Google Kubernetes Engine (GKE) Reference
Kip Landergren
(Updated: )
My cheat sheet for Google Kubernetes Engine covering common cluster operations, explanation of configuration options, and documentation links.
Contents
Creating a Cluster
Notes below are based on creating a cluster through the Cloud Console.
Kubernetes Versions: Standard and Advanced
“Advanced” has an improved SLA and reliability. More info in this blog post.
Regional vs Zonal
From the official documentation:
“A regional cluster provides a single static endpoint for the entire cluster and spreads your cluster's Pods across multiple zones of a given region. This allows you to access the cluster's control plane even during an outage or downtime involving one or more (but not all) individual zones.”
As of August 2019 you cannot switch a Zonal cluster to Regional. If you choose Zonal, the zone is permanent.
Master Version
Choose the Kubernetes version appropriate to your application and any dependencies it may have.
Node Pools
Select “More Options” to better customize.
Network Configuration
HTTP Load Balancing
If disabled, default Ingress objects will not be created using Google Cloud Load Balancer. However, a network load balancer will still be provisioned via a forwarding-rule for any IP address to target-pool mapping. This will incur an hourly charge.
Connecting to Cluster
Copy the command from the Cloud Console button “Connect” to update your local kubeconfig file with appropriate credentials and endpoint information. Generally of form:
gcloud container clusters get-credentials CLUSTER-NAME
Pulling Images From gcr.io
For a newly created cluster, things should “just work” as the service-account should be provisioned with permissions that allow read-only access to dev storage. If not, review the service account permissions and access policy on the bucket storing the images.