Red Hat OpenShift KubeCon EU 2021 Office Hours, 7 May 2021

Previous Meeting

Next Meeting

⏯

youtube image

►

From YouTube: KubeCon EU Office Hour: Replatforming to Kubernetes With Move2Kube

Description

Moving your existing applications to Kubernetes is hard, but tooling exists to make it easier. Meet the Move2Kube team, who will talk about how to handle your cloud native migration, and take your questions about migrating apps.

A

Now, good morning, good afternoon, good evening, wherever you're hailing from welcome to another, the last actually cubecon eu session uh office hours for us here on the red hat team, we're talking to our conveyor team right, the move to cube team. uh Yes, this is this- is the conveyor talk that everybody has been waiting for and I've been anxiously waiting for. So I'm chris short, I'm going to hand it over to josh burkus from the open source program office. Josh. Take it away.

B

Thanks, uh I'm josh burkus, I'm red hat's community person for everything cloud native, um and particularly that includes helping a lot of people move their applications onto kubernetes.

B

um If you are in operations, I am sure you have at least one application um that is still on some other platform that you would like to move and to help you do that. We have here the team behind move to cube um a a project for enabling applications. um Amit, you want to introduce the team.

C

Sure josh hi everybody, I'm amit singhi from ibm research and uh my team here, ashok nava, akash, pablo and hari. All of us have been working on this exciting new project move to cube. Actually it's not new for us, but it's new in the community and we're looking forward to talk about that.

C

So should we get going, I should be wanna share the slides.

C

Okay, so I think everyone's kind of familiar with kubernetes, I'm sure, uh so what we're going to talk about is, if you're, not on kubernetes today in your organization, uh how can this tool called move to cube help you get there with less pain and less time? Okay, and uh so, let's move forward to the next slide. Please um so move to cube is part of this open community called conveyor that we uh kicked off uh with a few different projects, contributed initially by red hat and by ibm research last year.

C

So you can find it online at conveyor.io, it's essentially a community of people who are really passionate about kubernetes and helping others modernize and migrate, their apps to the hybrid cloud leveraging communities, uh and so it's also a collection of tools and best practices on how you can re-host your applications over to kubernetes how you can read platform now we can refactor them even re-architect them so that they run uh really well on kubernetes.

C

um So moving to the next slide, please uh it's just like a quick view of what what's there in conveyor. Today we have uh five different projects. There um move to cube is one of them. We're going to talk about that that fits into the re-platform uh use case. So if you have an existing application infrastructure, you want to move it over to kubernetes, but you're, not really looking to redesign the code or the implementation, and that's what moviecube lets you do. On the re-host side, we have crane that helps.

C

You migrate, your applications between kubernetes clusters, so it migrates over the state.

C

You know so it's like a live migration and then we have forklift that will let you take your vms and move them over to cube word so that you've got your vms running on communities and then finally, there's tackle on the right-hand side that, let's you know that actually does uh detail source code analysis to analyze your applications and help you uh move them closer to containerization uh and then polaris that helps your organization measure the impact of changes that you're making in your software delivery so that you know you're actually moving in the right direction and improving the metrics around software delivery performance right.

C

So a variety of different tools uh with that I'm going to hand it over to ashok to take us through move to cube over to you.

D

Okay, thanks ahmed and thanks team thanks for everyone for joining in um so here, let's have a quick view of what motorcube is so that uh you have a preview of it and feel free to ask your questions uh about it. So, when you are writing your application or you have been hosting your application production, it might be in any of this platform right.

D

You might have already been in docker swarm where you have already containerized your application or you might have your application running in cloud foundry or you might have be having your j2ee application or just running in your bms right um motocube is about helping you automate the journey of getting to the kubernetes platform in the most cloud native way um by the most cloud native way.

D

What we mean is uh what is the most general way if you are developing an application from scratch for cloud native, you create your docker files or cloud native, build pack and containerize the application you create your deployment dmls and all of it. So multicube takes all the artifacts that you currently have and tries to automate the process of creating all these artifacts that you see in the right-hand side. uh It might be your kubernetes yammers, the dockerization scripts like docker file, cloud native, will pack s2i images, helm, charts, customize, yamls or their openshift templates.

D

It can even create an operator or k-native artifacts if you are interested in them and also uh if you are, if you are putting your application in production, you might be interested to see a ci pipelines like techcon, so it will create all those artifacts for you which will be bootstrapped and you are almost close to the deployment in a matter of minutes um at a high level.

