Cloud Native Computing Foundation KnativeCon EU 2022, 19 May 2022

Previous Meeting Next Meeting

⏯

youtube image

►

From YouTube: Lightning Talk: Docker-free Functions for Knative - Zbynek Roubalik, Red Hat

Description

Lightning Talk: Docker-free Functions for Knative - Zbynek Roubalik, Red Hat

Knative Functions allows developers to build and run their applications on Kubernetes without knowing anything about containers. Instead, the source code is built transparently using a local Docker or Podman installation and deployed as a Knative Service in a few simple steps. Sometimes, however, a local build in Docker or Podman is not possible or simply not the preferred option for a developer. Luckily there is now an alternative. In this lightning talk, Zbynek will present the latest feature of Knative Functions: On-cluster builds, which frees users from creating container images locally. The presentation will describe the different options for creating container images within Kubernetes clusters, the current status, and the plans for on-cluster function builds. A live demo will show the on-cluster builds in action.

A

Hello, everyone, uh my name, is maniac robotic. I know that my name is pretty hard to pronounce, so I don't feel about bad about this. I will talk about just briefly talk about how we can build functions without docker or podman, so why this is important because, as you probably all agree, some people don't want to have docker on their machine or port man. Some people are even not allowed to have have this kind of technology on the laptops, or some would like to have some flexibility.

A

So before I go, I will just just quickly give an overview of functions. I suppose you learn all the good things from lenses and mauricio's talk, but just to highlight some things. So basically we are using the kn func plugin to drive the functions, and the main thing is that the container images are being built uh by by build packs. So this this step requires podman or docker on the developer's machine. Once the container is produced, then it's deployed as a standard kubernetes creative service.

A

So basically the standard for flow and developing function is just I will create. Let's say I would like to create a node function, so I will run the first command. Then I will implement the business logic and then I will deploy my function and I can repeat those steps still forever, and this is the step that requires spot man or docker on developers, machine how we can solve this, so uh we have already kubernetes on the cloud somewhere on internet. So why don't we use resources on the on the kubernetes cluster?

A

There are a couple of options. uh Some of them is some of one of the option is basically, we can use tecton to define pipeline that will drive the build and deployment of the function on cluster.

A

There is obvious a problem: how to get the source code from developers machine on the cluster. So again we have multiple options.

A

So the first, let's say obvious option is that we can push the source code to some git repository and then just fetch the git repo on to the cluster. We are tekton and then execute build and deployment. The second option could be to directly upload the source code from the developer's machine to the to the cluster.

A

This might be beneficial for let's say uh when we are developing the function and we would like to iterate fast, so we don't want to go through for a good repo.

A

Another option might be not to use tecton at all and we can just package the source code as a as an image and push it to some container registry.

A

This has been proposed by mink, for example, but I suppose that the tecton tecton approaches several benefits, because this way we can easily plug the ci cd solution to our system. We can have some stability, then we can also cache cache the built artifacts or even the image layers on the cluster, so the builds are done, faster, etc. So I will just quickly show you a demo.

A

So basically this is, uh this is not function. It's very simple, just generated function is not important, but the important thing is that, basically, that I will I will use the approach with the git repository. So all I need to do is to specify the git repo that contains the source code of the function and specify a build type to git, because we have limited time I already pushed the source code to the repository. So all I need to do is to execute execute the deployment.

A

So I will just write func deploy and, as you can see, it starts doing something I will go back to my presentation and this diagram might help us understand, what's happening under the hood, so basically that guy on the left, that might be me or maybe you or some other doctor hater, so the funk cli then creates all the resources for me. So basically it creates tecton pipeline with tasks. So there are three stars at the moment.

A

So first task basically clones the repo and stores this on the persistent volume that's been as well created for me, then I can execute the build text task that will build the function, build the container and store the image digest. And finally, I can use this image digest to reference a container in my container registry and deploy deploy the function as a cognitive service.

A

If I go back to my to my pipeline, it's still, as you can see now it is deploying to the cluster. It is running on mini cube on my laptop, so it might be a little bit slow anyway. So this is the approach with the git repo and if we would like to use the direct go upload, we can just swi. We can just change the first task in this in this pipeline, where we can implement some, let's say basic receiver application that will wait for open connections and then we can from developer machine.

A

We can open a connection to this to this part and basically copy copy, a diff of source code or something like that and then execute the build again and deployment again. Okay, so let's see if my function has been built, so as you can see, my function has been built I'll, just quickly check.

A

Sorry typing my hand.

A

So, as you can see, as you can see, the function is up and running. We can. We can also check the pipeline, uh the tecton pipeline stuff, so we can uh describe yeah and we can check the latest by planner and, as you can see, this is the techno resources that has been generated for me. So me as a developer, don't need to know, but, as you can see here, I can check that the status of all all steps so the first. This is the fetch resources, then building and deploying.

A

Okay, so that's all thank you for your attendance and.