Getting Started with Helm
In this exercise, you will take your first steps with Helm. If you have not installed it yet, please install Helm on your machine. Once installed, you should be able to run the helm command.
$ helm version
version.BuildInfo{Version:"v3.11.3", GitCommit:"...", GitTreeState:"clean", GoVersion:"go1.20.3"}
Install Wordpress
Helm uses the Artifact Hub as the built-in chart repository. The website is searchable, search for wordpress and read the basic installation instructions.
To install Wordpress, run the following command.
$ helm install wp1 oci://registry-1.docker.io/bitnamicharts/wordpress
Pulled: registry-1.docker.io/bitnamicharts/wordpress:16.0.4
Digest: sha256:e76dcaf9f200d21db6afb9708a40ae3cb5dff3174d63b4059919b3a71182a1e1
NAME: wp1
...
After a successful execution, several instructions are printer to the terminal. The most important instruction tells you how to find the admin password.
echo Password: $(kubectl get secret --namespace default wp1-wordpress \
-o jsonpath="{.data.wordpress-password}" | base64 -d)
This will reveal the admin password for your wp1 instance or wordpress.
To check details and the status of your release, you can use the list command.
$ helm list -A
NAME NAMESPACE REVISION UPDATED STATUS CHART APP VERSION
wp1 default 1 2023-05-07 16:22:48 deployed wordpress-16.0.4 6.2.0
You can also check the status of your cluster to see what has been deployed.
$ $ kubectl get pods
NAME READY STATUS RESTARTS AGE
wp1-mariadb-0 1/1 Running 0 4m14s
wp1-wordpress-cfc756bb4-2dhhm 1/1 Running 0 4m14s
$ kubectl get services
NAME TYPE CLUSTER-IP EXTERNAL-IP PORT(S) AGE
wp1-mariadb ClusterIP 10.107.95.158 <none> 3306/TCP 4m19s
wp1-wordpress LoadBalancer 10.100.26.173 <pending> 80:30364/TCP,443:32016/TCP 4m19s
So, apparently, the chart has installed a database and a webserver for the wordpress installation. To access the service you could manually create an Ingress, but we will take a shortcut for this exercise and use minikube to automatically create such a tunnel for us. Under the hood, the command will create a temporary Ingress for all configured Services.
$ minikube service list
|----------------------|------------------------------------|--------------|-----|
| NAMESPACE | NAME | TARGET PORT | URL |
|----------------------|------------------------------------|--------------|-----|
...
| default | wp1-wordpress | http/80 | |
| | | https/443 | |
...
|----------------------|------------------------------------|--------------|-----|
$ minikube tunnel
Wordpress is configured for ports 80 and 443, as such, the tunnel command requires admin priviledges. As long as the tunnel is up, you can then access localhost or localhost:443 to access Wordpress. The admin backend is reachable via localhost/wp-admin, use the username user and the password that you have discovered before.
For this exercise, it is not relevant to understand how to use Wordpress. The important learning is that Helm can be used to install complex deployments in your cluster with a single command. It is even possible to run multiple releases of the same chart.
$ helm install wp2 oci://registry-1.docker.io/bitnamicharts/wordpress
...
$ helm list
NAME NAMESPACE REVISION UPDATED STATUS CHART APP VERSION
wp1 default 1 2023-05-07 16:22:48 deployed wordpress-16.0.4 6.2.0
wp2 default 1 2023-05-07 16:49:16 deployed wordpress-16.0.4 6.2.0
$ kubectl get pods
NAME READY STATUS RESTARTS AGE
wp1-mariadb-0 1/1 Running 0 23m
wp1-wordpress-cfc756bb4-2dhhm 1/1 Running 0 23m
wp2-mariadb-0 0/1 Running 0 5s
wp2-wordpress-7679687978-w8z5w 0/1 Running 0 5s
To finish this part of the exercise, uninstall your releases to free-up the resources again.
$ helm uninstall wp1 wp2
release "wp1" uninstalled
release "wp2" uninstalled
$ helm list
NAME NAMESPACE REVISION UPDATED STATUS CHART APP VERSION
Create Your First Chart
Similar to Docker images, it is not required to share charts in a repository, they can as well be defined and used locally. To practice this, create a folder somewhere on your machine. Within that folder, create the default template for a Helm chart.
$ helm create my_chart
Creating my_chart
Inspecting the folder will reveal the folder structure that has been discussed in the lecture. For now, delete all content and reduce the folder to the following three files (create the nginx.yaml file).
my_chart/
├── Chart.yaml
├── templates
│ └── nginx.yaml
└── values.yaml
For the exercise, you can leave the Chart.yaml file as is, but add the following contents to values.yaml.
# values.yml
nginx:
version: latest
port: 80
Also, add the following contents to nginx.yaml. This is the same definition of a simple pod that we used in the Kubernetes lecture, but some parts are taken from variables.
# templates/nginx.yml
apiVersion: v1
kind: Pod
metadata:
name: nginx
spec:
containers:
- name: nginx
image: nginx:{{.Values.nginx.version}}
ports:
{{- with .Values.nginx }}
- containerPort: {{.port}}
{{- end}}
The example illustrates two ways to access variables. Either, by simply using the full path in the values.yaml (e.g., {{.Values.nginx.version}}), or by binding the scope first (e.g., {{- with .Values.nginx }}) and then only refering to a relative path (e.g., {{.port}}). You can find more information about values in the Helm documentation on Values.
You can install this chart like every chart with the install command. For the exercise, use the --dry-run and --debug flag to see the generated files, without directly applying them to your cluster. You will see from the output that the variables have been replaced.
$ helm install myrelease --dry-run --debug ./my_chart/
...
# Source: my_chart/templates/nginx.yml
apiVersion: v1
kind: Pod
metadata:
name: nginx
spec:
containers:
- name: nginx
image: nginx:latest
ports:
- containerPort: 80
Default values can be overridden on install by using the --set parameter. Run the following command to use a different version of nginx.
$ helm install myrelease --dry-run --debug ./my_chart/ --set nginx.version=1.2.3
...
image: nginx:1.2.3
...