Add support for Vault as a secrets backend

CHANGELOG.md

@@ -3,11 +3,25 @@
 BREAKING CHANGES:
 * Control Plane
   * Update minimum go version for project to 1.17 [[GH-878](https://github.com/hashicorp/consul-k8s/pull/878)]
-  * Add boolean metric to merged metrics response `consul_merged_service_metrics_success` to indicate if service metrics were
-    scraped successfully. [[GH-551](https://github.com/hashicorp/consul-k8s/pull/551)]
+  * Add boolean metric to merged metrics response `consul_merged_service_metrics_success` to indicate if service metrics
+    were scraped successfully. [[GH-551](https://github.com/hashicorp/consul-k8s/pull/551)]
+
+FEATURES:
+* Vault as a Secrets Backend: Add support for Vault as a secrets backend for Gossip Encryption, Server TLS certs and Service Mesh TLS certificates,
+    removing the existing usage of Kubernetes Secrets for the respective secrets. [[GH-904](https://github.com/hashicorp/consul-k8s/pull/904/)]
+
+  See the [Consul Kubernetes and Vault documentation](https://www.consul.io/docs/k8s/installation/vault)
+  for full install instructions.
+
+  Requirements: 
+  * Consul 1.11+
+  * Vault 1.19+ and Vault-K8s 0.14+ must be installed with the Vault Agent Injector enabled (`injector.enabled=true`)
+    into the Kubernetes cluster that Consul is installed into.
+  * `global.tls.enableAutoEncryption=true` is required for TLS support.
+  * If TLS is enabled in Vault, `global.secretsBackend.vault.ca` must be provided and should reference a Kube secret
+    which holds a copy of the Vault CA cert.
 
 IMPROVEMENTS:
-* CLI
   * Pre-check in the `install` command to verify the correct license secret exists when using an enterprise Consul image. [[GH-875](https://github.com/hashicorp/consul-k8s/pull/875)]
 * Control Plane
   * Add a label "managed-by" to every secret the control-plane creates. Only delete said secrets on an uninstall. [[GH-835](https://github.com/hashicorp/consul-k8s/pull/835)]

acceptance/framework/consul/consul_cluster.go

@@ -2,7 +2,6 @@ package consul
 
 import (
 	"context"
-	"encoding/json"
 	"fmt"
 	"strings"
 	"testing"
@@ -115,7 +114,7 @@ func (h *HelmCluster) Create(t *testing.T) {
 	})
 
 	// Fail if there are any existing installations of the Helm chart.
-	h.checkForPriorInstallations(t)
+	helpers.CheckForPriorInstallations(t, h.kubernetesClient, h.helmOptions, "consul-helm", "chart=consul-helm")
 
 	helm.Install(t, h.helmOptions, config.HelmChartPath, h.releaseName)
 
@@ -281,49 +280,6 @@ func (h *HelmCluster) SetupConsulClient(t *testing.T, secure bool) *api.Client {
 	return consulClient
 }
 
-// checkForPriorInstallations checks if there is an existing Helm release
-// for this Helm chart already installed. If there is, it fails the tests.
-func (h *HelmCluster) checkForPriorInstallations(t *testing.T) {
-	t.Helper()
-
-	var helmListOutput string
-	// Check if there's an existing cluster and fail if there is one.
-	// We may need to retry since this is the first command run once the Kube
-	// cluster is created and sometimes the API server returns errors.
-	retry.RunWith(&retry.Counter{Wait: 1 * time.Second, Count: 3}, t, func(r *retry.R) {
-		var err error
-		// NOTE: It's okay to pass in `t` to RunHelmCommandAndGetOutputE despite being in a retry
-		// because we're using RunHelmCommandAndGetOutputE (not RunHelmCommandAndGetOutput) so the `t` won't
-		// get used to fail the test, just for logging.
-		helmListOutput, err = helm.RunHelmCommandAndGetOutputE(t, h.helmOptions, "list", "--output", "json")
-		require.NoError(r, err)
-	})
-
-	var installedReleases []map[string]string
-
-	err := json.Unmarshal([]byte(helmListOutput), &installedReleases)
-	require.NoError(t, err, "unmarshalling %q", helmListOutput)
-
-	for _, r := range installedReleases {
-		require.NotContains(t, r["chart"], "consul", fmt.Sprintf("detected an existing installation of Consul %s, release name: %s", r["chart"], r["name"]))
-	}
-
-	// Wait for all pods in the "default" namespace to exit. A previous
-	// release may not be listed by Helm but its pods may still be terminating.
-	retry.RunWith(&retry.Counter{Wait: 1 * time.Second, Count: 60}, t, func(r *retry.R) {
-		consulPods, err := h.kubernetesClient.CoreV1().Pods(h.helmOptions.KubectlOptions.Namespace).List(context.Background(), metav1.ListOptions{})
-		require.NoError(r, err)
-		if len(consulPods.Items) > 0 {
-			var podNames []string
-			for _, p := range consulPods.Items {
-				podNames = append(podNames, p.Name)
-			}
-			r.Errorf("pods from previous installation still running: %s", strings.Join(podNames, ", "))
-		}
-	})
-
-}
-
 // configurePodSecurityPolicies creates a simple pod security policy, a cluster role to allow access to the PSP,
 // and a role binding that binds the default service account in the helm installation namespace to the cluster role.
 // We bind the default service account for tests that are spinning up pods without a service account set so that

