Automate remaining graceful recovery tests

.github/workflows/functional.yml

@@ -119,6 +119,13 @@ jobs:
           make test${{ inputs.image == 'plus' && '-with-plus' || ''}} PREFIX=${ngf_prefix} TAG=${ngf_tag} GINKGO_LABEL=telemetry GW_SERVICE_TYPE=LoadBalancer
         working-directory: ./tests
 
+      - name: Run functional graceful-recovery tests
+        run: |
+          ngf_prefix=ghcr.io/nginxinc/nginx-gateway-fabric
+          ngf_tag=${{ steps.ngf-meta.outputs.version }}
+          make test${{ inputs.image == 'plus' && '-with-plus' || ''}} PREFIX=${ngf_prefix} TAG=${ngf_tag} GINKGO_LABEL=graceful-recovery GW_SERVICE_TYPE=LoadBalancer CLUSTER_NAME=${{ github.run_id }}
+        working-directory: ./tests
+
       - name: Run functional tests
         run: |
           ngf_prefix=ghcr.io/nginxinc/nginx-gateway-fabric

tests/Makefile

@@ -124,7 +124,7 @@ test: ## Runs the functional tests on your default k8s cluster
 		--image-tag=$(TAG) --version-under-test=$(NGF_VERSION) \
 		--plus-enabled=$(PLUS_ENABLED) --ngf-image-repo=$(PREFIX) --nginx-image-repo=$(NGINX_PREFIX) --nginx-plus-image-repo=$(NGINX_PLUS_PREFIX) \
 		--pull-policy=$(PULL_POLICY) --service-type=$(GW_SERVICE_TYPE) \
-		--is-gke-internal-lb=$(GW_SVC_GKE_INTERNAL)
+		--is-gke-internal-lb=$(GW_SVC_GKE_INTERNAL) --cluster-name=$(CLUSTER_NAME)
 
 .PHONY: test-with-plus
 test-with-plus: PLUS_ENABLED=true

tests/suite/graceful_recovery_test.go

@@ -5,6 +5,7 @@ import (
 	"errors"
 	"fmt"
 	"net/http"
+	"os/exec"
 	"strings"
 	"time"
 
@@ -15,6 +16,7 @@ import (
 	core "k8s.io/api/core/v1"
 	metav1 "k8s.io/apimachinery/pkg/apis/meta/v1"
 	"k8s.io/apimachinery/pkg/types"
+	ctlr "sigs.k8s.io/controller-runtime"
 	"sigs.k8s.io/controller-runtime/pkg/client"
 	"sigs.k8s.io/yaml"
 
@@ -28,7 +30,7 @@ const (
 
 // Since checkContainerLogsForErrors may experience interference from previous tests (as explained in the function
 // documentation), this test is recommended to be run separate from other tests.
-var _ = Describe("Graceful Recovery test", Ordered, Label("functional", "graceful-recovery"), func() {
+var _ = Describe("Graceful Recovery test", Ordered, Label("graceful-recovery"), func() {
 	files := []string{
 		"graceful-recovery/cafe.yaml",
 		"graceful-recovery/cafe-secret.yaml",
@@ -45,10 +47,11 @@ var _ = Describe("Graceful Recovery test", Ordered, Label("functional", "gracefu
 
 	var ngfPodName string
 
-	BeforeAll(func() {
+	BeforeEach(func() {
 		// this test is unique in that it will check the entire log of both ngf and nginx containers
 		// for any errors, so in order to avoid errors generated in previous tests we will uninstall
-		// NGF installed at the suite level, then re-deploy our own
+		// NGF installed at the suite level, then re-deploy our own. We will also uninstall and re-install
+		// NGF between each graceful-recovery test for the same reason.
 		teardown(releaseName)
 
 		setup(getDefaultSetupCfg())
@@ -64,9 +67,7 @@ var _ = Describe("Graceful Recovery test", Ordered, Label("functional", "gracefu
 		if portFwdHTTPSPort != 0 {
 			teaURL = fmt.Sprintf("%s:%d/tea", baseHTTPSURL, portFwdHTTPSPort)
 		}
-	})
 
-	BeforeEach(func() {
 		ns = core.Namespace{
 			ObjectMeta: metav1.ObjectMeta{
 				Name: "graceful-recovery",
@@ -98,8 +99,97 @@ var _ = Describe("Graceful Recovery test", Ordered, Label("functional", "gracefu
 	It("recovers when nginx container is restarted", func() {
 		runRecoveryTest(teaURL, coffeeURL, ngfPodName, nginxContainerName, files, &ns)
 	})
+
+	It("recovers when drained node is restarted", func() {
+		runRestartNodeWithDrainingTest(teaURL, coffeeURL, files, &ns)
+	})
+
+	It("recovers when node is restarted abruptly", func() {
+		runRestartNodeAbruptlyTest(teaURL, coffeeURL, files, &ns)
+	})
 })
 
