In this step you will add a chart registry to your Helm installation.

Task

Let's look first at what registries are available from our Helm client:

helm repo list

Helm installs the official k8s chart registry called stable and also a local chart registry called local.

Now let's add another repository to our Helm client. Copy the following command and paste it into the terminal:

helm repo add talend https://sgandon.github.io/katacoda-scenarios/

Execute again helm repo list and see the new talend registry in the list.

You can now list the charts from your talend registry:

helm search talend

If you want to find out what else helm repo can do, execute the following command:

helm repo --help

Now that you have added talend registry to your Helm client let's fetch the chart hello-world from this registry. This chart packages the same application you have worked with in the Kubernetes workshop. Before you download the chart, let's look at some details of this chart (version, app version, description):

helm search talend/hello-world

Use the following command to fetch the hello-world chart and in the same time unpack it in the current folder:

helm fetch --untar talend/hello-world

The command above has created a hello-world folder inside the current folder.

In this step you will install the hello-world chart in the k8s cluster.

Task

Helm provides two commands which are recommended to be used before installing a chart:

• helm lint checks the formatting of the manifest files
• helm --debug --dry-run simulates an install. The flag "--debug" displays the output of the install in the console.

Let's execute these commands and check their output:

helm lint hello-world

helm install hello-world --debug --dry-run

The chart is now ready to be installed. Every time a chart is installed, Helm will assign it a new release name unless you explicitly overwrite it. You can overwrite the default name with the "--name" flag. Let's use helm-training as our release name:

helm install hello-world --name helm-training

List all Helm/k8s applications/releases with the following command:

helm ls

Please note the columns REVISION, CHART and APP VERSION.

You can also see the resources deployed by a release with the following command:

helm status helm-training

Check the status of pods to see when the application has been full deployed. If you don't remember the command take a look at the solution below.

When the hello-world pod is running find out the service port exposed by k8s (the port will be in the range 30000-32767):

kubectl get --namespace default -o jsonpath='{.spec.ports[0].nodePort}{"\n"}' services helm-training-hello-world

In the upper side of the terminal (title bar) you have a tab called "Display 8080". Click on it and a new tab will open in your browser. The URL will have the following format: https://177078276-8080-599693266-training1.environments.katacoda.com/ but the application will not work. Replace 8080 in the host name with the port displayed by the previous command. The hello-world welcome page will appear which confirms that the application is up and running in the k8s cluster.

Let's move to the next step and see how we can perform an upgrade and a rollback on the Helm release.

In this step you will learn how to upgrade and rollback a Helm release.

Task

You will start by checking the number of running pods:

kubectl get pods

At this stage only one pod is running.

Open the file hello-world/values.yaml and change the value replicaCount from 1 to 3. Then open the file hello-world/Chart.yaml and change the field version from 0.1.0 to 0.1.1.

You are now ready to upgrade the Helm release that you created in the previous step.

helm upgrade helm-training hello-world

Check again the release information by executing:

helm ls

You can see that the fields REVISION and CHART have been modified. This means that you should have now three pods running. Check that again:

kubectl get pods

To view the full release history, execute the following command:

helm history helm-training

Let's see now what happens if we rollback to revision 1:

helm rollback helm-training 1

Execute again helm history helm-training. The REVISION field is incremented to 3 but the chart version is the same as in revision 1.

You should now also have only one pod running:

kubectl get pods

You have successfuly upgraded and rollbacked a release. Now you can delete the release before we move to the next step:

helm delete --purge helm-training

This command removes the release and all k8s objects belonging to it.

In this step you will learn how to check for chart dependencies and how to update chart dependencies.

Task

Let's fetch the chart talend-infrastructure from the talend registry and untar it. Try to do this on your own before you check the solution below. But don't feel discouraged if you can't. :-)

For the purpose of learning about dependencies let's rename the folder talend-infrastructure/charts to talend-infrastructure/subcharts:

mv talend-infrastructure/charts talend-infrastructure/subcharts

To check if the chart can be installed, simulate an install and display the output in the terminal. Use infra as Helm Release name.

Right, the chart is missing dependencies! You can list the chart's dependencies with the following command:

helm dep ls talend-infrastructure

Take a look at talend-infrastructure/requirements.yaml to see where and how dependencies are defined. All dependencies should have been in the talend-infrastructure/charts folder but, of course, they don't exist because we deleted that folder.

Let's recreate the talend-infrastructure/charts folder by updating the chart's dependencies:

helm dep up talend-infrastructure

And now look again at the talend-infrastructure/charts folder.

You can now proceed to install the chart talend-infrastructure with the release name infra in k8s. Do you remember the command for installing a chart?

Unfortunately that install has failed because, by default, the infrastructure chart doesn't create any objects. You will have to explicitly set which infrastructure service you want to create. This can be done through tags and/or conditions which we are going to tackle down in the next step.

However, even if the installed has failed, Helm has created a release which we need to delete.

Try to list existing releases and delete the infra one.

In this step you will learn about tags and conditions, which are a means to enable and disable chart dependencies.

Task

The talend-infrastructure chart has five dependencies and by default they are all disabled. Look again at talend-infrastructure/requirements.yaml and especially pay attention to the fields condition and tags. Then look at talend-infrastructure/values.yaml to see the default values for the tags. The condition global.infra.enabled is not defined at all.

Default values can be overriden at install time either by specifying a new values file with the -f parameter or by specifying an new individual value with the --set parameter.

Let's enable the talend-redis dependency by using the --set parameter:

helm install talend-infrastructure --name infra --set tags.talend-redis=true

Alternatively, you can use a new values file as follows:

helm install talend-infrastructure --name infra -f values-standalone.yaml

Check the status of the release with helm ls. Redis is now up and running inside the k8s cluster.

Do you remember how to check what pods are available in k8s? What about how to attach to a pod and run a command? Attach to the redis pod and run the command redis-cli ping. You should receive a pong message back.

You can now delete the infra release.

Congratulations, you have finished your Helm workshop!