D

If you are looking at what motor cube is it helps you discover all your artifacts helps you containerize it translate and create the right destination, artifacts and customize it for your particular deploy. We will see a quick demo of it in a minute, um so let's uh motor cube is completely into open source. You can just head over head over to this uh project and look at everything there.

D

It's a command line tool, but we have a web interface which grabs the command line tool and all this functionality, let's quickly look at the web interface, so this is the motor cube website to get your web interface going. All you need to do is go here, just create an empty folder and copy this command over here. So let me just quickly do that for you and um motor cube will be up and running in a matter of seconds right. Okay, now motor cube is up and running. Let's head over to the website,.

D

Okay, so here is a motor cube interface: let's create a new application, um since kubecon is going on. Let's call it coupon demo.

D

And what I'm going to do is to now take some source artifacts right, the platforms that we saw, let's take an application which is multi-component, which has multiple languages, let's say an application which has some blue line, components, java, gradle, node.js, java, maven, pythons and stuff right. Let's say you are an application which is really polyglot and has all these languages. All I'm going to do is to zip them all into a zip file and upload it and let motor cube do the job for so what is motor cube doing now?

D

So what it's doing is it's going through each and every folder and every file in here trying to analyze. Okay, is that a manufacturer? Is there some runtime information about cloud foundry? That is there or does? Is that a docker file? uh um Is there a go language application? It is trying to find all of that and tries to find any links right. If the server is talking to something else, it tries to find all those links and then it will come up with a proposal.

D

Okay, these are the applications I found in your folders, and this is how I think it can be ported to kubernetes. So in this case it's saying, okay, I found a service named golang and it can be containerized in three weeks either with the docker file, s2i images or cnb, and it has more details of that. um It's just a ui to uh yaml that the command line tool generates. This is the yaml. You will see if you are using a command line version of it.

D

So and then you have your java applications and all of it are listed over here. So what we will do now is: let's go to the next step: let's create the artifacts based on the plan that we got so what I'm going to do is to give next and then I'm going to tell it to containerize all my applications. uh These are the services that it found and then what I'm going to do is to okay.

D

I am interested in a docker file, let's see so I am telling you to create a docker file and then it says that okay, I find a java application which I can host to either jboss liberty or tomcat. So let's choose the jboss one in this case, and then let's say it's asking: which platform would you want to really deploy to motor cube, understands the variance of kubernetes in its most elements right, the kinds, the versions and stuff?

D

So we can really target the artifacts for that particular version that you're targeting it can be even your custom. Cluster motorcab has base where you can collect information about your particular cluster and really target artifacts for that version, and then you just click next and it asks basically the questions. It cannot find an answer to from the source. Right, for example, uh you have your application, which uh ingress url. Do you want to post it to? So it's ask those questions. Let's keep pressing next.

D

Okay and then um what it does is it asks the more uh specific questions about your destination like in this case it's asking which image registry do you want to deploy it to? um Let's say I will go to us.icr.io, which is the ibm cloud registry, and then I will put my images into an interface called n2k demo in the registry, and then I'm going to use the existing full secret.

D

And it's asking for the ingress url, I'm just going to head to my cluster and copy the url from there and put it over here and just copy this in a minute, and let you know why so. The next thing it asks is a tls secret which, in case of ibm cloud, is just the first part of your url by default. um So I'm just putting that in and then what it is doing is it's now creating all the artifacts that you are interested in.

D

Let's see what it has created for us uh now, so I'll just close this for a minute, and um you will see that all the artifacts that we need are will be generated.

D

D

So I'm just going to download those.

D

D

And let's see where the effects are.

D

Okay, here here is a set of artifacts that it has generated.

D

um As you can see, let's explore this.

D

E

What it has generated for us is.

D

It has basically taken your source folder and it has uh created the docker files for uh your.

E

D

Components in this case for golang, it has created a docker file for each of this conference. It would have created them in addition to that, it has also created all the scripts, uh all the yamas required for your deployment, like the tekton pipelines and uh the docker compose file for your local testing, the health charts for deploying your helm chart and the k native services.

D

If you are interested in serverless, uh the customized uh base, uh mls overlays for your different deployments and uh the openshift template uh the operator, helm-based operators and the normal yamas, if you're just uh interested in deploying it directly- and it has also created some helper scripts for you. If you want to try locally right so.

