CLI

Google Kubernetes Engine (GKE) is a fully managed Kubernetes service provided by Google Cloud. An integration is developed with GKE to ensure that users can provision GKE Clusters in any region and Google Cloud project using CLI (RCTL).

Create Cluster Via RCTL¶

Step 1: Cloud Credentials¶

Use the below command to create a GCP credential via RCTL

./rctl create credential gcp credentials-name <Location of credentials JSON File>

On successful creation, use this credential in the cluster config file to create a GKE cluster

Step 2: Create Cluster¶

Users can create the cluster based on a version controlled cluster spec that you can store in a Git repository. This enables users to develop automation for reproducible infrastructure.

./rctl apply  -f cluster-spec.yml

GKE config (v3 API version and Recommended)¶

Here's an illustrative example of a GKE Cluster v3 YAML file, featuring the following components:

One (1) node pool
Kubernetes control plane version set to 1.26
Cluster location set to zonal in us-central1-a
Maximum pods per node set to 110
Compute Engine Persistent Disk CSI driver enabled
Reservation Affinity configured

apiVersion: infra.k8smgmt.io/v3
kind: Cluster
metadata:
  name: my-gke-cluster
  project: defaultproject
spec:
  cloudCredentials: "demo-cred"
  type: gke
  config:
    gcpProject: demo-gcp-project
    location:
      type: zonal
      config:
        zone: us-central1-a
    controlPlaneVersion: '1.26'
    network:
      name: default
      subnetName: default
      access:
        type: public
        config:
      enableVPCNativetraffic: true
      maxPodsPerNode: 110
    features:
      enableComputeEnginePersistentDiskCSIDriver: true
    nodePools:
    - name: default-nodepool
      nodeVersion: '1.26'
      size: 2
      machineConfig:
        imageType: COS_CONTAINERD
        machineType: e2-standard-4
        bootDiskType: pd-standard
        bootDiskSize: 100
        reservationAffinity:
          consumeReservationType: specific
          reservationName: reservation-name
  blueprint:
    name: minimal
    version: latest

To configure either DataPlaneV2 or Calico Kubernetes Network Policy, include the parameters dataPlaneV2 and set enableDataPlaneV2Observability and enableDataPlaneV2Metrics to true. Alternatively, set the networkPolicyConfig to true and specify the network policy as calico. Below is a demonstration of the configuration for network details

network:
  name: default
  subnetName: default
  access:
    type: public
    config:
  enableVPCNativetraffic: true
  maxPodsPerNode: 110
  # dataPlaneV2: "ADVANCED_DATAPATH"
  # enableDataPlaneV2Observability: true
  # enableDataPlaneV2Metrics: true
  networkPolicyConfig: true
  networkPolicy: "CALICO"

Important

Users are allowed to enable either DataPlane V2 or Calico Kubernetes Network Policy during GKE cluster creation, but not both simultaneously.

For the V3 Config Schema, refer here

GKE config (v2 API version)¶

Here's an illustrative example of a YAML file for a regional GKE cluster, featuring the following components:

two (2) node pools
auto-upgrade enablement
node pool upgrade strategy set to Surge

apiVersion: infra.k8smgmt.io/v2
kind: Cluster
metadata:
    name: gke-cluster
    project: default-project
spec:
    blueprint:
        name: default
        version: latest
    cloudCredentials: gke-cred
    config:
        controlPlaneVersion: "1.22"
        location:
            region:
                region: us-east1
                zone: us-east1-b
            type: regional
        name: gke-cluster
        network:
            enableVPCNativeTraffic: true
            maxPodsPerNode: 75
            name: default
            networkAccess:
                privacy: public
            nodeSubnetName: default
        nodePools:
            - machineConfig:
                bootDiskSize: 100
                bootDiskType: pd-standard
                imageType: COS_CONTAINERD
                machineType: e2-medium
                reservationAffinity:
                  consumeReservationType: specific
                  reservationName: reservation-name
              name: default-nodepool
              nodeMetadata:
                gceInstanceMetadata:
                    - key: org-team
                      value: qe-cloud
                kubernetesLabels:
                    - key: nodepool-type
                      value: default-np
              nodeVersion: "1.22"
              size: 2
              management:
                autoUpgrade: true
              upgradeSettings:
                strategy: SURGE
                surgeSettings:
                  maxSurge: 0
                  maxUnavailable: 1
            - machineConfig:
                bootDiskSize: 60
                bootDiskType: pd-standard
                imageType: COS_CONTAINERD
                machineType: e2-medium
                reservationAffinity:
                  consumeReservationType: specific
                  reservationName: reservation-name
              name: pool2
              nodeMetadata:
                gceInstanceMetadata:
                    - key: org-team
                      value: qe-cloud
                kubernetesLabels:
                    - key: nodepool-type
                      value: nodepool2
              nodeVersion: "1.22"
              size: 2
        project: project1
        security:
            enableLegacyAuthorization: true
            enableWorkloadIdentity: true
    type: Gke

