README
¶
Integration Tests
Integration tests should be used to test a specific functionality of k3s that exists across multiple Go packages, either via exported function calls, or more often, CLI commands. Integration tests should be used for "black box" testing.
Framework
All integration tests in K3s follow a Behavior Diven Development (BDD) style. Specifically, K3s uses Ginkgo and Gomega to drive the tests.
To generate an initial test, the command ginkgo bootstrap can be used.
To facilitate K3s CLI testing, see tests/util/cmd.go helper functions.
Format
All integration tests should be placed under tests/integration/<TEST_NAME>.
All integration test files should be named: <TEST_NAME>_int_test.go.
All integration test functions should be named: Test_Integration<TEST_NAME>.
See the local storage test as an example.
Running
Integration tests can be run with no k3s cluster present, each test will spin up and kill the appropriate k3s server it needs.
Note: Integration tests must be run as root, prefix the commands below with sudo -E env "PATH=$PATH" if a sudo user.
go test ./tests/integration/... -run Integration
Additionally, to generate JUnit reporting for the tests, the Ginkgo CLI is used
ginkgo --junit-report=result.xml ./tests/integration/...
Integration tests can be run on an existing single-node cluster via compile time flag, tests will skip if the server is not configured correctly.
go test -ldflags "-X 'github.com/k3s-io/k3s/tests/integration.existingServer=True'" ./tests/integration/... -run Integration
Integration tests can also be run via a Sonobuoy plugin on an existing single-node cluster.
./scripts/build-tests-sonobuoy
sudo KUBECONFIG=/etc/rancher/k3s/k3s.yaml sonobuoy run --plugin ./dist/artifacts/k3s-int-tests.yaml
Check the sonobuoy status and retrieve results
sudo KUBECONFIG=/etc/rancher/k3s/k3s.yaml sonobuoy status
sudo KUBECONFIG=/etc/rancher/k3s/k3s.yaml sonobuoy retrieve
sudo KUBECONFIG=/etc/rancher/k3s/k3s.yaml sonobuoy results <TAR_FILE_FROM_RETRIEVE>
Documentation
¶
Index ¶
- func CheckDeployments(deployments []string) error
- func GetEndpointsAddresses() (string, error)
- func GetPersistentVolume(name string) (*corev1.PersistentVolume, error)
- func GetPersistentVolumeClaim(namespace, name string) (*corev1.PersistentVolumeClaim, error)
- func GetPod(namespace, name string) (*corev1.Pod, error)
- func IsExistingServer() bool
- func IsRoot() bool
- func K3sCleanup(k3sTestLock int, dataDir string) error
- func K3sCmd(inputArgs ...string) (string, error)
- func K3sDefaultDeployments() error
- func K3sKillServer(server *K3sServer) error
- func K3sSaveLog(server *K3sServer, dump bool) error
- func K3sServerArgs() []string
- func K3sStopServer(server *K3sServer) error
- func K3sTestLock() (int, error)
- func ParseNodes() ([]corev1.Node, error)
- func ParsePods(namespace string, opts metav1.ListOptions) ([]corev1.Pod, error)
- func RunCommand(cmd string) (string, error)
- func SearchK3sLog(k3s *K3sServer, target string) (bool, error)
- func ServerArgsPresent(neededArgs []string) bool
- type K3sServer
Constants ¶
This section is empty.
Variables ¶
This section is empty.
Functions ¶
func CheckDeployments ¶
CheckDeployments checks if the provided list of deployments are ready, otherwise returns an error
func GetEndpointsAddresses ¶
func GetPersistentVolume ¶
func GetPersistentVolume(name string) (*corev1.PersistentVolume, error)
func GetPersistentVolumeClaim ¶
func GetPersistentVolumeClaim(namespace, name string) (*corev1.PersistentVolumeClaim, error)
func IsExistingServer ¶
func IsExistingServer() bool
func K3sCleanup ¶
K3sCleanup unlocks the test-lock and attempts to cleanup networking and files leftover from an integration test. This is similar to the k3s-killall.sh script, but we dynamically generate that on install, so we don't have access to it during testing.
func K3sCmd ¶
K3sCmd launches the provided K3s command via exec. Command blocks until finished. Command output from both Stderr and Stdout is provided via string. Input can be a single string with space separated args, or multiple string args cmdEx1, err := K3sCmd("etcd-snapshot", "ls") cmdEx2, err := K3sCmd("kubectl get pods -A") cmdEx2, err := K3sCmd("kubectl", "get", "pods", "-A")
func K3sDefaultDeployments ¶
func K3sDefaultDeployments() error
K3sDefaultDeployments checks if the default deployments for K3s are ready, otherwise returns an error
func K3sKillServer ¶
K3sKillServer terminates the running K3s server and its children. Equivalent to k3s-killall.sh
func K3sSaveLog ¶
func K3sServerArgs ¶
func K3sServerArgs() []string
K3sServerArgs returns the list of arguments that the k3s server launched with
func K3sStopServer ¶
K3sStopServer gracefully stops the running K3s server and does not kill its children. Equivalent to stopping the K3s service
func K3sTestLock ¶
func ParseNodes ¶
func ServerArgsPresent ¶
ServerArgsPresent checks if the given arguments are found in the running k3s server
Types ¶
type K3sServer ¶
type K3sServer struct {
// contains filtered or unexported fields
}
func K3sStartServer ¶
K3sStartServer acquires an exclusive lock on a temporary file, then launches a k3s cluster with the provided arguments. Subsequent/parallel calls to this function will block until the original lock is cleared using K3sKillServer
Source Files