E

Let's quickly, try it.

D

Locally um in the meanwhile, if there are any questions, I can definitely take it.

D

So, let's go to this folder.

D

Going to do um again.

B

Yeah that I prefer yeah for one of the questions when you actually get to this, if you can um well actually you're already showing some of this, so one of the questions was about whether or not you could customize the artifacts um that it creates after it creates them, and it appears you're showing that now.

D

Absolutely uh these are the artic. So the way we look at motor cube is it tries to get you to ever 95 percent of where you want to be, and it's a one-time process, and then you can edit them as much as you want and deploy. That is one kind of customization right to cater to your.

E

D

In addition to that, what moody cube also allows is to customize using some scripting languages. We will get to that in a minute.

D

Okay, um so let me quickly show what it does and then um I'll just answer your password okay. So here there is a script which can be used to build the images I have already built it uh so that it does not take time. So I'm not just going to do it, and then there is a script for pushing the images and once you do that, all you need to do is to deploy the helm chart.

D

um I already have a cubectl in my context, so what I will do is, I will just click on a helm, deploy help and it will deploy the helm chart and get me the urls that I can access the application in a minute.

D

So it's um so um while uh it is deploying the artifacts uh joshua the base that we think of customizing or uh the move to cube for your needs is there are two ways that we generally think of. One is the scripting. Basically so you have you have your default multicube and then there are some starlock scripts that we use to say. Okay, this yaml is fine, but I need some annotations that needs to be added to it um or you. You need to um change okay.

D

I use this particular um full policy, so it allows you to use the starlock scripts to do it. Starlog is essentially the one that is used in basil. The second kind of customization cube also allows us. You have your project structure in different way. Right. Motor cube by default understands a lot of um languages and different folder structures, but you might have a specific way in you in which you put your customizations, so we can allow plugin based containers. We will also look at that in a minute.

D

Okay, so here is. It has deployed the application, so I can just head over here and I can see that the application is already posted in the cluster. I can and I can access the link over here. Okay, so that's a very quick demo of it. So what we will look at next is to answer josh question. So, let's look at how do you customize the artifacts that are generated by motor cube for that? I would like to invite hari to give a quick demo hari.

F

Hi, um so can I share the screen or.

D

C

F

You see the screen.

D

F

Okay, so hi everyone, so uh I've prepared a small application. um It's a it's a website with a rest api. uh The website is using java and the rest api is using a python server. um So we have a deployment for the website and the rest api. We have services that expose those uh deployments and we have an ingress so that we can reach them from outside the cluster.

F

um So basically somebody asked about: can we customize the artifacts, uh so we provide a method for customizing uh using a scripting language called starlark, it's very similar to python. uh So here we have a starlog transform and I'll just show. The we have two transforms here. um One transform is doing uh simply adding a new annotation, uh it's a common annotation to all the resources or the deployments and another one is simply setting the replicas.

F

These transformations can also interact with the user. So here we are asking the user uh what should be the number of replicas for each service and uh we are setting that replica and um uh the starlight transformations, uh so you might be familiar with tools like customized and so on. uh Those allow you to customize those resources, but star-lord gives you even more control uh because it has access to for loops uh if conditions and so on and uh more complex data structures.

F

So here we are going through all the containers and setting some resource constraints. So if your organization has some default uh limits, you can set that uh so these uh two files are the transformations that I'm going to demo. So um I already have mood game installed, uh we'll just do a translate and with the minuses flag, we'll give the source uh yamls and then, with the minus t flight we'll just give the transformations.

F

So now it's asking this is these are the same questions as we saw uh in the previous demo, um uh but fewer questions because we're not doing a full translation. We are simply uh doing some transformation. So here these are the questions that we configure.

F

What should be the number of replicas for the service rest api, um so we can give like uh two and for this three um and you can send defaults and uh hints and everything uh so now it has done the translation and uh we can look at the output, so the output, my project. So here uh we see all the same artifacts as before, uh we'll just look at the kubernetes that it generated. So uh what changes did it make so here?

F

First of all, we see the common annotation that it added uh to the deployments.

F

uh We can see that it changed the replicas uh to the number that we specified um and also uh it also set those resource constraints for each container in the deployments, and also um it's more powerful than that, because uh sometimes you might have um older versions of this yaml. So if you look at these deployments, these are this. One is extensions, v1 b1, which has been deprecated.

