We need a cluster in order to continue with this workshop. There are two alternatives here. We recommend the first alternative. With the second you must skip the tasks to create build triggers and apply changes to the application code.
- Set up your own cluster on Google Cloud Platform. In order to do this, you need to register your credit card, but you will not be charged anything for this workshop. To do this, follow the steps in ./gcp-setup.md
- The second alternative is to use a service account to authenticate against a cluster we have already created. To do this, follow the steps in ./1b-service-account-setup.md
- To operate our cluster, we will use the Kubernetes command line tool, kubectl:
gcloud components install kubectl
The cloud SDK installs the tool for you. This tool is not Google Cloud specific, but is used to operate Kubernetes clusters regardless of where they are hosted.
- You can see if your cluster is created by this command:
gcloud container clusters list
If the status of your cluster is RUNNING, you are good to go. If there is no output, you might have the wrong project set in your config file. Do this to set the correct project:
-
Go back to your browser and click on the dropdown next to
Google Cloud Platform. This should open a modal where at least one project is listed. -
Copy the ID of the active project
-
Type this in your terminal:
gcloud config project set INSERT_PROJECT_ID -
Run:
gcloud container clusters list
-
We want to set the default zone of our application, this tells google cloud where to look for the cluster. We created our cluster in europe-north1-a and will set our default zone to this.
gcloud config set compute/zone europe-north1-a -
The next step is to make sure that the Kubernetes command line tool is authenticated against our new cluster. This is easily done by this neat gcloud command:
gcloud container clusters get-credentials cv-cluster
What this does is to write credentials to the file ~/.kube/config. You can take a look at that file too see what is written to it.
Extra task: If you want bash autocompletion for kubectl, follow these steps.
Now that we are authenticated, we can look at the components in our cluster by using the kubectl command.
-
Remember how Kubernetes consists of nodes? List them by this command:
kubectl get nodes -
If you want you can get more details about them by describing one of them:
kubectl describe nodes <INSERT_NODE_NAME> -
We also have different namespaces:
kubectl get namespace
This should list the namespaces kube-system, kube-public and default. The namespace defaultis where we will deploy our applications. kube-system is used by Kubernetes, kube-public is for resources that does not need authentication and default is, as the name says, the default namespace for resources. You can create your own namespaces, e.g. for test and prod.
Only if you did step 2. in the first setup task, and did not set up your own Google Cloud project and cluster
The output from the previous command listed more than the default 3 namespaces. This is because we have created one namespace for each service acccount in this workshop. We have given you read/write rights to your own namespace and read rights to the others. This we have done using RBAC (role-based access control) which is a built in feature in Kubernetes. This is not covered in this workshop, but feel free to ask us about it or explore it on your own.
- Change namespace to the namespace provided in the email you got:
kubectl config set-context $(kubectl config current-context) --namespace=<insert-namespace-name-here>
# Validate it
kubectl config view | grep namespace:
The second line should output your namespace.