+func runRestartNodeWithDrainingTest(teaURL, coffeeURL string, files []string, ns *core.Namespace) {
+	runRestartNodeTest(teaURL, coffeeURL, files, ns, true)
+}
+
+func runRestartNodeAbruptlyTest(teaURL, coffeeURL string, files []string, ns *core.Namespace) {
+	runRestartNodeTest(teaURL, coffeeURL, files, ns, false)
+}
+
+func runRestartNodeTest(teaURL, coffeeURL string, files []string, ns *core.Namespace, drain bool) {
+	nodeNames, err := getNodeNames()
+	Expect(err).ToNot(HaveOccurred())
+	Expect(nodeNames).To(HaveLen(1))
+
+	kindNodeName := nodeNames[0]
+
+	Expect(clusterName).ToNot(BeNil(), "clusterName variable not set")
+	Expect(*clusterName).ToNot(BeEmpty())
+	containerName := *clusterName + "-control-plane"
+
+	if portFwdPort != 0 {
+		close(portForwardStopCh)
+	}
+
+	if drain {
+		_, err := exec.Command(
+			"kubectl",
+			"drain",
+			kindNodeName,
+			"--ignore-daemonsets",
+			"--delete-local-data",
+		).CombinedOutput()
+		Expect(err).ToNot(HaveOccurred())
+
+		_, err = exec.Command("kubectl", "delete", "node", kindNodeName).CombinedOutput()
+		Expect(err).ToNot(HaveOccurred())
+	}
+
+	_, err = exec.Command("docker", "restart", containerName).CombinedOutput()
+	Expect(err).ToNot(HaveOccurred())
+
+	// need to wait for docker container to restart and be running before polling for ready NGF Pods or else we will error
+	Eventually(
+		func() bool {
+			output, err := exec.Command(
+				"docker",
+				"inspect",
+				"-f",
+				"{{.State.Running}}",
+				containerName,
+			).CombinedOutput()
+			return strings.TrimSpace(string(output)) == "true" && err == nil
+		}).
+		WithTimeout(timeoutConfig.CreateTimeout).
+		WithPolling(500 * time.Millisecond).
+		Should(BeTrue())
+
+	// ngf can often oscillate between ready and error, so we wait for a stable readiness in ngf
+	var podNames []string
+	Eventually(
+		func() bool {
+			podNames, err = framework.GetReadyNGFPodNames(k8sClient, ngfNamespace, releaseName, timeoutConfig.GetStatusTimeout)
+			return len(podNames) == 1 && err == nil
+		}).
+		WithTimeout(timeoutConfig.CreateTimeout * 2).
+		WithPolling(500 * time.Millisecond).
+		MustPassRepeatedly(20).
+		Should(BeTrue())
+
+	ngfPodName := podNames[0]
+	Expect(ngfPodName).ToNot(BeEmpty())
+
+	if portFwdPort != 0 {
+		ports := []string{fmt.Sprintf("%d:80", ngfHTTPForwardedPort), fmt.Sprintf("%d:443", ngfHTTPSForwardedPort)}
+		portForwardStopCh = make(chan struct{})
+		err = framework.PortForward(ctlr.GetConfigOrDie(), ngfNamespace, ngfPodName, ports, portForwardStopCh)
+		Expect(err).ToNot(HaveOccurred())
+	}
+
+	checkNGFFunctionality(teaURL, coffeeURL, ngfPodName, "", files, ns)
+}
+
 func runRecoveryTest(teaURL, coffeeURL, ngfPodName, containerName string, files []string, ns *core.Namespace) {
 	var (
 		err       error
@@ -127,36 +217,7 @@ func runRecoveryTest(teaURL, coffeeURL, ngfPodName, containerName string, files
 			Should(Succeed())
 	}
 
-	Eventually(
-		func() error {
-			return checkForWorkingTraffic(teaURL, coffeeURL)
-		}).
-		WithTimeout(timeoutConfig.RequestTimeout).
-		WithPolling(500 * time.Millisecond).
-		Should(Succeed())
-
-	Expect(resourceManager.DeleteFromFiles(files, ns.Name)).To(Succeed())
-
-	Eventually(
-		func() error {
-			return checkForFailingTraffic(teaURL, coffeeURL)
-		}).
-		WithTimeout(timeoutConfig.RequestTimeout).
-		WithPolling(500 * time.Millisecond).
-		Should(Succeed())
-
-	Expect(resourceManager.ApplyFromFiles(files, ns.Name)).To(Succeed())
-	Expect(resourceManager.WaitForAppsToBeReadyWithPodCount(ns.Name, 2)).To(Succeed())
-
-	Eventually(
-		func() error {
-			return checkForWorkingTraffic(teaURL, coffeeURL)
-		}).
-		WithTimeout(timeoutConfig.RequestTimeout * 2).
-		WithPolling(500 * time.Millisecond).
-		Should(Succeed())
-
-	checkContainerLogsForErrors(ngfPodName, containerName == nginxContainerName)
+	checkNGFFunctionality(teaURL, coffeeURL, ngfPodName, containerName, files, ns)
 }
 
 func restartContainer(ngfPodName, containerName string) {
@@ -254,11 +315,41 @@ func expectRequestToFail(appURL, address string) error {
 	return nil
 }
 