F

uh The newer version is apps v1 deployment. But if you look at the output, uh we have uh automatically uh converted it to the latest version, so these are apps v1, apps, v1 uh and same for the ingress. The ingressors are extensions, v1 beta1, and now it's a it's a networking kxio v1. So this is more than just a simple string change, uh because if you look at the older uh increase it, I had a different, especially here it had a different format: syntax um in the newer version.

F

uh It has a different syntax, so it is actually understanding uh the services, the paths and uh doing uh the correct translation. It's not simply changing the api version string so and um yeah. So that's basically what I want yeah if anyone.

E

Asks any questions.

D

Yeah also, I I see an interesting question on the windows containers. Yes, um motocute can handle windows, dot, net framework and stuff, and we have an interesting demo of that too. We can look at it george. Whenever that's the time is right.

B

Okay, um yeah and actually so we're gonna. Are we gonna do that demo later yeah? We can do it now: okay, cool yeah, because there's a there's, a follow-up question in that right, obviously, because it's more complicated than just creating the home charts.

D

Absolutely um uh naba can I invite you to give a quick demo of the windows platform, translation.

G

Sure ashok, thank you now. Let me share my screen.

G

G

Is my screen visible.

A

It's coming hang on. Okay, there we go.

G

Yeah, thank you so uh before I start giving a demo about the windows container, uh I just wanted to give you uh a brief introduction of it. So uh windows container support uh in openshift started off in the later part of last year. Sometime around december, and uh one of the intentions uh was, there are different parts in which uh existing windows workloads, uh which have been there around and being, are being used by enterprise applications and so on.

G

uh How could we find a path of for these applications to be re-platformed so that they can avail the benefits of openshift or kubernetes environments?

G

So uh one one one approach is to use open shift virtualization and lift and shift the workloads directly and put them into vms as it is, but they won't be using the benefits of the country, containerized mechanism and uh mechanisms and orchestration mechanisms that openshift provides. So the second approach is to provide a windows container uh wrapper around this uh framework applications, especially for legacy application. uh Sorry, uh the existing framework applications, which are there such as dotnet, four point dot net framework, four point, eight and so on. The third approach is uh re redesign.

G

These applications around dot net five, which has emerged recently and avail the linux containers wrapper and then deploy them onto open openshift. So these are some of the parts that uh windows containers as a windows workflows can avail to run on openshift, like advanced platforms for container orchestration.

G

Now uh let uh the demo that I'm going to give you is uh for both of these types of uh frameworks. One is for dot net 5 I there is, I I'll, be using a dot net file, uh app and also a windows communication framework service app, which is uh making use of dot net 4.8.

G

Now uh these apps will be accompanied with certain plugins, uh just as what uh ashok mentioned. So without actually changing move to cube, we can insert plugins which can understand these apps and generate the desired containerization that is required for this app. So these are the two plugins one for dot net buy and one for windows, wcf app and now, uh let's use the ui to upload this app to the uh to move, to cube and see how it processes it.

G

So, let's create an app.

G

G

And let's upload our applications and the corresponding plugins to this app.

G

So the flow is very similar to what ashok just showed, ah so the applications and the corresponding plugins are processed in multicube and it will generate a plan and it will show the various strategies as shown below. So once this is shown uh here, you might have noticed that there are two apps are detected, one for dot net file and one for the wcf and new docker files generation format is recommended for each of these apps.

G

Now, once this is done, we move on to the translate phase to consume this plan and go through the translation process which was shown before so we select both the services. Each of them have been detected separately by the plugins and we choose the new docker file generation mechanism.

G

We use kubernetes as the target platform select the services, both the services to be exposed and diving response, and so on will be generated.

G

G

So this process, once it is done, it will generate the artifacts that we want.

G

And we download it yeah and we have a downloaded version of this already to save time. um So, as you can see, this downloaded uh re-platformed uh artifacts target artifacts contains the docker files that are required to contain constructor containers for each of these applications, so, for instance, for the wcf app let it load close.

G

Wcf app it's a two phase container.

G

Docker file that is generated want to build the app uh a scene can can be seen from line 15 to 19. The second one is the run stage where the built binary can be run as part of the runtime runtime container uh and uh similarly for the w- uh and you know you can write the 4.8 uh framework that has been automatically detected and used to then find the appropriate base image.

G

Similarly, for the dot net 5 app, we have a docker file which uh use which produces the, uh which has the images for generating the generating the published artifacts for netfi asp app and also to run the runtime to run this published artifacts.

