Multicluster Hello World Helidon

The Hello World Helidon example is a Helidon-based service that returns a “Hello World” response when invoked. The example application is specified using Open Application Model (OAM) component and application configuration YAML files, and then deployed by applying those files. This example shows how to deploy the Hello World Helidon application in a multicluster environment.

Before you begin

Create a multicluster Verrazzano installation with one admin and one managed cluster, and register the managed cluster, by following the instructions here.

NOTE: The Hello World Helidon application deployment files are contained in the Verrazzano project located at <VERRAZZANO_HOME>/examples/multicluster/hello-helidon, where <VERRAZZANO_HOME> is the root of the Verrazzano project.

Create the application namespace

Apply the VerrazzanoProject resource on the admin cluster that defines the namespace for the application. The namespaces defined in the VerrazzanoProject resource will be created on the admin cluster and all the managed clusters.

$ kubectl --kubeconfig $KUBECONFIG_ADMIN apply \
    -f https://raw.githubusercontent.com/verrazzano/verrazzano/v1.0.4/examples/multicluster/hello-helidon/verrazzano-project.yaml

Deploy the Hello World Helidon application

1. Apply the hello-helidon multicluster resources to deploy the application. Each multicluster resource is an envelope that contains an OAM resource and a list of clusters to which to deploy.

   $ kubectl --kubeconfig $KUBECONFIG_ADMIN apply \
       -f https://raw.githubusercontent.com/verrazzano/verrazzano/v1.0.4/examples/multicluster/hello-helidon/mc-hello-helidon-comp.yaml
   $ kubectl --kubeconfig $KUBECONFIG_ADMIN apply \
       -f https://raw.githubusercontent.com/verrazzano/verrazzano/v1.0.4/examples/multicluster/hello-helidon/mc-hello-helidon-app.yaml
   

2. Wait for the application to be ready on the managed cluster.

   $ kubectl --kubeconfig $KUBECONFIG_MANAGED1 wait \
       --for=condition=Ready pods \
       --all -n hello-helidon \
       --timeout=300s
   

Explore the example application

Follow the instructions for exploring the Hello World Helidon application in a single cluster use case. Use the managed cluster kubeconfig for testing the example application.

Troubleshooting

Follow the instructions for troubleshooting the Hello World Helidon application in a single cluster use case. Use the managed cluster kubeconfig for troubleshooting the example application.

1. Verify that the application namespace exists on the managed cluster.

   $ kubectl --kubeconfig $KUBECONFIG_MANAGED1 get namespace hello-helidon
   

2. Verify that the multicluster resources for the application all exist.

   $ kubectl --kubeconfig $KUBECONFIG_MANAGED1 get MultiClusterComponent -n hello-helidon
   $ kubectl --kubeconfig $KUBECONFIG_MANAGED1 get MultiClusterApplicationConfiguration -n hello-helidon
   

Locating the application on a different cluster

By default, the application is located on the managed cluster called managed1. You can change the application’s location to be on a different cluster, which can be the admin cluster or a different managed cluster. In this example, you change the placement of the application to the admin cluster by patching the multicluster resources.

1. To change the application’s location to the admin cluster, specify the change placement patch file.

   # To change the placement to the admin cluster
   $ export CHANGE_PLACEMENT_PATCH_FILE="https://raw.githubusercontent.com/verrazzano/verrazzano/v1.0.4/examples/multicluster/hello-helidon/patch-change-placement-to-admin.yaml"
   

   This environment variable is used in subsequent steps.

2. To change their placement, patch the hello-helidon multicluster resources.

   $ kubectl --kubeconfig $KUBECONFIG_ADMIN patch mcappconf hello-helidon-appconf \
       -n hello-helidon \
       --type merge \
       --patch "$(cat $CHANGE_PLACEMENT_PATCH_FILE)"
   $ kubectl --kubeconfig $KUBECONFIG_ADMIN patch mccomp hello-helidon-component \
       -n hello-helidon \
       --type merge \
       --patch "$(cat $CHANGE_PLACEMENT_PATCH_FILE)"
   

3. To verify that their placement has changed, view the multicluster resources.

   $ kubectl --kubeconfig $KUBECONFIG_ADMIN get mccomp hello-helidon-component \
       -n hello-helidon \
       -o jsonpath='{.spec.placement}';echo
   $ kubectl --kubeconfig $KUBECONFIG_ADMIN get mcappconf hello-helidon-appconf \
       -n hello-helidon \
       -o jsonpath='{.spec.placement}';echo
   

   The cluster name, local, indicates placement in the admin cluster.

4. To change its placement, patch the VerrazzanoProject.

   $ kubectl --kubeconfig $KUBECONFIG_ADMIN patch vp hello-helidon \
       -n verrazzano-mc \
       --type merge \
       --patch "$(cat $CHANGE_PLACEMENT_PATCH_FILE)"
   

5. Wait for the application to be ready on the admin cluster.

   $ kubectl --kubeconfig $KUBECONFIG_ADMIN wait \
       --for=condition=Ready pods \
       --all -n hello-helidon \
       --timeout=300s
   

   Note: If you are returning the application to the managed cluster, then instead, wait for the application to be ready on the managed cluster.

   $ kubectl --kubeconfig $KUBECONFIG_MANAGED1 wait \
       --for=condition=Ready pods \
       --all -n hello-helidon \
       --timeout=300s
   

6. Now, you can test the example application running in its new location.

To return the application to the managed cluster named managed1, set the value of the CHANGE_PLACEMENT_PATCH_FILE environment variable to the patch file provided for that purpose, then repeat the previous numbered steps.

   # To change the placement back to the managed cluster named managed1
   $ export CHANGE_PLACEMENT_PATCH_FILE="https://raw.githubusercontent.com/verrazzano/verrazzano/v1.0.4/examples/multicluster/hello-helidon/patch-return-placement-to-managed1.yaml"

Undeploy the Hello World Helidon application

Regardless of its location, to undeploy the application, delete the multicluster resources and the project from the admin cluster:

# Delete the multicluster application configuration
$ kubectl --kubeconfig $KUBECONFIG_ADMIN delete \
    -f https://raw.githubusercontent.com/verrazzano/verrazzano/v1.0.4/examples/multicluster/hello-helidon/mc-hello-helidon-app.yaml
# Delete the multicluster components for the application
$ kubectl --kubeconfig $KUBECONFIG_ADMIN delete \
    -f https://raw.githubusercontent.com/verrazzano/verrazzano/v1.0.4/examples/multicluster/hello-helidon/mc-hello-helidon-comp.yaml
# Delete the project
$ kubectl --kubeconfig $KUBECONFIG_ADMIN delete \
    -f https://raw.githubusercontent.com/verrazzano/verrazzano/v1.0.4/examples/multicluster/hello-helidon/verrazzano-project.yaml