+func checkNGFFunctionality(teaURL, coffeeURL, ngfPodName, containerName string, files []string, ns *core.Namespace) {
+	Eventually(
+		func() error {
+			return checkForWorkingTraffic(teaURL, coffeeURL)
+		}).
+		WithTimeout(timeoutConfig.RequestTimeout * 2).
+		WithPolling(500 * time.Millisecond).
+		Should(Succeed())
+
+	Expect(resourceManager.DeleteFromFiles(files, ns.Name)).To(Succeed())
+
+	Eventually(
+		func() error {
+			return checkForFailingTraffic(teaURL, coffeeURL)
+		}).
+		WithTimeout(timeoutConfig.RequestTimeout).
+		WithPolling(500 * time.Millisecond).
+		Should(Succeed())
+
+	Expect(resourceManager.ApplyFromFiles(files, ns.Name)).To(Succeed())
+	Expect(resourceManager.WaitForAppsToBeReadyWithPodCount(ns.Name, 2)).To(Succeed())
+
+	Eventually(
+		func() error {
+			return checkForWorkingTraffic(teaURL, coffeeURL)
+		}).
+		WithTimeout(timeoutConfig.RequestTimeout * 2).
+		WithPolling(500 * time.Millisecond).
+		Should(Succeed())
+
+	checkContainerLogsForErrors(ngfPodName, containerName == nginxContainerName)
+}
+
 // checkContainerLogsForErrors checks both nginx and ngf container's logs for any possible errors.
-// Since this function retrieves all the logs from both containers and the NGF pod may be shared between tests,
-// the logs retrieved may contain log messages from previous tests, thus any errors in the logs from previous tests
-// may cause an interference with this test and cause this test to fail.
-// Additionally, when the NGINX process is killed, some errors are expected in the NGF logs while we wait for the
+// When the NGINX process is killed, some errors are expected in the NGF logs while we wait for the
 // NGINX container to be restarted.
 func checkContainerLogsForErrors(ngfPodName string, checkNginxLogsOnly bool) {
 	nginxLogs, err := resourceManager.GetPodLogs(
@@ -349,6 +440,24 @@ func getContainerRestartCount(ngfPodName, containerName string) (int, error) {
 	return restartCount, nil
 }
 
+func getNodeNames() ([]string, error) {
+	ctx, cancel := context.WithTimeout(context.Background(), timeoutConfig.GetTimeout)
+	defer cancel()
+	var nodes core.NodeList
+
+	if err := k8sClient.List(ctx, &nodes); err != nil {
+		return nil, fmt.Errorf("error listing nodes: %w", err)
+	}
+
+	names := make([]string, 0, len(nodes.Items))
+
+	for _, node := range nodes.Items {
+		names = append(names, node.Name)
+	}
+
+	return names, nil
+}
+
 func runNodeDebuggerJob(ngfPodName, jobScript string) (*v1.Job, error) {
 	ctx, cancel := context.WithTimeout(context.Background(), timeoutConfig.GetTimeout)
 	defer cancel()

tests/suite/system_suite_test.go

@@ -58,6 +58,7 @@ var (
 	serviceType              = flag.String("service-type", "NodePort", "Type of service fronting NGF to be deployed")
 	isGKEInternalLB          = flag.Bool("is-gke-internal-lb", false, "Is the LB service GKE internal only")
 	plusEnabled              = flag.Bool("plus-enabled", false, "Is NGINX Plus enabled")
+	clusterName              = flag.String("cluster-name", "kind", "Cluster name")
 )
 
 var (
@@ -138,6 +139,10 @@ func setup(cfg setupConfig, extraInstallArgs ...string) {
 		Skip("GW_SERVICE_TYPE must be 'LoadBalancer' for NFR tests")
 	}
 
+	if clusterInfo.IsGKE && strings.Contains(GinkgoLabelFilter(), "graceful-recovery") {
+		Skip("Graceful Recovery test must be run on Kind")
+	}
+
 	if *versionUnderTest != "" {
 		version = *versionUnderTest
 	} else if *imageTag != "" {