'm going to add that step inside my pattern, one, but since I'm really bad at writing, I'm just going to copy and paste. So so I'm going to put that here, for example, I'm sorry, let's see that works, I'm not going to go to uh vi. Let's try to get 12.

D

Yes, okay, all right! Here we go so where did he get like? I just add, like a new pipeline, uh a new step, that's going to refund the styling pipeline. uh One is like I said like this. This, like it's different, the piling uh task that filing task like I need normally, I should normally install it. What I'm going to do is I'm going to add styling here and ultimately, it's going to fetch that pipeline from the text on the hub and for the second catalog and you're going to to have it.

D

So that's the only thing I need to add like to uh be able like to uh to uh to to do the ci on my uh on my python app. So I need a styling check and I'm going to push it.

D

Okay, usually like it doesn't work. I mean like I, I didn't write it out properly, so I'm going to come back to what was like on yeah, and so I just push. I just pushed it uh and now it's like normally, like you see the somebody should yeah there, you go it's coming up, it's just going to validate like a new push and it's going to normally. If it works, it's going to be used, uh it's going to reuse the pylink and it's going to do the painting.

D

So the thing like here is, like you just need to part of your pr. You can just iterate your pipeline and add your pipeline like doing uh doing the the check.

B

D

B

Question if I'm at dev- and I just want to test it on on my own branch- could I specify uh on push rather than on pull request.

D

That that's exactly yeah, so when you, when you saw before like uh when you're doing like pkn pack, people um that uh that that that tk impact people uh was uh creating like it was a ceo like. If you want to put on pull requests on push, then you can just do one push and you can say like on that release on another that type. But here's like I have well, I don't know, what's the problem, but you don't know.

D

Like it does, it did fail. I was expecting that the only thing is like that error. It's like it's a fine interior, but I guess uh it's because I didn't specify properly my uh my uh my name but.

A

D

Like he did, uh he did launch. uh He has like videos the pilot it is like started like two to. uh Let me let me try locally, because I'm not sure why it's taking that thing.

D

Well, there's a there's a thing with the task in itself, but um but yeah like there's uh like is like we have. We have introduced like a new task inside it and uh if you wanted like to fix it, for example, we say like. Finally, it's like uh like what I've introduced like. I don't want like to have it, because I don't want to validate the python code. You can just uh redo, you can just delete, say I'm going to commit it, I'm going to disable the styling check, because uh it's not accurate.

D

I think, there's a problem with my uh with my structure. Here. It's been a long time for that item and I'm going to push it and now it's like it's like it's just it. It's going to pick up like the change and it's going to be the pipeline, usually uh usually that's pipeline. That changes like you need to do live on the cluster, but here's like you can just do like inside the code and that's the main point that uh that's where it is like. It should be green because we removed the tail there actually.

B

D

B

Point I wanted to point: is that you're doing all the git based workflow, which developers are very familiar with, like you're, not doing any cube, ctl commands, so it's very natural to developers.

D

That's correct yeah we can, uh you can do that and uh it's it's very native, like usually you can use like most of the things you can just use the console and uh and guitar directly like you don't like to use the tli, even if you really wanted, but uh there's as well like uh the the two packs, like mtv view status.

D

uh So if I do a consonant that will create it here,.

D

All right so yeah, it's like I have. I can list my repos, like all my epos and the events. I don't know if you see it properly here. Maybe it's better! uh You! Yes, like you say like it's uh like that's my all my repo, like, usually you like a discount people. It's going to handle like that event, which is the full request, is like in this case like it's infective.

D

If I start to do a describe on the on it, that is going to pick up the last one, it can tell you like. That's the last run, that's successful on that commit, so that's the commit that you defend, but uh the the other one that was before like I was failing, and it gives you all the status uh in there. So usually you would you can use that uh we are working on uh integrating uh that status uh inside the openshift console tool. So hopefully we get.

D

We we're going to have it uh uh the signal rather than later, and uh and hopefully you can see like uh soon enough, you'll be able to see all the resources exactly so so you're going to apply all your lists of your story inside the inside uh inside the console and uh and uh the embellished detailed like which sha is being validated and anything seen that. So that's uh that's the demo. uh I don't know if I'm out of time for a question or yeah.

A

You got about 10, more minutes folks, so if you got questions, let us know put them in we'll get them answered.

D

Please feel free, I know, there's a lot of concepts here and uh I I yeah I I think I could have spent more time on from the concept, but uh there is uh there is uh this thing that I wanted to talk about which I did break it down? uh There's the ecl with the echoes which I was telling you about before. It's like you can have like a you, can specify like axles uh inside uh inside your repo and say like.