G

Now, let's see how we can build this app using the container docker file that has been produced by motor.

G

Cube so we build.

G

It and, as you can see, the image has been successfully built now we run this image and please note that uh this is a windows container and not in its linear container, and this is running natively on a windows bare metal machine.

G

So I'm going to run this service.

G

And this is a native wcf dot net framework service and I'll run the client to hit this service and get a response from another container.

G

So, as you can see, the service container has been detected and a request has been sent and a response has been obtained, so the exposed ports. So this is about the windows, containers and the support that motor cube, provides to containerize windows applications and available windows container facility and eventually find their way into openshift platform. So if you have any questions, please let me know.

B

Cool um one of the ones from the original question also is that obviously, one of the things that um somebody with a.net app is going to be looking at is migrating not just from not on kubernetes to on kubernetes, but also at the same time, they have to move the application from.net framework to net core so that it will run on kubernetes.

G

Yeah so uh currently motor2 provides a re-platforming solution.

G

What is being suggested by the question is a re factoring solution in the sense that you have a dotnet framework which uses a older version of dotnet.net, let's say 4.8, and you want to convert that to dotnet5 now that is a that requires redesigning the application to consume the advanced features of dot net file.

G

Now, that is not what multi cube provides, but if there is a uh uh the 4.8 version of framework or 4.7 version of framework or similarly, if I need five version of the framework motor cube, detects the uh apps requirements and accordingly generates the uh containerization uh strategy so that that is what multicube tries to. That is the gap that motorcycle tries to fill in.

B

Okay, we have another question unrelated to windows, containers or net containers, um and I think this is probably more of a migration advice thing, although I'm curious as to whether or not moved cube can help with us. We have um an admin here whose co-workers have do a lot of work in jupiter notebooks.

B

um You know the data analytics platform um and the process of moving that infrastructure onto kubernetes is complicated. Is there any part of that that move can help with.

D

um Currently, we are not uh handling any um uh jupiter notebooks. We are not tried it, but if there is a interesting use case, we can discuss about it in the community, so I will request the team to come, come over to the conveyor slack channel uh in the kubernetes workspace and we can have a discussion around it. The exact use case on what we are trying to manage.

B

Yeah, um I mean I'll say from my own personal experience that there are some major issues with running jupiter hub on kubernetes that are limitations of jupiter hub itself, um which is is um particularly around um access control. um The um which is, is not something that we're gonna overcome and conveyor. That's kind of up to the jupiter team, um the um uh followed question.net um I in order to move on to kubernetes, do you have to re-architect as microservices?

B

Or can you just take the monolith and turn it into a monolithic pod service.

G

B

It is possible to.

G

Yeah, please please continue.

D

Yeah, so basically it is about this right like what, in, as you saw, the different tools in the conveyor community. Each has a particular scenario that we are trying to uh decorate. um The first thing that we are trying to do with re-platform is take the monolith and run it as it is in the kubernetes platform. So that's what motorcube will have once you are there.

D

You can then refactor your application, big monolith into multiple components and deploy it, as there are tools like tackle and a few more which can help you with that in as part of the conveyor community.

C

I may add one one more point there.

C

You know the refactoring to microservices tends to be driven a lot more by the needs of the business rather than whether the black weather platform or technology considerations, uh because refactoring to microservices means additional work, not just for the refactoring, but also for them. How you change your devops processes around the microservices. So in a sense it's uh it's an orthogonal point. Although it comes together with re-platform kubernetes, often because the microservices, let you leverage the benefits of kubernetes uh better.

C

So it's not mandatory, but if the business requires it, then you know it's good to consider. If the business doesn't require it, it may actually be unnecessary overhead. If your app is not that large that it, you know, won't run in a container.

B

Okay, um so another um again, not a net question, but another question in general about migration and move to cube uh out of slack, which is um somebody has uh source code, repos that have scripting for building vms and deploying them, um and uh they wanted to know whether or not there's any tooling move to kube to help them. Convert that vm generation code into um docker file generation code.

D

Sure so the way mutual cube handles this up. So, for example, they have a vm generation code, but there is some uh source code that is at the end of the day, contributing to the vm. So if you have an application which is being built and then the binary is being put there, the way we do it in motorcube is that we target the source directly. So let's say you have five or ten gig repositories where you are taking the code from and then compiling it and putting it into the vm.

