Skip to content

Latest commit

 

History

History
129 lines (98 loc) · 4.61 KB

README.md

File metadata and controls

129 lines (98 loc) · 4.61 KB
Raw

Installation

  • Setup Cloud Resources

    If using AWS, create VPC and AMI. Similarly for other providers create the necessary resources.

  • Setup Kubernetes cluster in the cloud

    At least one node in the cluster must have the "worker" role. Verify by executing the following command.

     kubectl get nodes

    You should see "worker" under the "ROLES" column as shown below:

     NAME             STATUS   ROLES                         AGE   VERSION
 testk-master-0   Ready    control-plane,master,worker   37h   v1.25.0

    If "worker" role is missing, execute the following command to set the role.

    export NODENAME=<node-name>
kubectl label node $NODENAME node.kubernetes.io/worker=

Deploy webhook

Please refer to the instructions available in the following doc.

Configure and deploy CoCo and the cloud-api-adaptor

Deploy the CoCo operator, CC runtime CRD and the cloud-api-adaptor daemonset

Using make deploy

You can deploy the CoCo operator and cloud-api-adaptor with the Makefile by running

  • set CLOUD_PROVIDER

    export CLOUD_PROVIDER=<aws|azure|ibmcloud|ibmcloud-powervs|libvirt|vsphere>
    • RESOURCE_CTRL is set to true by default to allow the peerpod-ctrl to run, monitor and delete dangling cloud resources

  • make deploy deploys operator, runtime and cloud-api-adaptor pod in the configured cluster

    • validate kubectl is available in your $PATH and $KUBECONFIG is set

Note: make delete deletes the cloud-api-adaptor daemonset from the configured cluster (and peerpod-ctrl if RESOURCE_CTRL=true is set)

Manually

Alternatively the manual approach, if you want to pick a specific CoCo release/reference is:

  • Deploy the CoCo operator

  • Create the peer pods variant of the CC custom resource to install the required pieces of CC and create the kata-remote RuntimeClass

    • Again, either deploy a release version of the Confidential Containers peer pod customer resource with, by running the following command where <RELEASE_VERSION> needs to be substituted with the desired release tag:

    Note: the release version needs to be v0.6.0 or after

    export RELEASE_VERSION=<RELEASE_VERSION>
kubectl apply -k github.com/confidential-containers/operator/config/samples/ccruntime/peer-pods?ref=<RELEASE_VERSION>
    • Alternatively install the latest development version with:
    kubectl apply -k "github.com/confidential-containers/operator/config/samples/ccruntime/peer-pods"

  • Wait until all the pods are running with:

    kubectl get pods -n confidential-containers-system --watch

  • Wait until the kata-remote runtime class has been created by running:

    kubectl get runtimeclass --watch

  • Apply the kustomize.yaml configuration that you modified earlier with:

    kubectl apply -k install/overlays/ibmcloud

  • Wait until all the pods are running with:

    kubectl get pods -n confidential-containers-system --watch

Verify

  • Check POD status

    kubectl get pods -n confidential-containers-system

    A successful install should show all the PODs with "Running" status under the confidential-containers-system namespace.

    NAME                                              READY   STATUS    RESTARTS   AGE
cc-operator-controller-manager-546574cf87-phbdv   2/2     Running   0          43m
cc-operator-daemon-install-pzc4b                  1/1     Running   0          42m
cc-operator-pre-install-daemon-sgld6              1/1     Running   0          42m
cloud-api-adaptor-daemonset-mk8ln                 1/1     Running   0          37s

  • View cloud-api-adaptor logs

    kubectl logs pod/cloud-api-adaptor-daemonset-mk8ln -n confidential-containers-system

Building custom cloud-api-adaptor image

  • Set CLOUD_PROVIDER

    export CLOUD_PROVIDER=<aws|azure|ibmcloud|ibmcloud-powervs|libvirt|vsphere>

  • Set container registry and image name

    export registry=<namespace>/<image_name>

  • Build the container image and push it to $registry

    make image