D

If I have a corner file and uh going to say honor, equal, let's say d1 and then then d1 would be able to submit a full request on that uh on my repo and automatically like it would uh it would uh it would pick it up and it would work, but if not like it would be denied, because uh because, uh as far as I know like uh the one is not part of the official python organization but uh yeah, so that's uh that's.

D

That's uh that's one of the thing of the article that we do. uh We have uh other features like uh we have uh the the the resolution for the resolutions like I was saying before, like you can specify like uh different things inside the resolution.

D

You can say that uh that you can play the you can say that, obviously you can say like something from the hub or you can say something from local uh from the repo from somewhere else in the repo or a remote url as well and there's one other thing which actually I wanted to say because you asked for it the one before you were asking me like. uh How do I do like for private people?

D

How do I check out private reports when I have private people yeah there is uh something magical uh with uh with uh with python at the code. Is that um with private repository every time you um every time you uh you start like a pack and run every time.

D

You start a pipeline, run you uh the what what uh python is to create a secret inside that uh namespace, where you run it, but a secret that uh that belongs to you, the one who's thinking right now and that secret is called backpack, give back more open, qd market. So basically it's going to give you basic up. That's your organization, which one indicates and openshift and that's the name of the pipeline. So that's that's created automatically that that token is coming from the github app. But when you install a github, you say like.

D

I want to give you the ability like to generate tokens from ideas to the app so since python's record. As knows the token, so it's going to generate that and what you would do inside your pipeline uh type inside your pipeline, like it's all documented here, maybe we'll make it a little bit better better. You just need to add a workspace to the git clone task as a documentary called basic, or that's going to reference that secret. That gets auto-generated.

D

So so that's that's some viable. I was saying before so it's called dynamics, so you can transport between different things and the magic here is like git. Clone is going to see that workspace, which is a feature of section called optional workspace and uh and that optional workspace is generated with like git, config and http with the token. So it would be able like to uh run uh to check out the code from file triple so, for example, like I'm, I'm developing on uh on github enterprise.

D

So we spoke with an enterprise and I'm using it a lot as well. I do need it as well automatically like to make it uh to check out my code.

B

That is like. I can really appreciate that, because I remember like before I had to create a a secret and then attach it to the service account. Then.

A

B

It like all of these, like in order to appreciate how automation is helping you sometimes have to do things manually and then realize that, oh when, when it's automated how convenient it is, I agree.

D

And and there's there's one thing like I'm sure, like a lot of people, would ask like the competency, like security. You know like what's what's happening with situations and that's something we did uh think about a little bit so so we need to give the service act.

D

Not a service out of cyclone's account needs to have the ability to create secrets and delete secrets, but that's on on every name space, but we restrict like to only delete and uh and create and, and those secrets are only secrets uh are only created like inside the user, user or namespace. So it's only people who have access that username would be able to see them and uh delete them, but uh which I like to retrieve, but we did give an ability like to.

D

We have a way to configure yourself another code and uh you have a complete map and you can disable that feature if it's something that you are too scared about.

B

That is amazing, like the the amount of information you showed and demoed in a such a short time frame is, is amazing and like there are a lot of other integrations with tecton. I know shameless shows like that, the git clone and the lend, but there's also like slack integration. So let's say if your pipeline ran and then you want a notification to appear on your slack um yeah there you go like you, you can you reuse the the tasks in tekton hub.

D

Yes, that's correct, yeah. I I give I give an example. I actually write attack, but it's like if someone uh maybe undo a shameless plug here like which makes it very easy to uh to to plug your like on any pipeline like a passport partner. So we report the success like with an image secret and everything you can actually uh use and you will get like a nice uh nice little strategy, but uh but yes, uh sometimes, I got currently like support only github checks for notifications.

D

But if you want like to use like, I documented it here, because a lot of people have been asking the questions like how do I do like integration, but uh you can just grab grab the examples from this documentation here.

B

That's great so we're five minutes uh uh to to the end of the show. So if let's say, if I'm a dev and now I can't wait to get started with techno or or openshift pipelines, uh if you could recap the starting points like where to get started, the talks, the the links just so that they have it referenced. I'm.

D

Guessing like you're, going to have some phone nuts, so you're going to have a so. The the github report currently is a is the source of everything uh I did first like I did I did uh I did. uh I did um a work through video uh on youtube.

D

D

Well, I can add that I think, but uh but yeah I have a work for the video, but if you want like to do the inside install guide, the install guide is here like he has, like all the recommendations I have as well. I work through video to go like over the install process and which, because there is like you need to create a github applications and all that thing, but everything is uh these details here and uh you can you can just like.

