Harvester Testing

This path contains the basic testing framework and integration testing cases, the primary focus in here to covering all Harvester APIs with integration testing cases.

Integration test

The integration test is performed on every pull request and is run by the Drone CI. The script is stored in the scripts/test-integration file.

How to run

• Configure appropriate environment variables

  • set KUBECONFIG path and configure export USE_EXISTING_CLUSTER=true to use an existing cluster, default to create a local Kind cluster.
    • For the local Kind cluster, you can specify export KIND_IMAGE_MIRROR="http://your-mirror-address:port" to speed up pulling images.
  • configure export SKIP_HARVESTER_INSTALLATION=true to skip Harvester installation, default to install a new harvester chart or update the existing one.
  • set export DONT_USE_EMULATION=true if the harvester runtime supports KVM, default to using QEMU software emulation.
  • set export KEEP_TESTING_CLUSTER=true if you want to keep the testing cluster, default to delete the local Kind cluster.
  • set export KEEP_HARVESTER_INSTALLATION=true if you want to keep the harvester installation, default to delete it on exit, this will always be true when SKIP_HARVESTER_INSTALLATION=true.
  • set export KEEP_TESTING_RESOURCE=true if we want to keep testing resources, default to delete each testing resource after each test case finished.
  • set export ENABLE_E2E_TESTS=true if want to run e2e test, the user must have an existing Kubernetes cluster.

• Run ./scripts/test-integration or make test-integration.

  • if you want to use an existing cluster with a pre-installed harvester chart, make sure to config the replicas of harvester deployment to be 0, then run $ env USE_EXISTING_CLUSTER=true SKIP_HARVESTER_INSTALLATION=true ./scripts/test-integration.