work

README

ManifestWork

Support a primitive that enables resources to be applied to a managed cluster.

Community, discussion, contribution, and support

Check the CONTRIBUTING Doc for how to contribute to the repo.

---

Description

Work agent

Work agent is a controller running on the managed cluster. It watches the ManifestWork CRs in a certian namespace on hub cluster and applies the manifests included in those CRs on the managed clusters.

Work webhook

Work webhook is an addmission webhook running on hub cluster. It ensures the content of ManifestWorks created/updated are valid.

Getting Started

Prerequisites

These instructions assume:

• You have at least one running kubernetes cluster;
• The name of the managed cluster is cluster1;

Deploy

Deploy on one single cluster

Set environment variables.

export KUBECONFIG=</path/to/kubeconfig>

Override the docker image (optional)

export IMAGE_NAME=<your_own_image_name> # export IMAGE_NAME=quay.io/open-cluster-management/work:latest

And then deploy work webhook and work agent

make deploy

Deploy on two clusters

Set environment variables.

• Hub and managed cluster share a kubeconfig file

  export KUBECONFIG=</path/to/kubeconfig>
  export HUB_KUBECONFIG_CONTEXT=<hub-context-name>
  export SPOKE_KUBECONFIG_CONTEXT=<spoke-context-name>
  

• Hub and managed cluster use different kubeconfig files.

  export HUB_KUBECONFIG=</path/to/hub_cluster/kubeconfig>
  export SPOKE_KUBECONFIG=</path/to/managed_cluster/kubeconfig>
  

Set cluster ip if you are deploying on KIND clusters.

export CLUSTER_IP=<host_name/ip_address>:<port> # export CLUSTER_IP=hub-control-plane:6443

You can get the above information with command below.

kubectl --kubeconfig </path/to/hub_cluster/kubeconfig> -n kube-public get configmap cluster-info -o yaml

Override the docker image (optional)

export IMAGE_NAME=<your_own_image_name> # export IMAGE_NAME=quay.io/open-cluster-management/work:latest

And then deploy work webhook and work agent

make deploy

What is next

After a successful deployment, a namespace cluster1 is created on the hub cluster. You are able to create ManifestWorks in this namespace, and the workload will be distributed to the managed cluster and then be applied by the work agent.

Create a manifest-work.yaml as shown in this example:

apiVersion: work.open-cluster-management.io/v1
kind: ManifestWork
metadata:
  name: hello-work
  namespace: cluster1
  labels:
    app: hello
spec:
  workload:
    manifests:
      - apiVersion: apps/v1
        kind: Deployment
        metadata:
          name: hello
          namespace: default
        spec:
          selector:
            matchLabels:
              app: hello
          template:
            metadata:
              labels:
                app: hello
            spec:
              containers:
                - name: hello
                  image: quay.io/asmacdo/busybox
                  command: ['sh', '-c', 'echo "Hello, Kubernetes!" && sleep 3600']

Apply the yaml file to the hub cluster.

kubectl --kubeconfig </path/to/hub_cluster/kubeconfig> apply -f manifest-work.yaml

Verify that the ManifestWork resource was applied to the hub.

kubectl --kubeconfig </path/to/hub_cluster/kubeconfig> -n cluster1 get manifestwork/hello-work -o yaml

apiVersion: work.open-cluster-management.io/v1
kind: ManifestWork
metadata:
  labels:
    app: hello
  name: hello-work
  namespace: cluster1
spec:
  ... ...
status:
  conditions:
    - lastTransitionTime: '2021-06-15T02:26:02Z'
      message: Apply manifest work complete
      reason: AppliedManifestWorkComplete
      status: 'True'
      type: Applied
    - lastTransitionTime: '2021-06-15T02:26:02Z'
      message: All resources are available
      reason: ResourcesAvailable
      status: 'True'
      type: Available
  resourceStatus:
    manifests:
      - conditions:
          - lastTransitionTime: '2021-06-15T02:26:02Z'
            message: Apply manifest complete
            reason: AppliedManifestComplete
            status: 'True'
            type: Applied
          - lastTransitionTime: '2021-06-15T02:26:02Z'
            message: Resource is available
            reason: ResourceAvailable
            status: 'True'
            type: Available
        resourceMeta:
          group: apps
          kind: Deployment
          name: hello
          namespace: default
          ordinal: 0
          resource: deployments
          version: v1

As shown above, the status of the ManifestWork includes the conditions for both the whole ManifestWork and each of the manifest it contains. And there are two condition types:

• Applied. If true, it indicates the whole ManifestWork (or a particular manifest) has been applied on the managed cluster; otherwise reason/message of the condition will show more information for troubleshooting.
• Available. If true, it indicates the coressponding Kubernetes resources of the the whole ManifestWork (or a particular manifest) are available on the managed cluster; otherwise reason/message of the condition will show more information for troubleshooting

Check on the managed cluster and see the Pod has been deployed from the hub cluster.

kubectl --kubeconfig </path/to/managed_cluster/kubeconfig> -n default get pod
NAME                     READY   STATUS    RESTARTS   AGE
hello-64dd6fd586-rfrkf   1/1     Running   0          108s

Clean up

To clean the environment

make undeploy