Kubernetes : Google
Overview of Installing on GKE
Google Kubernetes Engine, or GKE for short, is Google Cloud’s managed Kubernetes service. Setting up GKE is somewhat complex due to all the configuration options that Google offers. However, they at least provide sane defaults for most of the settings. You can create a new cluster with as little as a name and a region.
Part 1 : Installing Prerequisites
- Install the
kubectl
andhelm
tools using the instructions here. - Install the
gcloud
CLI tool from Google using the instructions here. - Verify that
gcloud
is installed using the command:gcloud --version
- If not already installed, install the
gke-gcloud-auth-plugin
using the command:gcloud components install gke-gcloud-auth-plugin
- Verify that the
gke-gcloud-auth-plugin
is installed using the command:gke-cloud-auth-plugin --version
- Add the OpenSquiggly Helm chart repository to your list of Helm repositories:
helm repo add opensquiggly https://opensquiggly.github.io/helm-charts
Part 2 : Creating the GKE Cluster
Steps to Complete
- Visit your Google cloud console at console.cloud.google.com and login to your Google account.
- Select Kubernetes Engine from the Quick Access list, or search for “Kubernetes” in the search box at the top and select it from the list.
- Navigate to the “Clusters” section and click the “Create” button to start the process to create a new cluster.
- Choose between an Auto Pilot cluster or Standard cluster.
- The default option is Auto Pilot
- To switch to a Standard cluster, click on the “Switch to Standard Cluster” in the upper-right portion of the header bar at the top.
- Choose Auto Pilot cluster if you want the system to manage and auto-scale your node pools and other resources. The benefit of Auto Pilot is it can save costs by not over-allocating unneeded resources.
- Choose Standard cluster if you want precise control over your cluster resources, and want to manage your node pools and compute resources yourself. The benefit of a Standard cluster is increased control, at the risk of possibly overpaying for over-allocated or unused resources.
- Under the “Cluster basics” section, enter the desired name of the cluster and select the region.
- To set advanced options, navigate to the Networking, Advanced settings, and Fleet sections and select the desired options.
- When you’ve finished reviewing all of your options, navigate to the “Review and create” section and click the “Create Cluster” button.
Part 3 : Connecting kubectl to the GKE Cluster
Steps to Complete
- Retrieve the cluster connection credentials, which will add the connection information to your default
KUBECONFIG file, with the command:
gcloud container clusters get-credentials cluster-name-here --region=region-here --project=project-id-here
- Verify the the cluster has been added to your list of contexts and is your current context using the command:
kubectl config get-contexts
Part 4 : Install OpenSquiggly with the Helm Chart
Use the helm
tool to install the OpenSquiggly Helm chart into the cluster:
helm install release-name-here opensquiggly/allinone --set cloudType=google[,diskSize=xx]
where xx
is the desired disk size in gigabytes. The default is 32Gb if omitted.
Example:
helm install opensquiggly1 opensquiggly/allinone --set cloudType=google,50
The above command will create a Helm release named opensquiggly1
with a 50Gb persistent volume.
Part 5 : Testing the Install
Ensure that the pod has started by issuing the command:
kubectl get pods [--watch]
Note that is may take a minute or two for the persistent volume claim to be provisioned, and the pod won’t be able to install until it is completed. Be patient. We’ve noticed in our testing the PVCs take longer to provision on Auto Pilot than on the Standard cluster.
When the pod is running, retrieve the load balancer IP address with:
kubectl get svc
Hit the IP address with your browser and that should take you to the OpenSquiggly login screen.