Skip to main content

Deploy a GCP IaaS Cluster with Crossplane

Palette supports using Crossplane to create and manage Kubernetes host clusters across major infrastructure providers. This section guides you on how to use Crossplane to deploy a Palette-managed Kubernetes cluster in GCP.

Prerequisites

  • A Palette account and API key. Refer to the Create API Key page for instructions on creating an API key.
  • A public GCP service account with the required roles.
  • An SSH key pair available. Check out the Create an Upload an SSH Key page for guidance.
  • The GCP account must be registered in Palette. Follow the Register and Manage GCP Accounts guide to register your account in Palette.
  • A Kubernetes cluster with at least 2 GB of RAM. This guide uses a kind cluster as an example. Refer to the kind Quick Start to learn how to install kind and create a cluster.
  • The following software is required and must be installed:
    • kubectl
    • Helm version v3.2.0 or later
    • curl
    • A text editor such as Vi or Nano. This guide uses Vi as an example.

Deploy a GCP IaaS Cluster with Crossplane

  1. Open up a terminal session and set the kubectl context to your kind cluster. Replace <kind-cluster-name> with the name of your cluster.

    kubectl cluster-info --context <kind-cluster-name>
    Kubernetes control plane is running at https://127.0.0.1:65306
    CoreDNS is running at https://127.0.0.1:65306/api/v1/namespaces/kube-system/services/kube-dns:dns/proxy
    To further debug and diagnose cluster problems, use 'kubectl cluster-info dump'.
  2. Next, add the Crossplane Helm chart. This will enable Crossplane to install all its Kubernetes components.

    helm repo add \
    crossplane-stable https://charts.crossplane.io/stable
    helm repo update
    "crossplane-stable" has been added to your repositories
    Hang tight while we grab the latest from your chart repositories...
    ...Successfully got an update from the "ngrok" chart repository
    ...Successfully got an update from the "crossplane-stable" chart repository
    ...Successfully got an update from the "stable" chart repository
    Update Complete. ⎈Happy Helming!⎈
  3. Install the Crossplane components using the helm install command.

    helm install crossplane \
    crossplane-stable/crossplane \
    --namespace crossplane-system \
    --create-namespace

    You can verify the installation with the kubectl get pods command. The output must contain two Crossplane pods in the Running status.

    kubectl get pods --namespace crossplane-system
    NAME                                     READY   STATUS    RESTARTS   AGE
    crossplane-869d89c8f8-7jc6c 1/1 Running 0 20s
    crossplane-rbac-manager-784b496b-8mr6z 1/1 Running 0 20s
  4. Once Crossplane is installed, create a folder to store the Kubernetes configuration files.

    mkdir crossplane-gcp
  5. Use a text editor of your choice to create a file for the Palette Crossplane provider configuration.

    vi crossplane-gcp/provider-palette.yaml
  6. Paste the following Kubernetes configuration into the text editor window that opens. Press the Escape key, type :wq, and press Enter to save the file and exit.

    apiVersion: pkg.crossplane.io/v1
    kind: Provider
    metadata:
    name: provider-palette
    spec:
    package: xpkg.upbound.io/crossplane-contrib/provider-palette:v0.19.2
    tip

    Access the Upbound Marketplace to get the latest version of the Palette Crossplane provider.

  7. Issue the command below to install the Palette Crossplane provider. Crossplane will install the CRDs that allow you to create Palette resources directly inside Kubernetes.

    kubectl apply --filename crossplane-gcp/provider-palette.yaml

    You can check the installation with the kubectl get providers command.

    kubectl get providers
     NAME                          INSTALLED   HEALTHY   PACKAGE                                              AGE
    provider-palette True True crossplane-contrib/provider-palette:v0.12.0 61s
  8. Create a file to store a Kubernetes Secret containing your Palette API key. The Palette provider requires credentials to create and manage resources.

    vi crossplane-gcp/secret-gcp.yaml
  9. Paste the following Kubernetes configuration into the text editor window that opens. Replace <your-api-key> with your Palette API key and change the values of project_name and host according to your environment. Save the file and exit.

    apiVersion: v1
    kind: Secret
    metadata:
    name: palette-creds
    namespace: crossplane-system
    type: Opaque
    stringData:
    credentials: |
    {
    "api_key": "<your-api-key>",
    "project_name": "Default",
    "host": "console.spectrocloud.com"
    }
  10. Create the Kubernetes secret.