To configure the shared VPC Network, specify the exact network path names as demonstrated in the following configuration for network details, where podSecondaryRangeName and serviceSecondaryRangeName are mandatory

network:
  enableVPCNativeTraffic: true
  maxPodsPerNode: 110
  name: projects/kr-test-200723/global/networks/km-1
  networkAccess:
    privacy: public
  nodeSubnetName: projects/kr-test-200723/regions/us-central1/subnetworks/km-1
  podSecondaryRangeName: pod
  serviceSecondaryRangeName: service

To configure a private cluster with Firewall Rules, specify the cluster privacy as demonstrated in the following configuration

network:
  enableVPCNativeTraffic: true
  maxPodsPerNode: 110
  name: default
  networkAccess:
    privacy: private
    privateCluster:
      controlPlaneIPRange: 172.16.8.0/28
      disableSNAT: true
      enableAccessControlPlaneExternalIP: true
      enableAccessControlPlaneGlobal: true
      firewallRules:
      - action: ALLOW
        rules:
        - protocol: tcp
          ports:
          - "22284"
          - "9447"
        description: Allow Webhook Connections
        direction: EGRESS
        name: test-allow-webhook
        # network: default /
        priority: 1555
        destinationRanges:
        - 172.16.8.0/28
      - action: DENY
        rules:
        - protocol: tcp
          ports:
          - "22281"
          - "9443"
        description: Allow Webhook Connections
        direction: INGRESS
        name: deny-demo-test-allow-webhook-1
        priority: 1220
        sourceRanges:
        - 172.16.8.0/28

DataPlane V2 and Network Policy Config

To configure either DataPlaneV2 or Calico Kubernetes Network Policy, include the parameters dataPlaneV2 and set enableDataPlaneV2Observability and enableDataPlaneV2Metrics to true. Alternatively, set the networkPolicyConfig to true and specify the network policy as calico. Below is a demonstration of the configuration for network details

network:
      dataPlaneV2: "ADVANCED_DATAPATH"
      enableDataPlaneV2Observability: true
      enableDataPlaneV2Metrics: true
      # networkPolicyConfig: true
      # networkPolicy: "CALICO"
      enableVPCNativeTraffic: true
      maxPodsPerNode: 110
      disableDefaultSNAT: true
      name: default
      networkAccess:
        privacy: public
      nodeSubnetName: default

Important

To configure Calico network policy for nodes, the networkPolicyConfig must be set to true.

The parameters enableDataPlaneV2Observability and enableDataPlaneV2Metrics can be set to false as a Day-2 operation.
Day-2 operation is not supported for dataPlaneV2, but the parameters enableDataPlaneV2Observability and enableDataPlaneV2Metrics can be modified as a Day-2 operation.

On successful provisioning, you can view the cluster details as shown below

Successfully provisioned GKE cluster

For more GKE cluster spec examples, refer here

For the V2 Config Schema, refer here

For cluster sharing, add a new block to the cluster config (Rafay Spec) as highlighted in the below config file

apiVersion: infra.k8smgmt.io/v2
kind: Cluster
metadata:
  labels:
    clusterName: demo-gke-cluster
    clusterType: gke
  name: demo-gke-cluster
  project: defaultproject
spec:
  blueprint:
    name: minimal
    version: latest
  cloudCredentials: demo-cred
  config:
    controlPlaneVersion: "1.24"
    location:
      type: zonal
      zone: us-west1-c
    name: demo-gke-cluster
    network:
      enableVPCNativeTraffic: true
      maxPodsPerNode: 110
      name: default
      networkAccess:
        privacy: public
      nodeSubnetName: default
    nodePools:
    - machineConfig:
        bootDiskSize: 100
        bootDiskType: pd-standard
        imageType: COS_CONTAINERD
        machineType: e2-standard-4
      name: default-nodepool
      nodeMetadata:
        nodeTaints:
        - effect: NoSchedule
          key: k1
      nodeVersion: "1.24"
      size: 3
    - machineConfig:
        bootDiskSize: 100
        bootDiskType: pd-standard
        imageType: COS_CONTAINERD
        machineType: e2-standard-4
      name: pool2
      nodeVersion: "1.24"
      size: 3
    project: dev-382813
  sharing:
    enabled: true
    projects:
    - name: "demoproject1"
    - name: "demoproject2"
  type: Gke

