Cloud Native Computing Foundation Kubernetes Community Days Bangalore 2022, 11 May 2022

Previous Meeting Next Meeting

⏯

youtube image

►

From YouTube: Running local Kubernetes clusters using minikube, kind and microk8s

Description

In this talk, I will be explaining how one can run local Kubernetes clusters using various tools such as minikube, kind, and microk8s.

A

Hello, everyone welcome to kcd bangalore, I am miha and we will be discussing about running local kubernetes cluster using some of the tools. So a little about me, I am a junior pursuing bachelor's in computer engineering, currently working as an intern at cloud yoga 3 release. I am a doc shadow and kubernetes and also involved as a sick, docs member for this spring cohort. I am a lfx mentee at cncf, kubernetes majorly working on sig network. Also, you could connect with me on linkedin and twitter.

A

So let's look at what we will discuss. First of all, I will give you a little background on why we need a local kubernetes cluster and how they are useful.

A

Then we will move on to some of the tools that might help us to set up a local environment and finally, we will see what are the major differences so why local communities cluster are useful. It creates agile and safe application deployment process. Although most people run kubernetes in a cloud environment, running kubernetes locally has its own benefits. We can quickly try out kubernetes before deciding to use it as a primary platform to deploy our application. Users can focus on experimentation and proof of concepts in a very low risk.

A

Environment, like local environments, eliminate the need to provide any of the required cloud infrastructure and any other related administrative task.

A

It provides an environment isolation and we get quick feedback from the users. Thus, it helps to better understand how the system works. Ultimately, it develops in-depth practitioner skills when working with the platform.

A

So we have some tools to help us running a local kubernetes cluster. Each of these tools simplify setup and configuration to varying degrees. We will be looking uh about mini cube kind and microgrids in a while. So, let's start with mini cube, the kubernetes sick is the key developer. It all started in that is a year after the first kubernetes release.

A

It supports cross platform. That is, it is installed. It can be installed in linux, mac, os and windows.

A

It follows kubernetes table releases mini cube, handles some of the features uh like load, balancer, node port ingress and also supports all of the container runtimes. It also uses multi-nodes and a driver. Apart from the base, configuration and minicube also provides a set of add-ons.

A

We know that uh every tool has its own pros and cons. So some of the advantages of mini cube are it supports multi drivers that can match with any environment, and it can work with or without any intermediate vm on linux, that is, we can use a vm driver as equally as equal to none flag.

A

It installs several plugins by default as well, and it is very flexible on installation requirements and upgrades, and some of the disadvantages of minicube are that the installation and removal are not as streamlined as other solutions, and it can also create conflict with the local installation of the other tools like virtualbox, you can do a hands-on demo of mini cube tool by clicking on this link.

A

As soon as you come on the cloud you can portal here, you will see a lab set up button by clicking on the lab setup button. You will get a terminal and an id where you can perform certain commands and get your mini cube cluster.

A

Now, let's move on with the next tool that is kind kubernetes and docker. This tool is also developed by kubernetes. It started in 2019 kind, is a tool for running local kubernetes cluster using docker container nodes. So it's being said that it works anywhere where docker works kind was primarily designed for testing itself, but it can be used for local development, quality assurance or cr. It uses container nodes with systemd or cri and kubernetes as well. It supports multi-node cluster on the same machine in kind.

A

Creating and deleting a cluster is much more faster as compared to other tools. One of the major advantage is it uses container d instead of docker shim. Now question may arise like how we can say that it can be considered as an advantage.

A

It's because kubernetes is proceeding with the deprecation and removal of the docker ship in the upcoming 1.24 release. Next, it runs and most ci environments. Network external access to the cluster is more complicated and kind, so it can be considered as one of the core unkind.

A

You can try out a hands-on demo by clicking on this link, so before installing kind we will need to install docker and cube ctl on our lab. So let's update first.

A

Now we can turn the command to install the docker.

A

As soon as the docker installation is cutting complete, we can start installing cube ctl.

A

Now, as docker and cubesat installation is getting completed, we can start the kind installation, as there are many ways to install the kind you can refer the official documentation.

A

Given the link above there are mainly three ways: one can install kind like install with package manager installed from release binaries or directly from source code.

A

Here we are installing it from the release. Binaries, you can execute this command.

A

After getting kind installed, we can confirm the installation by running the kind version command you can see. You can execute this commands over here also.

A

Now, as we are completed with the installation of kind, we can begin by creating the kubernetes cluster.

A

Run kind create cluster.

A

Now here you can see the starting control plane, so the above command, which we have run kind create cluster, will create a single node cluster with the default name kind.

A

So now we can confirm the corresponding docker container by running the docker ps command.

A

Now, if you don't want a cluster with the default name, you can also use hyphen iphone name flag to give the custom name. Now we can, after setting up a single node cluster, we can also interact with it to list out all the cluster. We can use kind get cluster comma.

A

Here we are getting kind as we have set it up.

A

You can also use cube. Ctrl commands as well.

A

Now, once the cluster is ready, we can check the details using the cluster info command for cube ctl.