    kubectl apply --filename crossplane-gcp/secret-gcp.yaml
  11. Next, create a file to store the ProviderConfig object. This object configures the Palette Crossplane provider with the secret containing the Palette API key.

    vi crossplane-gcp/providerconfig-gcp.yaml
  12. Paste the content below into the text editor window, save the file and exit.

    apiVersion: palette.crossplane.io/v1beta1
    kind: ProviderConfig
    metadata:
    name: default
    spec:
    credentials:
    source: Secret
    secretRef:
    name: palette-creds
    namespace: crossplane-system
    key: credentials
  13. Create the Kubernetes ProviderConfig object.

    kubectl apply --filename crossplane-gcp/providerconfig-gcp.yaml
  14. Once the Palette Crossplane provider is installed and set up, create a file to store the GCP cluster profile configuration.

    vi crossplane-gcp/cluster-profile-gcp.yaml
  15. Paste the Kubernetes configuration below into the text editor window that opens. Save the file and exit.

    apiVersion: cluster.palette.crossplane.io/v1alpha1
    kind: Profile
    metadata:
    name: gcp-crossplane-cluster-profile
    namespace: crossplane-system
    spec:
    forProvider:
    cloud: "gcp"
    description: "GCP cluster profile"
    type: "cluster"
    pack:
    - name: "ubuntu-gcp"
    tag: "22.04"
    registryUid: "5eecc89d0b150045ae661cef"
    values:
    "# Spectro Golden images includes most of the hardening as per CIS Ubuntu Linux 22.04 LTS Server L1 v1.0.0
    standards\n\n# Uncomment below section to\n# 1. Include custom files to be copied over to the nodes
    and/or\n# 2. Execute list of commands before or after kubeadm init/join is
    executed\n#\n#kubeadmconfig:\n# preKubeadmCommands:\n# - echo \"Executing pre kube admin config
    commands\"\n# - update-ca-certificates\n# - 'systemctl restart containerd; sleep 3'\n# - 'while [ ! -S
    /var/run/containerd/containerd.sock ]; do echo \"Waiting for containerd...\"; sleep 1;
    done'\n# postKubeadmCommands:\n# - echo \"Executing post kube admin config commands\"\n# files:\n# -
    targetPath: /usr/local/share/ca-certificates/mycom.crt\n# targetOwner:
    \"root:root\"\n# targetPermissions: \"0644\"\n# content: |\n# -----BEGIN
    CERTIFICATE-----\n# MIICyzCCAbOgAwIBAgIBADANBgkqhkiG9w0BAQsFADAVMRMwEQYDVQQDEwprdWJl\n# cm5ldGVzMB4XDTIwMDkyMjIzNDMyM1oXDTMwMDkyMDIzNDgyM1owFTETMBEGA1UE\n# AxMKa3ViZXJuZXRlczCCASIwDQYJKoZIhvcNAQEBBQADggEPADCCAQoCggEBAMdA\n# nZYs1el/6f9PgV/aO9mzy7MvqaZoFnqO7Qi4LZfYzixLYmMUzi+h8/RLPFIoYLiz\n# qiDn+P8c9I1uxB6UqGrBt7dkXfjrUZPs0JXEOX9U/6GFXL5C+n3AUlAxNCS5jobN\n# fbLt7DH3WoT6tLcQefTta2K+9S7zJKcIgLmBlPNDijwcQsbenSwDSlSLkGz8v6N2\n# 7SEYNCV542lbYwn42kbcEq2pzzAaCqa5uEPsR9y+uzUiJpv5tDHUdjbFT8tme3vL\n# 9EdCPODkqtMJtCvz0hqd5SxkfeC2L+ypaiHIxbwbWe7GtliROvz9bClIeGY7gFBK\n# jZqpLdbBVjo0NZBTJFUCAwEAAaMmMCQwDgYDVR0PAQH/BAQDAgKkMBIGA1UdEwEB\n# /wQIMAYBAf8CAQAwDQYJKoZIhvcNAQELBQADggEBADIKoE0P+aVJGV9LWGLiOhki\n# HFv/vPPAQ2MPk02rLjWzCaNrXD7aPPgT/1uDMYMHD36u8rYyf4qPtB8S5REWBM/Y\n# g8uhnpa/tGsaqO8LOFj6zsInKrsXSbE6YMY6+A8qvv5lPWpJfrcCVEo2zOj7WGoJ\n# ixi4B3fFNI+wih8/+p4xW+n3fvgqVYHJ3zo8aRLXbXwztp00lXurXUyR8EZxyR+6\n# b+IDLmHPEGsY9KOZ9VLLPcPhx5FR9njFyXvDKmjUMJJgUpRkmsuU1mCFC+OHhj56\n# IkLaSJf6z/p2a3YjTxvHNCqFMLbJ2FvJwYCRzsoT2wm2oulnUAMWPI10vdVM+Nc=\n# -----END
    CERTIFICATE-----"
    uid: "63fdd138199bafb6b657b7e6"