D

Basically, you just need to install like a really the channel and uh create a github app and after you're, pretty much uh you're putting it on, but you need to do it properly, though. So that's, why, like you, need to produce mutation but feel free to go on the github repo and uh which is openshift pipeline python's, the code and you'll get uh you'll get documentation. Hopefully, we are working on documentation on a other type of documentation that supports the more details, but uh currently on this preview I mean there.

D

The next features uh we'll be working on. uh Currently it's going to be like one important thing like being able like to support different vcs and github, so uh so like like big buckets, so that's something that that we've been working lately, but a lot of the things they say. That's github! So there's a lot of feedback going to be done and hopefully you'll get these packages gitlab and all that stuff done. Whenever we get there.

B

Awesome if it's okay, I'm gonna, drop shmuel's, linkedin and twitter handle on the chat so that if folks have questions, they can and reach out and then tag on and and um has public slack channels as well.

D

We have three minutes left, but I have another thing like we are working as well on the concept of things I mean uh with some other people from the community. We are working on the on making, like all those. So there is one thing that I am not super happy about is the annotation annotation is not. um The annotations are not very scalable. You know kind of things like uh at first, but uh we're working on a on a stream like part of the workflow working group in tecton.

D

So if you want to join the the spectrum channels there like they are, we are always there and uh to make like all those concepts like a little bit more standard than annotation and uh inside the inside. uh It's called workflow, actually and uh but hopefully uh when uh well like it's another topic to talk about maybe one day I'll be able to speak about it here and uh but uh yeah feel free to drop on the slides on the on the upstream flight.

D

If you wanted and uh take some slack and uh you get, uh you can contact us there, hannah I'll be happy to answer too.

A

I'm trying to find the link to sign up for techton slack. I know I'm in it, but I don't remember the.

A

A

There we go, no, you have it, but I don't it's not loading. For some reason. Can you drop it.

B

I I pasted it in the youtube chat.

A

Oh, you did okay. Thank you.

B

All right thanks so much again shamoyel for for coming to the show and uh talking about pipeliners code. I definitely will try it again this week and chris so now that I'm back it's time for you to visit canada and travel across the coast of new brunswick when the border opens.

A

When the border opens dot, dot.

D

I think the the border just opened for europeans, or maybe I'll visit before.

A

I hope yes, please, please come visit all right folks. um Yes, thank you, shamal and thank you dawan for putting this on. This has been great, uh which one is the best option: f, techton or arc.

A

B

A

Coming in from chat or arc is that, like an azure thing, tecton or arc, I'm not sure I've heard of.

B

Is, if that's an abbreviation of something, I'm not too familiar.

A

A

Yeah, if you could clarify what's up with our kumari that'd, be great techton vs arc, I'm not coming up with anything. So I guess like.

D

Usually the answer is victim, but I don't know.

A

Right, I mean argo: okay, sorry, yeah.

D

A

Tekton versus argo.

D

Well, I'll argo is more like a github guitar like flow as like the question that you and I was asking as well at the beginning of the show uh where you, uh where you you, like, usually with gtops like you, would uh your manager, your full, uh your without logo, your knowledge, your full resources like for for, like a secret communities, and all that say like all that stuff when tekton is more like uh it's more, it's more focused on the on pipeline and on the on the developing pipeline. I mean at least organism.

D

Now it's like uh the differences between the products. I would not be able to say I'm not really an expert with argos, but uh but uh like usually like the way I would I use. The argos is mostly on cds, which was doing a great uh a great great thing like he would be able to uh deploy.

D

Like all my uh all my secrets and everything like that and uh my service accounts certain things and if I and uh and text on I would use the python you know and especially like at the detection catalog, which is uh which has a lot of tasks. So you can use all that solubility and comfortability.

B

I made a shameless plug on one of my uh recent argo videos on youtube chat, so so folks would like to check that out. Cool.

D

I think I need to I need to because I don't know very well, I just used it, but I didn't do it.

A

So I dropped a link in chat as well to the um a talk that cmok did on using tekton and argo together in kubernetes and how awesome it is so yeah. Hopefully folks find that helpful. Did you say you dropped something in chat down.

D

I don't see it.

B

Oh um in the in the youtube.

A

Chat: okay, let's see to the youtube chats right right below your comment: okay, it's just not coming through on restream, so I gotta.

B

A

Okay, like re-paste, it basically.

B

It's okay, the the link you paste for cm like that that encompasses.

A

Everything, so that's good. Okay, I mean, if you're sure I wouldn't want to discredit anything that you did anyway, because juan is awesome. Oh.

D

Thank you when you look.

A

Up awesome in the dictionary there's a picture dejuan next to it, um gracias, no problem, no problem at all, uh yeah cool, all right, um awesome, great show, thank you all and we will see y'all soon stay tuned to the calendar like and subscribe. As always, and um we will see you all soon.

A

Stay safe out there folks.

C