D

You just point motorcube to the source code and it will find how to containerize each one of the components and can put it into cooperatives.

D

So we do look at the artifact based artifacts. That's how we have.

B

Okay, uh we'll see whether or not they come in their actual. Their actual use case apparently, is vnfs, so so telco vms, um the um and they're trying to move to cnf, which sounds like the kind of use case that we might at some point have tooling to optimize for um given how much of it we do do we have that right now is that is that, like a project that we have right now, vnf to cnf.

D

uh We haven't tested that particular thing, but uh our colleagues in ibm research are all into our stuff, so cool yeah.

B

That almost feels like it might eventually be a separate conveyor project.

A

B

Embeds because it's kind of a special use case.

C

It is, and in fact that's a really interesting one. I think one of the challenges right now is that kubernetes itself is not necessarily optimized for networking workloads like vmfs, because they need certain performance guarantees at the networking level around jitter and latency, and things like that, yeah so yeah. So there's a lot of interest and work, but it's exploratory because we're also looking for the platform to catch up.

B

Well and that's why I was thinking we would probably need like it, a special tool set right, because when we're looking at vnf or we're looking at cnf, um I know from working the guys is you're not looking at just deploying regular pods and stuff you're generally going to be using something like kubernetes containers with multis. In order to have that whole network infrastructure in the deployed application.

C

Yeah yeah, the other other project, that's kind of interesting. There is forklift because it it moves your vms into cube work. So I mean you might consider that if you're not able to containerize some of these workloads again, you still have the challenge that the virtualized network may or may not satisfy your performance requirements.

B

Yeah, it's still a work in progress.

B

Okay, I think we're on top of the questions for now. So if you had a next part.

D

Sure so, actually we just saw a quick demo of uh network, so uh the other big platform request we generally get is cloud foundry and generally people uh use a spring boot along with it, and um we have a pablo from our team who is from our tokyo lab, who has been working really into it, trying to see uh how good you can translate. We can probably have a quick demo of that.

H

Sure let me share my screen.

H

Okay, can you see my screen.

A

H

So yeah um yeah: this is a quick demo on basically using multi-cue for migrating a cloud foundry springboard based application, um it's a kind of simple application, and you can see here the pum.xml file, there's a lot of information in terms of the packaging and the java version etc, um and the the the main challenge when we are trying to uh migrate.

H

This type of applications is the diverse source of sources of configuration files and in that sense the the challenge for multi-q is trying to capture all of them, integrate them aggregate them and then use that information to generate the um target artifact, for example the docker file. So based on this application, you can see we have a cloud foundry manifest. We have application properties which is uh common on extreme applications.

H

So let's see how much the cube works in this in this context, so um for this demo I will use the um command line version of the queue. So let's say I have already moved the cube installed too. So we can do move to cube.

H

Here you can see it's the um just a path to the actual application, so here we are just running the uh plan phase same as um as you again now I did before on the ui. Here we can see how the output that we obtained from the um from the command line. It takes a little bit of time, uh but it should be done um just in a moment.

H

Okay, so here you can see how multi-cube generated the plan file and it's located here so based on that we can just start and execute the second face, which is the translate, and here you can see how much cube identified this as an application based on cloud foundry, and we we go through the same questionnaire. We saw on the ui, but in this case we can interact directly with the command line.

H

Here you can see, move the cube. I tell us that one of the available options is springboot because it identifies the application using that framework. So we go with that and the rest of the questions. We, um let's uh keep it the default values. Just just for this demo.

H

So, based on that, we have um an output folder that was created with the name, my project and, let's, let's inspect what is inside. So we can do code.

H

So this is the my project folder, that I moved to cube just created and here specifically in the source folder. Let's take a look at the generated docker file, so here on the right, we can see how to move the cube first of all, to give kind of, like a reason, on several several pieces of information that capture from the application and, for example, it has to uh install maven, because we know this application is made and based right. Based on that.

H

Also move to cubes adds these instructions to the docker file to generate the um a final uh deployable file.

H

We know that it's a war based file, so I moved together, also takes care of that, um as this application is not based on the embedded server that is part of um or the default behavior on on, springboot uh multicube selects the more appropriate uh image from jboss wildfly that is compatible with the actual java version that this application has also multicube takes care of um all the ports that needs to be exposed and, finally, in the last line of this generated, docker file move the queue, identifies and builds the paths to the deployment file and copy it to the um deployment of the folder.