    - name: "kubernetes"
    tag: "1.28.3"
    registryUid: "5eecc89d0b150045ae661cef"
    values:
    "# spectrocloud.com/enabled-presets: Kube Controller Manager:loopback-ctrlmgr,Kube
    Scheduler:loopback-scheduler\npack:\n k8sHardening: True\n #CIDR Range for Pods in cluster\n # Note :
    This must not overlap with any of the host or service network\n podCIDR: \"192.168.0.0/16\"\n #CIDR
    notation IP range from which to assign service cluster IPs\n # Note : This must not overlap with any IP
    ranges assigned to nodes for pods.\n serviceClusterIpRange: \"10.96.0.0/12\"\n # serviceDomain:
    \"cluster.local\"\n\n# KubeAdm customization for kubernetes hardening. Below config will be ignored if
    k8sHardening property above is disabled\nkubeadmconfig:\n apiServer:\n extraArgs:\n # Note :
    secure-port flag is used during kubeadm init. Do not change this flag on a running
    cluster\n secure-port: \"6443\"\n anonymous-auth: \"true\"\n profiling:
    \"false\"\n disable-admission-plugins: \"AlwaysAdmit\"\n default-not-ready-toleration-seconds:
    \"60\"\n default-unreachable-toleration-seconds: \"60\"\n enable-admission-plugins:
    \"AlwaysPullImages,NamespaceLifecycle,ServiceAccount,NodeRestriction,PodSecurity\"\n admission-control-config-file:
    \"/etc/kubernetes/pod-security-standard.yaml\"\n audit-log-path:
    /var/log/apiserver/audit.log\n audit-policy-file:
    /etc/kubernetes/audit-policy.yaml\n audit-log-maxage: \"30\"\n audit-log-maxbackup:
    \"10\"\n audit-log-maxsize: \"100\"\n authorization-mode: RBAC,Node\n tls-cipher-suites:
    \"TLS_ECDHE_ECDSA_WITH_AES_128_GCM_SHA256,TLS_ECDHE_RSA_WITH_AES_128_GCM_SHA256,TLS_ECDHE_ECDSA_WITH_CHACHA20_POLY1305,TLS_ECDHE_RSA_WITH_AES_256_GCM_SHA384,TLS_ECDHE_RSA_WITH_CHACHA20_POLY1305,TLS_ECDHE_ECDSA_WITH_AES_256_GCM_SHA384,TLS_RSA_WITH_AES_256_GCM_SHA384,TLS_RSA_WITH_AES_128_GCM_SHA256\"\n extraVolumes:\n -
    name: audit-log\n hostPath: /var/log/apiserver\n mountPath:
    /var/log/apiserver\n pathType: DirectoryOrCreate\n - name: audit-policy\n hostPath:
    /etc/kubernetes/audit-policy.yaml\n mountPath: /etc/kubernetes/audit-policy.yaml\n readOnly:
    true\n pathType: File\n - name: pod-security-standard\n hostPath:
    /etc/kubernetes/pod-security-standard.yaml\n mountPath:
    /etc/kubernetes/pod-security-standard.yaml\n readOnly: true\n pathType:
    File\n controllerManager:\n extraArgs:\n profiling: \"false\"\n terminated-pod-gc-threshold:
    \"25\"\n use-service-account-credentials: \"true\"\n feature-gates:
    \"RotateKubeletServerCertificate=true\"\n scheduler:\n extraArgs:\n profiling:
    \"false\"\n kubeletExtraArgs:\n read-only-port : \"0\"\n event-qps: \"0\"\n feature-gates:
    \"RotateKubeletServerCertificate=true\"\n protect-kernel-defaults:
    \"true\"\n rotate-server-certificates: \"true\"\n tls-cipher-suites:
    \"TLS_ECDHE_ECDSA_WITH_AES_128_GCM_SHA256,TLS_ECDHE_RSA_WITH_AES_128_GCM_SHA256,TLS_ECDHE_ECDSA_WITH_CHACHA20_POLY1305,TLS_ECDHE_RSA_WITH_AES_256_GCM_SHA384,TLS_ECDHE_RSA_WITH_CHACHA20_POLY1305,TLS_ECDHE_ECDSA_WITH_AES_256_GCM_SHA384,TLS_RSA_WITH_AES_256_GCM_SHA384,TLS_RSA_WITH_AES_128_GCM_SHA256\"\n files:\n -
    path: hardening/audit-policy.yaml\n targetPath: /etc/kubernetes/audit-policy.yaml\n targetOwner:
    \"root:root\"\n targetPermissions: \"0600\"\n - path: hardening/90-kubelet.conf\n targetPath:
    /etc/sysctl.d/90-kubelet.conf\n targetOwner: \"root:root\"\n targetPermissions: \"0600\"\n -
    targetPath: /etc/kubernetes/pod-security-standard.yaml\n targetOwner:
    \"root:root\"\n targetPermissions: \"0600\"\n content: |\n apiVersion:
    apiserver.config.k8s.io/v1\n kind: AdmissionConfiguration\n plugins:\n - name:
    PodSecurity\n configuration:\n apiVersion:
    pod-security.admission.config.k8s.io/v1\n kind:
    PodSecurityConfiguration\n defaults:\n enforce:
    \"baseline\"\n enforce-version: \"v1.28\"\n audit:
    \"baseline\"\n audit-version: \"v1.28\"\n warn:
    \"restricted\"\n warn-version: \"v1.28\"\n audit:
    \"restricted\"\n audit-version: \"v1.28\"\n exemptions:\n # Array of
    authenticated usernames to exempt.\n usernames: []\n # Array of runtime class
    names to exempt.\n runtimeClasses: []\n # Array of namespaces to
    exempt.\n namespaces: [kube-system]\n\n preKubeadmCommands:\n # For enabling
    'protect-kernel-defaults' flag to kubelet, kernel parameters changes are required\n - 'echo \"====>
    Applying kernel parameters for Kubelet\"'\n - 'sysctl -p
    /etc/sysctl.d/90-kubelet.conf'\n #postKubeadmCommands:\n #- 'echo \"List of post kubeadm commands to
    be executed\"'\n\n# Client configuration to add OIDC based authentication flags in
    kubeconfig\n#clientConfig:\n #oidc-issuer-url: \"{{
    .spectro.pack.kubernetes.kubeadmconfig.apiServer.extraArgs.oidc-issuer-url }}\"\n #oidc-client-id: \"{{
    .spectro.pack.kubernetes.kubeadmconfig.apiServer.extraArgs.oidc-client-id }}\"\n #oidc-client-secret:
    1gsranjjmdgahm10j8r6m47ejokm9kafvcbhi3d48jlc3rfpprhv\n #oidc-extra-scope: profile,email"
    uid: "659a47b28b7673a7dff73658"

