- Create a basic Helm chart for the
sentences
application
This exercise will create a simple Helm chart for the sentence application. The chart will be 'simple' in the sense that it will not provide support for customizing application parameters.
More information
In the sentences-app/deploy/kubernetes/
folder
we have Kubernetes YAML definitions for the three
microservices that make up the sentence
application (three Deployments and three
Services):
$ ls -1 sentences-app/deploy/kubernetes/
sentences-age-deployment.yaml
sentences-age-svc.yaml
sentences-deployment.yaml
sentences-name-deployment.yaml
sentences-name-svc.yaml
sentences-svc.yaml
- Running the sentences application on Kubernetes
- Create a skeleton Helm chart
- Copy
sentences
kubernetes yaml into the chart - Lint and deploy our new chart
More details
Deploy the sentences application to Kubernetes
First, let's run the application in bare Kubernetes to see that our YAML is right.
kubectl apply -f sentences-app/deploy/kubernetes
This will create three microservice deployments with a single POD instance each.
Test the deployed application
kubectl get pods
💡 The front-end microservice for the sentences application is exposed with a Kubernetes service of type
NodePort
.
When all three PODs are in a running state, look up the actual NodePort used by the frontend microservice:
kubectl get svc sentence
Output:
NAME TYPE CLUSTER-IP EXTERNAL-IP PORT(S) AGE
sentences NodePort 10.15.245.208 <none> 8080:30250/TCP 37s
In the example above, the relevant NodePort is
30250
.
-
look up an external accessible IP address that can be used to access the front-end microservice.
-
kubectl get nodes -o wide
Any of the IP addresses from the
EXTERNAL-IP
-column can be used.
To request a sentence from the sentences
application, use curl with the external IP address
and NodePort
found above:
curl <EXTERNAL-IP>:<NodePort>
Output:
John is 73 years
💡 in the above example
NodePort
should be changed with your nodeport found above
- Clean up the application deployed with
kubectl delete -f sentences-app/deploy/kubernetes/
Create a skeleton Helm chart
First we create a new directory for our Helm chart, and then use the helm create
command to create the chart skeleton:
mkdir helm-chart
cd helm-chart
helm create sentence-app
The helm create
command we just issued created a lot of files that you might want to use when creating a new Helm chart.
We do not need all of those files for the chart we will be creating, therefore we will remove the files we do not need:
rm -rf sentence-app/templates/*
echo "" > sentence-app/values.yaml
This provides us with skeleton chart without any template files.
Copy sentences
kubernetes yaml into the
chart
Next, we copy the original Kubernetes YAML files to the template folder:
cp -v ../sentences-app/deploy/kubernetes/*.yaml sentence-app/templates/
That's it - now we have a Helm chart for our sentences application.
💡 It is a simple Helm chart in the sense that it has no configurable values, but it is a complete installable chart and it will use the correct sentence application Kubernetes YAML definitions.
Lint and deploy our new chart
Before deploying the chart, we run a static validation of it:
helm lint sentence-app/
Running this command produces the following output:
==> Linting sentence-app/
[INFO] Chart.yaml: icon is recommended
1 chart(s) linted, 0 chart(s) failed
💡 Normally a chart is fetched from a chart registry (like a container registry), however, a chart stored locally can also be deployed with Helm.
To deploy the chart from the newly created chart run the following:
helm install sentences sentence-app/
Running this command produces the following output:
NAME: sentences
LAST DEPLOYED: Wed Apr 21 10:43:55 2021
NAMESPACE: user1
STATUS: deployed
REVISION: 1
TEST SUITE: None
To see all the different objects that Helm has created, use:
kubectl get pods,services,deployments
Expected output:
NAME READY STATUS RESTARTS AGE
pod/sentence-age-78fc854dd5-w9gdq 1/1 Running 0 64s
pod/sentence-name-ff4c584b9-txp5n 1/1 Running 0 64s
pod/sentences-746cc46db8-khp85 1/1 Running 0 64s
NAME TYPE CLUSTER-IP EXTERNAL-IP PORT(S) AGE
service/age ClusterIP 10.191.240.60 <none> 8080/TCP 66s
service/name ClusterIP 10.191.251.238 <none> 8080/TCP 66s
service/sentence NodePort 10.191.245.72 <none> 8080:32665/TCP 66s
NAME READY UP-TO-DATE AVAILABLE AGE
deployment.apps/sentence-age 1/1 1 1 66s
deployment.apps/sentence-name 1/1 1 1 66s
deployment.apps/sentences 1/1 1 1 66s
To see the applications installed with Helm use
the helm ls
operation:
helm ls
expected output:
NAME NAMESPACE REVISION UPDATED STATUS CHART APP VERSION
sentences user1 1 2021-04-21 10:43:55.789048706 +0000 UTC deployed sentence-app-0.1.0 1.16.0
To see the Kubernetes YAML which Helm used to
install the application use the helm get
operation:
helm get all sentences
In our case this will be identical to the YAML files we copied previously since we haven't provided any means of customizing the application installation.
Try to reach it again like we did with the raw kubernetes objects application to begin with.
- Note down the NodePort from the service
kubectl get svc
Look up an external accessible IP address that can be used to access the front-end microservice.
kubectl get nodes -o wide
Any of the IP addresses from the EXTERNAL-IP
-column can be used.
curl <EXTERNAL-IP>:<NodePort>
and see that your application is running once again.
Output:
John is 47 years
In this exercise we created a single Helm chart for the complete application even though it is based on three microservices. When would it make sense to have a Helm chart for each microservice?
Uninstall the application release with Helm:
helm uninstall sentences