posts tools docker kubernetes kubectl aws-cli kind helm environment-setup 2023-11-17

Quick setup your DevOps environment on Fedora 39

Last week i did a SSD upgrade on my openSUSE Tumbleweed box and decided to not clone the old disk, doing a fresh install instead. This one i had KDE instead of GNOME, but decided to put the new Fedora 39 on it. The KDE spin.

Sooner i realized out the tools i had to install back in order to get it ready for daily work.

Installation process itself i'll lend no comments because it's dead simple, the Fedora media installation is also a live cd, so you can try before touch your system.

Hostname

One thing i recommend if to proper set the hostname before install chrome, if you're a chrome user:

# replace 'thanatos' with a cool name for your machine
sudo hostnamectl hostname thanatos

Do this in order to not get hit by this bug.

Install developer tools just in case

Fedora has the concept of package groups, a set of packages which are commonly installed together therefore there is an alias to do so more quickly:

sudo dnf groupinstall "Development Tools"

Check dnf grouplist for other cool package groups.

AWS cli

Fedora offers several cloud tools directly from its repositories.

Just install aws command line client:

sudo dnf install awscli2

This fedora package also delivers proper cli completion out of the box.

Docker

Follow the official docker guide and don't forget the post-installation steps.

# installation
sudo dnf -y install dnf-plugins-core
sudo dnf config-manager --add-repo https://download.docker.com/linux/fedora/docker-ce.repo
sudo dnf install docker-ce docker-ce-cli containerd.io docker-buildx-plugin docker-compose-plugin

# proper configuration
sudo systemctl enable docker
sudo systemctl start docker
sudo groupadd docker
sudo usermod -aG docker $USER
newgrp docker

# testing installation
docker run hello-world

Kubectl

kubectl is the tool needed to interact with kubernetes clusters. Everything runs on kubernetes clusters nowadays so you have to have it ready for tests, emergency deploys or any other cryptic need.

Official tool

One way to get this is to follow the official guide:

# configure yum repo for installation and updates
# This overwrites any existing configuration in /etc/yum.repos.d/kubernetes.repo
cat <<EOF | sudo tee /etc/yum.repos.d/kubernetes.repo
[kubernetes]
name=Kubernetes
baseurl=https://pkgs.k8s.io/core:/stable:/v1.28/rpm/
enabled=1
gpgcheck=1
gpgkey=https://pkgs.k8s.io/core:/stable:/v1.28/rpm/repodata/repomd.xml.key
EOF

# install the cli
sudo dnf install kubectl

From fedora repo

Another way is to simply:

sudo dnf install kubernetes-client

Bash completion

Remember to configure completion for kubectl:

# configure autocomplete
kubectl completion bash | sudo tee /etc/bash_completion.d/kubectl > /dev/null
sudo chmod a+r /etc/bash_completion.d/kubectl

Helm

Helm is a kind of package manager for kubernetes. it makes easier to install popular apps. It's available directly from official fedora 39 repos as first-class citizen:

# install the cli
sudo dnf install helm

It's up to you however to proper configure helm autocomplete:

# configure autocomplete
helm completion bash | sudo tee /etc/bash_completion.d/helm > /dev/null
sudo chmod a+r /etc/bash_completion.d/helm

Kind

Kind is the easiest and cleaner way to install kubernetes into a development machine.

Real deal kubernetes can be a resource hog, minikube and others relies on virtual machines. Kind is designed to be light. But most important, i like it more than the other tools, :-)

Install kind using the "source" installation, but before do that you need to...

Install go

Kind (and several other tools from the DevOps ecosystem) is written in Go and it's dead easy to install go on Fedora 39:

sudo dnf install golang

Use go to install kind

Then install kind:

go install sigs.k8s.io/kind@latest

Configure your $PATH

This installation isn't system-wide like we did for the other tools. That one is just for you.

Therefore add those two lines on your .bashrc file:

export PATH="$PATH:$HOME/go/bin"
source <(kind completion bash)

Check ~/.kube/config

In order to interact with the cluster using kubectl, you will need a valid ~/.kube/config.

Kind generates one for you when you create a cluster.

Use k0s instead of kind

Regarding the local cluster, other cool option is k0s.

But keep in mind, pick only one provider for local cluster. You need a clean house in order to proper mess around and find out, ;-)

Installing

Directly from quickstart guide:

curl -sSLf https://get.k0s.sh | sudo sh
sudo k0s install controller --single
sudo k0s start

One cool thing about k0s is it already comes with metrics-server:

[sombriks@lucien ~]$ sudo k0s kubectl get deployments --all-namespaces -o wide
NAMESPACE     NAME             READY   UP-TO-DATE   AVAILABLE   AGE   CONTAINERS       IMAGES                                                 SELECTOR
kube-system   coredns          0/1     1            0           20h   coredns          quay.io/k0sproject/coredns:1.11.1                      k8s-app=kube-dns
kube-system   metrics-server   0/1     1            0           20h   metrics-server   registry.k8s.io/metrics-server/metrics-server:v0.6.4   k8s-app=metrics-server

Bash completion and ~/.kube/config

And as you can imagine, you can install completion for that tool too:

source <(k0s completion bash)

But unlike kind, it does not perform automatic setup local user to interact with the cluster. Do this:

touch ~/.kube/config
sudo k0s kubeconfig admin > ~/.kube/config
chmod 600 ~/.kube/config

There! We're good to go.

Use k3s instead of kind or k0s

Regarding the local cluster, another cool option is k3s.

curl -sfL https://get.k3s.io | sh -

And as you can imagine, you can install completion for that tool too:

source <(k3s completion bash)

The cool thing about k3s is it comes with metrics and a ingress controller out of the box, so you don't have to install one. You can see what else comes bundled with k3s here.

~/.kube/config

Once installed, you can set up your local user to interact with the cluster by using this command:

touch ~/.kube/config
sudo cat /etc/rancher/k3s/k3s.yaml > ~/.kube/config
chmod 600 ~/.kube/config

Combine kubectl configs with yaml-merge

Dealing with several clusters, local and remote, can be cumbersome.

Each tool has a way to provide ~/.kube/config for you, but they don't talk to each other out of the box.

One simple solution is to generate several configs and merge them using yaml-merge.

Quite easy to install:

sudo npm -g install yaml-merge

Quite easy to use:

# let's take an eks cluster
aws eks update-kubeconfig --name my-cluster
# and our local k0s cluster
sudo k0s kubeconfig admin > ~/.kube/config-k0s
# then merge those
mv ~/.kube/config ~/.kube/config-eks
yaml-merge ~/.kube/config-k0s ~/.kube/config-eks > ~/.kube/config

Of course, there are several ways to get tha merge done, but the tool is useful!

Install VSCode and/or Intellij Ultimate

Both IDE's offer a decent kubernetes plugins which comes to be very handy when inspecting cluster.

You, like me, will need an IDE for coding anyways.

Install k9s

k9s is a must-have tool to use alongside kubectl itself. It helps to better visualize what's happening and to perform quick actions on the cluster.

The simpler way to install is using go:

go install github.com/derailed/k9s@latest

In order to proper work all you have to do is to setup your ~/.kube/config.

Wrap up

There! next time i need to gear up for DevOps, this is the minimum i need. The minimum you need.

Happy Hacking!