    - name: "cni-calico"
    tag: "3.27.0"
    uid: "661cc4f20aa79b7543637fa9"
    registryUid: "5eecc89d0b150045ae661cef"
    values:
    "pack:\n content:\n images:\n - image:
    gcr.io/spectro-images-public/packs/calico/3.27.0/cni:v3.27.0\n - image:
    gcr.io/spectro-images-public/packs/calico/3.27.0/node:v3.27.0\n - image:
    gcr.io/spectro-images-public/packs/calico/3.27.0/kube-controllers:v3.27.0\n\nmanifests:\n calico:\n images:\n cni:
    \"\"\n node: \"\"\n kubecontroller: \"\"\n # IPAM type to use. Supported types are
    calico-ipam, host-local\n ipamType: \"calico-ipam\"\n\n calico_ipam:\n assign_ipv4:
    true\n assign_ipv6: false\n\n # Should be one of CALICO_IPV4POOL_IPIP or
    CALICO_IPV4POOL_VXLAN \n encapsulationType: \"CALICO_IPV4POOL_IPIP\"\n\n # Should be one of Always,
    CrossSubnet, Never\n encapsulationMode: \"Always\"\n\n env:\n # Additional env variables for
    calico-node\n calicoNode:\n #IPV6: \"autodetect\"\n #FELIX_IPV6SUPPORT:
    \"true\"\n #CALICO_IPV6POOL_NAT_OUTGOING: \"true\"\n #CALICO_IPV4POOL_CIDR:
    \"192.168.0.0/16\"\n #IP_AUTODETECTION_METHOD: \"first-found\"\n\n # Additional env variables
    for calico-kube-controller deployment\n calicoKubeControllers:\n #LOG_LEVEL:
    \"info\"\n #SYNC_NODE_LABELS: \"true\""