You can also use the wildcard operator "*" to share the cluster across projects

sharing:
    enabled: true
    projects:
    - name: "*"

Notes: When passing the wildcard operator, users cannot pass other projects name

To remove any cluster sharing from the project(s), remove that specific project name(s) and run the apply command

List Clusters¶

To retrieve a specific GKE cluster, use the below command

./rctl get cluster <gkecluster_name>

Output

./rctl get cluster demo-gkecluster
+------------------------+-----------+-----------+---------------------------+
| NAME                   | TYPE      | OWNERSHIP | PROVISION STATUS          |
+------------------------+-----------+-----------+---------------------------+
| demo-gkecluster        | gke       | self      | INFRA_CREATION_INPROGRESS |
+------------------------+-----------+-----------+---------------------------+

To retrieve a specific v3 cluster details, use the below command

./rctl get cluster demo-gkecluster --v3

Example

./rctl get cluster demo-gkecluster --v3
+------------------------+-------------------------------+-----------+----------+-----------+---------------------------+
| NAME                   | CREATED AT                    | OWNERSHIP | TYPE     | BLUEPRINT | PROVISION STATUS          |
+------------------------+-------------------------------+-----------+----------+-----------+---------------------------+
| demo-gkecluster        | 2023-06-05 10:54:08 +0000 UTC | self      | gke      | minimal   | INFRA_CREATION_INPROGRESS |
+------------------------+-------------------------------+-----------+----------+-----------+---------------------------+

To view the entire v3 cluster config spec, use the below command

./rctl get cluster <gkecluster_name> --v3 -o json

(or)

./rctl get cluster <gkecluster_name> --v3 -o yaml

Download Cluster Config¶

Use the below command to download the v3 Cluster Config file

./rctl get cluster config <cluster-name> --v3

Important

Download the cluster configuration only after the cluster is completely provisioned

Node Pool Management¶

To add/edit/scale/upgrade/delete node pool(s), make the required changes in the GKE Cluster config spec and run the apply command

Delete Cluster¶

Delete cluster will clean up the resources in Google Cloud as well

./rctl delete cluster <cluster_name>

Dry Run¶

The dry run command is utilized for operations such as upgrades, control plane upgrades, blueprint upgrades, and node pool operations. It provides a pre-execution preview of changes, enabling users to assess potential modifications before implementation. This proactive approach is beneficial for identifying and addressing issues, ensuring that the intended changes align seamlessly with infrastructure requirements. Whether provisioning a new cluster or managing updates, incorporating a dry run enhances the predictability, reliability, and overall stability of your infrastructure.

./rctl apply -f <clusterfile.yaml> --dry-run

Examples

Controlplane Upgrade

Below is an example of the output from the dry run command when a user successfully attempts to upgrade a control plane.

./rctl apply -f demo_cluster.yaml --dry-run
{
  "operations": [
    {
      "operation": "ControlPlane upgrade",
      "resourceName": "ControlPlane version: 1.26",
      "comment": "ControlPlane will be upgraded. Diff shown in json patch"
    }
  ],
  "diffJsonPatch": {
    "patch": [
      {
        "op": "replace",
        "path": "/spec/config/controlPlaneVersion",
        "value": {
          "Kind": {
            "string_value": "1.26"
          }
        }
      }
    ]
  }
}

New update request when cluster provision in progress

Below is an example of the output from the dry run command when a user tries to send an update request to a cluster while the cluster provision is IN PROGRESS:

./rctl apply -f demo_cluster.yaml --dry-run
{
  "operations": [
    {
      "operation": "Cluster Update",
      "resourceName": "gke-np-upgradesetting",
      "comment": "Cluster object exists. But provision is IN-PROGRESS.",
      "error": ["Cannot run dry-run on the cluster now. Try again later.."]
    }
  ],
  "diffJsonPatch": {
    "patch": [
      {
        "op": "replace",
        "path": "/spec/config/controlPlaneVersion",
        "value": {
          "Kind": {
            "string_value": "1.27"
          }
        }
      }
    ]
  }
}

CLI

Create Cluster Via RCTL¶

Step 1: Cloud Credentials¶

Step 2: Create Cluster¶

GKE config (v3 API version and Recommended)¶

GKE config (v2 API version)¶

Cluster Sharing¶

List Clusters¶

Download Cluster Config¶

Node Pool Management¶

Delete Cluster¶

Dry Run¶