GKE

To create a Google Kubernetes Engine you have to do so in a Google Cloud Project. Assuming you have already do so and that you have installed the gcloud CLI you can proceed by creating the kubernetes cluster with following characteristics (that you may change):

We recommend disable auto upgrade

  • Project: my-project
  • Cluster name: stackgres
  • GKE version: 1.21.3-gke.2001
  • Zone: us-west1
  • Node locations: us-west1-a,us-west1-b,us-west1-c
  • Machine type: e2-standard-4
  • Number of nodes: 3
  • Disk size 100GB
  • Node auto upgrade/repair disabled
gcloud container clusters create "stackgres" \
 --project "my-project" \
 --region "us-west1" \
 --no-enable-basic-auth \
 --cluster-version "1.21.3-gke.2001" \
 --no-enable-ip-alias \
 --release-channel "None" \
 --machine-type "e2-standard-4" \
 --image-type "COS_CONTAINERD" \
 --disk-type "pd-standard" \
 --disk-size "100" \
 --metadata disable-legacy-endpoints=true \
 --num-nodes "3" \
 --no-enable-autoupgrade \
 --no-enable-autorepair \
 --node-locations "us-west1-a","us-west1-b","us-west1-c"

The output will be similar to the following:

Note: Your Pod address range (`--cluster-ipv4-cidr`) can accommodate at most 1008 node(s).
Creating cluster stackgres in us-west1... Cluster is being health-checked (master is healthy)...done.                                                                                                              
Created [https://container.googleapis.com/v1beta1/projects/my-project/zones/us-west1/clusters/stackgres].
To inspect the contents of your cluster, go to: https://console.cloud.google.com/kubernetes/workload_/gcloud/us-west1/stackgres?project=my-project
kubeconfig entry generated for stackgres.
NAME       LOCATION            MASTER_VERSION    MASTER_IP      MACHINE_TYPE   NODE_VERSION      NUM_NODES  STATUS
stackgres  southamerica-east1  1.20.10-gke.1600  34.95.157.116  e2-standard-4  1.20.10-gke.1600  9          RUNNING

...

To access your cluster via kubectl, you can conveniently configure your kubeconfig via the following command:

gcloud container clusters get-credentials --region us-west1 --project my-project stackgres

# if your gcloud / GKE version expects to use the GKE gcloud auth plugin, you might need to set this env variable:
export USE_GKE_GCLOUD_AUTH_PLUGIN=True; gcloud container clusters get-credentials --region us-west1 --project my-project stackgres

You may also want to cleanup compute disks used by persistence volumes that may have been created:

$ kubectl get nodes
NAME                                       STATUS   ROLES    AGE   VERSION
gke-stackgres-default-pool-2b3329f9-1v10   Ready    <none>   91s   v1.25.6-gke.200
gke-stackgres-default-pool-2b3329f9-4lgg   Ready    <none>   91s   v1.25.6-gke.200
gke-stackgres-default-pool-2b3329f9-8z0h   Ready    <none>   90s   v1.25.6-gke.200
gke-stackgres-default-pool-6efa0dd8-243j   Ready    <none>   91s   v1.25.6-gke.200
gke-stackgres-default-pool-6efa0dd8-mmnn   Ready    <none>   91s   v1.25.6-gke.200
gke-stackgres-default-pool-6efa0dd8-qc56   Ready    <none>   90s   v1.25.6-gke.200
gke-stackgres-default-pool-e04d99f3-79cr   Ready    <none>   92s   v1.25.6-gke.200
gke-stackgres-default-pool-e04d99f3-d4f7   Ready    <none>   91s   v1.25.6-gke.200
gke-stackgres-default-pool-e04d99f3-g1gr   Ready    <none>   91s   v1.25.6-gke.200

To clean up the Kubernetes cluster you can run the following command:

gcloud container clusters delete stackgres \
 --project my-project \
 --region us-west1 \
 --quiet

You may also want to clean up compute disks used by persistence volumes that may have been created:

gcloud compute disks list --project my-project --filter "zone:us-west1" --quiet | tail -n+2 | sed 's/ \+/|/g' | cut -d '|' -f 1-2 \
  | grep '^gke-stackgres-[0-9a-f]\{4\}-pvc-[0-9a-f]\{8\}-[0-9a-f]\{4\}-[0-9a-f]\{4\}-[0-9a-f]\{4\}-[0-9a-f]\{12\}|' \
  | xargs -r -n 1 -I % sh -ec "gcloud -q compute disks delete --project my-project --zone \"\$(echo '%' | cut -d '|' -f 2)\" \"\$(echo '%' | cut -d '|' -f 1)\""