H

So you can see how much the cube is able to support this. This type of migration, where we have several technologies, such as cloud foundry spring boot and and what is the kind of output that it can provides to the user, and from here let's say the user can continue right. This is uh probably there are some other changes that can be done, but I probably give the multi-cube gives a good candidate for for the deployment process so that that is basically how how we are handling currently this. uh This scenario.

D

Sure, thanks pablo, so, uh as we just saw, it was a very a small uh cloud, foundry app and it was able to find it was a cloud foundry springboard and get all the right parameters for you and, as we usually just talked about in the start of the sessions, um this can be customized for your specific setups, uh based on the names.

D

So uh yeah, so the uh the other thing I would like to probably show is uh um there was an initial question on: um how do I try out uh motor cube right? um You can definitely head over to motorcube dot, conveyor, dot, io and try it. In addition to that, we also have a category scenarios that you can try. We are building more and more scenarios, but we initially have a seed. um We can have a quick preview of that.

D

uh Josh uh sorry, uh can we have a quick preview of that? Is that fine, okay.

B

Yeah um well you're, setting that up uh quick question again with a lot of questions that are more general migration and not necessarily things that are included in the movie cube um functionality. Right now, um which was somebody was actually asking about. You know they have a set of rules about um network access. They have a bunch of network access rules around their current vm infrastructure and they were asking about how uh how to convert that to uh network policy in kubernetes.

D

Sure uh so by default, so for example, uh as you might see here uh in this particular demo, where it's about docker composed to kubernetes, so the docker compose environment has a concept of network. So if that source platform has a concept which is very similar to network mutual, can understand that and create the writings for you. So, for example, if there is a network here, it will create the right network policies for it, and you can, um we can even be extended to create, is theo policies and stuff.

D

If that is what the organization requires so um by it all depends on the information that is available in the source code and if you want.

A

To give additional.

D

Information, you can always always query the user to get additional information and import them as their policies.

D

Thanks um akash, do you want to quickly control.

I

This yeah yeah sure thanks ashok. So if you are new to move to cube, then you don't need to worry. Move to cube is available on catacorta. You just need to go to move to catacoda.com, and then you can quickly get hints on using move to queue so I'll quickly uh go over one of the scenario that we have.

I

So you don't need to install anything on your laptop on your local machine and you can try, try out move to queue directly on the catacomb. So the first step we are installing move to queue.

I

And now move to cube is installed, so we have a full slice demo of move to cube here using our sample docker compose file, but for this office hours we will use an open source project and try to generate the deployment artifacts for this open source project, using move to cube so I'll, just clone that.

I

Repository and next I'll run move to cube, translate over the.

I

I

So multicube is asking some questions and it is trying to internally. It is creating a plan for you and it is going through the folders and trying to identify whatever services it can find, based on the docker files and the docker compose files which are present in the source.

I

D

So yeah and, as you can see like motor cube, um supports the multiple sources like cnb and stuff. So uh if uh for a particular client, you are interested in only one of these sources, uh you can select or deselect it to speed up the process. For example, cloud native buildback images generally take a lot of time to bring in the gb of image sizes right. So based on that, you can speed up the process.

I

Sure so these are the services that move to cube, has identified based on the docker files and docker compose files, and now it's asking whatever services we want to select and then what container is the technique? We want to.

I

Use and then it has to select the cluster type and.

E

Whatever services.

I

We want to expose, we can select or deselect here and then provide the paths to be exposed.

I

For all the services it's asking on which path we should expose the service and then some more questions. I'm just quickly going with the default answers here and we will go over to see what are the artifacts that move to cube has generated for us. So here are the artifacts. We have a readme file which, which is giving the instruction like how we can deploy the application to kubernetes following these steps, and we have the tecton ci cd pipeline related artifacts and the help chart and emails which are required for deploying to kubernetes.

I

So in this way we can, if we have docker compose file, then within minutes we can generate the target artifacts using move to queue, and I would encourage all of you to go to uh our tutorial category scenarios and provide your uh experience and feedback on communities select channels.

I

Over to your show, thanks.

E

D

E

D

There any questions you can take it uh or I will quickly flash one screen there. With that uh context, you can have more questions.

B

Yeah, I have a question about a completely well. Actually, it's not completely different topic. This is a getting started. Question um which is somebody wants to know, do they have to have docker installed um in order to use move to coupe.