A

You can see kubernetes control plane is running at this ip address. If you don't need cluster at some point, you can also delete it.

A

Like now, if someone wants to provide some custom specifications to the clusters, then we might need to work with the config files. So, let's begin with the basic config file, the left has been configured with the desired config file such as config.yaml and many others. You can check it by running the following command.

A

We can create a cluster using this config file.

A

We can also define the nodes in the config file like. If I want to create a multi-node cluster with one control, plane and two child nodes, then I will run this command.

A

And it will create a multi-node cluster for me, as I have already a multi-node.ml file.

A

So, yes, our cluster is set up now. What if one wants to change the version of the image like we can have two different versions of the clusters running on the same machine? We can run this command.

A

To fetch the image of this v1.21.1.

A

Version you can also consider checking the kinds release note if you want to use a different image.

A

So, at a high level, we use feature gate to turn and features on and off technically feature gates are a set of key value pairs that describes uh kubernetes features. You can turn this features on or off cluster wide, using the hyphen for feature gates flag on each kubernetes component.

A

Same way we have runtime config, which is one of the advanced functionality like it is a set of key value pairs that enable or disable built-in kubernetes apis. We can write a group. Slash version is equal to true or false. As per your requirement for a specific api group and version, we can have a look at some of the advanced concepts of kubernetes like setting up engraves or creating a local registry through kind.

A

So what is ingress ingress is an object that manages external access to the services running on the cluster, to set up an increased controller. There are mainly two steps. First of all, we will create a cluster by running the following command.

A

Next, we need to deploy an ingress controller. There are many implementations of ingress controller, like ambassador contour, glue, ingress engine x and many more, but here we will use an ingress engine x, which provides a deployment which we can leverage through greater so run. The calling commands.

A

Finally, we are all set to deploy our services for the demo purpose. We are using a simple http eco web server, which is available as a docker image on the docker hub.

A

You can see over here.

A

So, yes, you can consider checking the service.tml file, which contains foo and bar services.

A

We can deploy our service by running the following command. So now, as we have deployed our services, we can check the status of all the services using cube. Ctrl get services.

A

We can verify the application running under the lab urls.

A

We need to write, slash foo, so here you can see foo as an output. Similarly, we need to write bar to see the bar service running.

A

Now we may have a look on local registry. Our goal is to keep everything locally, including our local docker registry. So a registry is a storage and content delivery system holding named docker images available in different tagged versions. We can check a cell script file in the terminal, which creates a kubernetes cluster with local docker registry. Enabled the above cell script creates a docker registry called the kind registry which runs locally on the port 5000.

A

It first inspects the current environment for any existing registry and set up a new registry. If it couldn't find any.

A

This configuration file will run a single known, kubernetes cluster and add some configuration to the containerity interface to allow pulling images from a local docker registry.

A

Once the docker registry is available, the shell script configures kind to use this local registry for pulling container images during deployments.

A

The last step is to connect the client's network with the local docker registries network, run the following command to execute the process of cluster creation using the local registry. As we can see, our local registry gets created. Also, we need to make sure that the docker registry is running by executing the following command.

A

Now we will try to use this registry. We can pull some images from the docker hub.

A

By using the docker pull command, we have completed the pulling of the image. We can also take the image so that it points to our registry.

A

We can push by using the docker push command and finally, we can use this image.

A

So, as you can see, we have created our deployment using the local registry.

A

We can verify our deployment as well. Finally, in this hands-on lab, we have learned enough about kind like starting from the installation process to deploying services and setting up a local registry which are some of the advanced concepts.

A

Let's see something about micro gates, canonical is the key developer of microgate. It is one of the kubernetes distribution, which is mainly aimed for developers and iot devices. Microgreat is a snap deploying kubernetes.

A

It is better tested in ubuntu than in any other distributions. Flannel is a tool which gives a subnet to each host, for use with container runtimes micro creates supports the functionality of multi-node clusters on different machines. There are some situations where you may need additional services.

A

Microcade scatters with the concept of add-ons that is extra services, is provided, which can be easily added to microgreats, so in micro creates. Everything is easy, like installation, if we need to upgrade, and also the removal process, is too easy.

A

As you can see in the architecture of microgreats, it is completely isolated from the other tools in your machine and also it does not need a vm. All services runs locally.

A

You can try out hands-on lab on micro kx by clicking on this link. So here you will be learning about the installation process of micro gates how add-ons work a brief introduction about most widely used add-ons. And, finally, you will host your first service in kubernetes, using micro gates.

A

Let's have a quick look at how they are different from each other, so many cube and kind can manage nodes, while micro kits don't manage many manicure uses, virtualbox docker, podman, ssh, node management system. According to the operating system, kind uses docker as a node management system, mini cube and micro gates have the functionality of add-ons while kind don't have the default. Cni in mini cube is bridge in kind. It is kind net d and micro creates its planet.

A

If we talk about multi-node support, then all three supports multi-node management, but in different ways like in mini cube, it is through the driver in kind it can be done on the same physical machine and in micro k8. It may be same or different container run time in mini cube, can be docker container d or creo in kind. It can be container d or creo and in micro kits it should be continuity.

A