    - name: "csi-gcp-driver"
    tag: "1.12.4"
    uid: "661cc4f80aa79b75d4440676"
    registryUid: "5eecc89d0b150045ae661cef"
    values:
    "pack:\n content:\n images:\n - image:
    gcr.io/spectro-images-public/packs/csi-gcp-driver/1.12.4/csi-provisioner:v3.6.2\n - image:
    gcr.io/spectro-images-public/packs/csi-gcp-driver/1.12.4/csi-attacher:v4.4.2\n - image:
    gcr.io/spectro-images-public/packs/csi-gcp-driver/1.12.4/csi-resizer:v1.9.2\n - image:
    gcr.io/spectro-images-public/packs/csi-gcp-driver/1.12.4/csi-snapshotter:v6.3.2\n - image:
    gcr.io/spectro-images-public/packs/csi-gcp-driver/1.12.4/gcp-compute-persistent-disk-csi-driver:v1.12.4\n -
    image:
    gcr.io/spectro-images-public/packs/csi-gcp-driver/1.12.4/csi-node-driver-registrar:v2.9.2\n \nmanifests:\n storageclass:\n #Flag
    to denote if this should be the default storage class for dynamic provisioning\n isDefaultClass:
    \"true\"\n\n parameters:\n #Possible values : pd-standard or pd-ssd\n type:
    \"pd-standard\"\n \n #Possible values: none or regional-pd\n replication-type:
    \"none\"\n \n #Supported binding modes are Immediate,
    WaitForFirstConsumer\n volumeBindingMode: \"WaitForFirstConsumer\"\n\n #Set this flag to true to
    enable volume expansion\n allowVolumeExpansion: true\n\n #Allowed reclaim policies are Delete,
    Retain\n reclaimPolicy: \"Delete\"\n\n #allowedTopologies\n zones:\n #-
    us-central1-a\n #- us-central1-b\n\n k8sVersion: \"{{ .spectro.system.kubernetes.version
    }}\"\n\n controller:\n args:\n csiProvisioner:\n - \"--v=5\"\n -
    \"--csi-address=/csi/csi.sock\"\n - \"--feature-gates=Topology=true\"\n -
    \"--http-endpoint=:22011\"\n - \"--leader-election-namespace=$(PDCSI_NAMESPACE)\"\n -
    \"--timeout=250s\"\n - \"--extra-create-metadata\"\n #-
    \"--run-controller-service=false\" # disable the controller service of the CSI driver\n #-
    \"--run-node-service=false\" # disable the node service of the CSI driver\n -
    \"--leader-election\"\n - \"--default-fstype=ext4\"\n -
    \"--controller-publish-readonly=true\"\n \n csiAttacher:\n - \"--v=5\"\n -
    \"--csi-address=/csi/csi.sock\"\n - \"--http-endpoint=:22012\"\n -
    \"--leader-election\"\n - \"--leader-election-namespace=$(PDCSI_NAMESPACE)\"\n -
    \"--timeout=250s\"\n\n csiResizer:\n - \"--v=5\"\n -
    \"--csi-address=/csi/csi.sock\"\n - \"--http-endpoint=:22013\"\n -
    \"--leader-election\"\n - \"--leader-election-namespace=$(PDCSI_NAMESPACE)\"\n -
    \"--handle-volume-inuse-error=false\"\n\n csiSnapshotter:\n - \"--v=5\"\n -
    \"--csi-address=/csi/csi.sock\"\n - \"--metrics-address=:22014\"\n -
    \"--leader-election\"\n - \"--leader-election-namespace=$(PDCSI_NAMESPACE)\"\n -
    \"--timeout=300s\"\n\n csiDriver:\n - \"--v=5\"\n - \"--endpoint=unix:/csi/csi.sock\""

    providerConfigRef:
    name: default

    The cluster profile contains the following core infrastructure layers.