D

um So was the question to do: is docker required for trying out motor queue.

B

It was docker required if they want to use move to cube on their desktop in order to convert source code.

D

Not necessarily um so, um if you need to use uh um cloud native, build pack containers which are based on images, then at least you need to have either pod map or docker either one of it it supports both of it. You have a cloud native, build pack is not required in your base machine. You don't really require uh containers as such, uh you can use the command line, tool and uh you'll be all set.

B

Okay, but a developer, whose laptop is, for example, running windows home, would need to find some other platform to work on then.

D

Yes, be yes, uh we have tried it on wsn, so that could be something uh they can try it out. Okay, so.

E

D

They have the windows edition, they can just install wsl2, which will give them a linux environment to work with or the other way around is they can bring up the docker image in the windows mission in which.

E

Case they are just using the.

B

Browser: okay, go ahead.

D

Okay, great so um we saw a lot of functionalities of motor cube. Many of them are already open source. It has been tested in a lot of scenarios and uh right now we have the 0.2.0 alpha version and 0.1.0 release versions in the open. uh We have exciting plan over the next few months, uh where we are looking at taking all this functionalities and enhancing those based on what the community needs.

D

So, if you have any requirements or any questions or any use, cases do head over and feed us the input, either in the slack channel in ktes, dot, slack dot, io um slack.ktas.ivo and uh the conveyor community channel over there or in mood cube hyphen dev, google group you can head over there all the links are there in the moodlecube.convert dot website right. So the few of the things that are coming up is we are planning to support custom templating for our yamas.

D

We support a lot of artifacts right now, but what if you have a crd so that functionality is coming in quite soon and we are adding more enriched containerizers, which can use more um interesting techniques, machine learning techniques and stuff to get more parameters from your applications and um all the things that you saw, which are plugins right now like spring boot, support windows, container, support and stuff, uh we'll be adding it as part of the base uh motor cube tool tooling itself, and um we are looking at netflix voices, argo, cds and more- um and this is the prioritization at this point.

D

All these things are subject to change uh depending on your needs. So do head to the channels- and let us know.

D

Yeah so josh, that's. uh I think I would like.

B

Oh yeah, so I was just looking to see whether or not we had any additional questions from that. um The um I have a couple, but we're actually kind of at the end of our time, and people are starting to sign off for the day, because it's the end of google.

A

B

um So um so I'll just say, if you want to wind it up any any last thoughts on migrating to kubernetes and move to kube on where we are with cloud native today um uh from the team.

B

Oh wait: wait, wait.

B

Ooh, okay, actually one following up on the earlier uh spring boot thing.

B

Oh no sorry this is advice for working around the windows, home restrictions, um so I'll. Just let that user read that um so last thoughts.

B

D

Yeah, so um this is a community effort. All of us are there uh to get your feedback and uh improve it, and if you have any contributions, we have um interesting stuff in cube right and go language javascript and lot more. So we value your contributions to come over. We will help you out getting started and uh taking your contributions into motor cube, yeah and a.

E

Team, if you have any.

D

Other points please do.

C

Well, thank you to the folks who spend the time to listen to us and for the great questions. I think that they'll definitely make us think a bit more about some of these dimensions.

A

A

So that's all the time we have, and I appreciate everybody coming on to talk about move to cube. I appreciate everybody listening uh and tuning in out there uh kubecon might be over, but openshift tv lives on forever uh at noon. Eastern today, uh 1800 cest we're gonna, have an open shift commons briefing with kristin mclemore from red hat talking about scaling the portfolio wall, so please feel free to tune in to that um and as always thank you josh for running these office hours during cubecon. They are.

B

Always informative, thank you, and and thank you for for hosting and streaming us, and we will actually have future office hours. Events on openshift tv. um uh We have some teams who are interested in doing that. So just um uh subscribe to openshifttv on, say, twitch, it's a good way to be informed about upcoming, shows.

A

We have a streaming calendar. I drop the link in chat if you're curious go ahead. Subscribe to that, if you want youtube, is a good place to.

B

And if you are in the process of migrating to kubernetes, um like I said conveyor, io is not just a set of tools: it's a community for practitioners for people to help each other around all of the many complicated problems we have with migrating. um So please join that uh for some peer-to-peer help and sharing um with the community of people who are facing the same problems that you are and.

A

Thank you very much. Thank you all for your time today. Thank you very much.

D

E

E