Skip to main content

Control Prep

As you can imagine, k8s can get outrageously complex. To deploy any application on Kubernetes you may need a half dozen to a few hundred resource definitions, scripts, templates and all sorts of complication. This guide is not meant to be a deep dive into Kubernetes—heck, we've been working with it for years and at times only feel we have scratched the surface—so we will be consolidating our interactions with k3s as much as possible.

As such, to really make the most out of k3s, as novices, we need to have some tools installed first that will make our lives easier. Most importantly, we will be using Helm—the Kubernetes package manager—to provide organization, separations of concerns, and simplicity to Kubernetes. But, there are a handful of other tools that are must-haves.

We will now install all these tools on our local control computer, and get them set up appropriately.

Install Kubectl​

Kubectl is how you control and interact with a Kubernetes cluster. You may recall from How? that Kubernetes adds centralized command and control to a distributed system. Kubectl ("Kube Control") is that centralized control. Whoever has the cert which gives them full access to the cluster, fully owns that cluster.

Make sure you have kubectl installed on your local/control machine (it also gets installed on the leader node, but it's best to have locally, so you don't have to ssh into the cluster) Install.

Once installed, it's a great idea to enable bash/zsh completion though not required.

You can make sure Kubectl is working by running:

% kubectl
kubectl controls the Kubernetes cluster manager.

 Find more information at: https://kubernetes.io/docs/reference/kubectl/overview/

Basic Commands (Beginner):
  create        Create a resource from a file or from stdin
...

Install Krew​

Krew is a plugin manager for Kubectl (yes, both Kubernetes and Kubectl have their ecosystem managers, did we mention that the Kubernetes landscape has ten of billions of funding behind it?). Krew makes life with Kubectl, and thus Kubernetes, much better.

  1. Install Krew from here and make sure it's working by running:

    % kubectl krew
    krew is the kubectl plugin manager.
    You can invoke krew through kubectl: "kubectl krew [command]..."
    ...

    Once installed, poke around, you might find a plugin you find interesting.

  2. We highly recommend adding ns which allows you to change namespace via the command:

    kubectl ns bitcoin

    This will deeply come in handy as without it, nearly every command will need a namespace flag, which is cumbersome.

    To install:

    kubectl krew install ns

  3. Lastly, make sure you are good to go by running

    kubectl krew update

Install Kube-ps1 (optional)​

Kube-ps1 puts your current namespace in the prompt of you terminal. When combined with ns installed above, it is the best way to use k8s (that we have seen, so far). Take it from us, being in the wrong namespace and/or context has killed production websites in real time (and we know this, because we have mistakenly done it).

Install from here

You can edit it accordingly, but your command prompt should now look something like this:

(⎈|default:default) %

The first default is the context. If you had multiple clusters, like a test cluster and production cluster, that is where context comes in handy. We won't be paying much attention to this at this time. The second default is the namespace, which we will be utilizing, and which will affect where apps get installed.

Install Helm​

Helm consolidates app deployments from a myriad of resources, definitions, variables, and redundancies into single line install, update, and uninstall scripts. We will be using it to install and manage as much as we can. Install. Ansible will primarily use Helm at first, but we will use it directly once we get to Bitcoin.

% helm
The Kubernetes package manager

Common actions for Helm:

- helm search:    search for charts
...

What are namespaces?​

For a full definition, check out this article. For our novice purposes, we can think of namespaces as folders which isolate apps. Apps in k8s are generally not solitary programs, but a collection of containers, storage, pods, tasks and other resources, that all work together in replaceable ways to deliver a functional App with minimized (or completely removed) single points of failure. It is helpful to isolate these resources into namespaces so that the cluster does not become overly cluttered with definitions and thus hard (or impossible) to navigate. Namespaces exist cluster wide, they are not assigned to hosts.

All set? Let's spin up Kubernetes!