    Pack TypeRegistryPack NamePack Version
    OSPublic Repoubuntu-gcp22.04
    KubernetesPublic Repokubernetes1.28.3
    NetworkPublic Repocni-calico3.27.0
    StoragePublic Repocsi-gcp-driver1.12.4
    tip

    If you want to use different packs in your cluster profile, you can access the Palette UI and simulate creating a cluster profile to gather the packs' required values. During the cluster profile creation, click the API button to display the API payload. You will find the values of each pack's name, tag, uid, registryUid, and values parameters.

  16. Create the cluster profile.

    kubectl apply --filename crossplane-gcp/cluster-profile-gcp.yaml
  17. Issue the commands below to get the ID of the cluster profile once its creation is complete.

    kubectl wait --for=condition=Ready profile.cluster.palette.crossplane.io/gcp-crossplane-cluster-profile
    clusterProfileId=$(kubectl get profile.cluster.palette.crossplane.io gcp-crossplane-cluster-profile --output jsonpath='{.status.atProvider.id}')
    echo Cluster Profile ID: $clusterProfileId
    profile.cluster.palette.crossplane.io/gcp-crossplane-cluster-profile condition met
    Cluster Profile ID: 6638e0eb8f42b00cb4d1d22bb
  18. Next, get the ID of your GCP cloud account registered in Palette by invoking the cloudaccounts Palette API. Replace <your-api-key> with your Palette API key and <gcp-account-name> with the name under which you registered your GCP account in Palette.

    curl --location --request GET 'https://api.spectrocloud.com/v1/cloudaccounts/gcp' \
    -H 'Accept: application/json' \
    -H 'ApiKey: <your-api-key>' | jq '.items[] | select(.metadata.name == "<gcp-account-name>") | .metadata.uid'
    "645981f0ab3ab8105fabc982"

    Copy the API response containing your GCP cloud account ID.

  19. Create a file to store the GCP IaaS cluster configuration.

    vi crossplane-gcp/cluster-gcp.yaml
  20. Paste the cluster configuration displayed below into the text editor window that opens:

    • Replace <project-name> with the name of your GCP project.
    • Replace <cluster-profile-id> and <cloud-account-id> with the IDs obtained in steps 17 and 18 of this guide.

    Optionally, you can edit the region, availability zone, instance type, and number of nodes of your cluster according to your workload.

    Once you are done making the alterations, save and exit the file.

    apiVersion: cluster.palette.crossplane.io/v1alpha1
    kind: GCP
    metadata:
    name: gcp-crossplane-cluster
    namespace: crossplane-system
    spec:
    forProvider:
    cloudConfig:
    - project: "<project-name>"
    region: "us-east1"
    machinePool:
    - azs:
    - us-east1-b
    count: 2
    instanceType: n1-standard-4
    name: machinepool1
    - azs:
    - us-east1-b
    count: 1
    instanceType: n1-standard-4
    name: controlplanepool
    controlPlane: true
    clusterProfile:
    - id: <cluster-profile-id>
    cloudAccountId: <cloud-account-id>
    providerConfigRef:
    name: default
  21. Create the GCP IaaS cluster.

    kubectl apply --filename crossplane-gcp/cluster-gcp.yaml
  22. Wait for the cluster to be created. The cluster provisioning might take up to 20 minutes.

    kubectl wait --for=condition=Ready gcp.cluster.palette.crossplane.io/gcp-crossplane-cluster --timeout=1h

    Once ready, you should receive an output similar to the following.

    gcp.cluster.palette.crossplane.io/gcp-crossplane-cluster condition met

Validate

  1. Log in to Palette.

  2. From the left Main Menu, click Clusters.

  3. Verify the deployed cluster named gcp-crossplane-cluster is displayed and has the Running and Healthy status.