acceptance/framework/helpers/helpers.go

@@ -2,6 +2,7 @@ package helpers
 
 import (
 	"context"
+	"encoding/json"
 	"fmt"
 	"os"
 	"os/signal"
@@ -10,6 +11,8 @@ import (
 	"testing"
 	"time"
 
+	"github.com/gruntwork-io/terratest/modules/helm"
+
 	terratestk8s "github.com/gruntwork-io/terratest/modules/k8s"
 	"github.com/gruntwork-io/terratest/modules/random"
 	"github.com/hashicorp/consul-k8s/acceptance/framework/logger"
@@ -25,14 +28,56 @@ func RandomName() string {
 	return fmt.Sprintf("test-%s", strings.ToLower(random.UniqueId()))
 }
 
+// CheckForPriorInstallations checks if there is an existing Helm release
+// for this Helm chart already installed. If there is, it fails the tests.
+func CheckForPriorInstallations(t *testing.T, client kubernetes.Interface, options *helm.Options, chartName, labelSelector string) {
+	t.Helper()
+
+	var helmListOutput string
+	// Check if there's an existing cluster and fail if there is one.
+	// We may need to retry since this is the first command run once the Kube
+	// cluster is created and sometimes the API server returns errors.
+	retry.RunWith(&retry.Counter{Wait: 1 * time.Second, Count: 3}, t, func(r *retry.R) {
+		var err error
+		// NOTE: It's okay to pass in `t` to RunHelmCommandAndGetOutputE despite being in a retry
+		// because we're using RunHelmCommandAndGetOutputE (not RunHelmCommandAndGetOutput) so the `t` won't
+		// get used to fail the test, just for logging.
+		helmListOutput, err = helm.RunHelmCommandAndGetOutputE(t, options, "list", "--output", "json")
+		require.NoError(r, err)
+	})
+
+	var installedReleases []map[string]string
+
+	err := json.Unmarshal([]byte(helmListOutput), &installedReleases)
+	require.NoError(t, err, "unmarshalling %q", helmListOutput)
+
+	for _, r := range installedReleases {
+		require.NotContains(t, r["chart"], chartName, fmt.Sprintf("detected an existing installation of %s %s, release name: %s", chartName, r["chart"], r["name"]))
+	}
+
+	// Wait for all pods in the "default" namespace to exit. A previous
+	// release may not be listed by Helm but its pods may still be terminating.
+	retry.RunWith(&retry.Counter{Wait: 1 * time.Second, Count: 60}, t, func(r *retry.R) {
+		pods, err := client.CoreV1().Pods(options.KubectlOptions.Namespace).List(context.Background(), metav1.ListOptions{LabelSelector: labelSelector})
+		require.NoError(r, err)
+		if len(pods.Items) > 0 {
+			var podNames []string
+			for _, p := range pods.Items {
+				podNames = append(podNames, p.Name)
+			}
+			r.Errorf("pods from previous installation still running: %s", strings.Join(podNames, ", "))
+		}
+	})
+}
+
 // WaitForAllPodsToBeReady waits until all pods with the provided podLabelSelector
 // are in the ready status. It checks every 5 seconds for a total of 20 tries.
 // If there is at least one container in a pod that isn't ready after that,
 // it fails the test.
 func WaitForAllPodsToBeReady(t *testing.T, client kubernetes.Interface, namespace, podLabelSelector string) {
 	t.Helper()
 
-	logger.Log(t, "Waiting for pods to be ready.")
+	logger.Logf(t, "Waiting for pods with label %q to be ready.", podLabelSelector)
 
 	// Wait up to 10m.
 	// On Azure, volume provisioning can sometimes take close to 5 min,
@@ -41,6 +86,7 @@ func WaitForAllPodsToBeReady(t *testing.T, client kubernetes.Interface, namespac
 	retry.RunWith(counter, t, func(r *retry.R) {
 		pods, err := client.CoreV1().Pods(namespace).List(context.Background(), metav1.ListOptions{LabelSelector: podLabelSelector})
 		require.NoError(r, err)
+		require.NotEmpty(r, pods.Items)
 
 		var notReadyPods []string
 		for _, pod := range pods.Items {
@@ -55,7 +101,7 @@ func WaitForAllPodsToBeReady(t *testing.T, client kubernetes.Interface, namespac
 	logger.Log(t, "Finished waiting for pods to be ready.")
 }
 
-// Sets up a goroutine that will wait for interrupt signals
+// SetupInterruptHandler sets up a goroutine that will wait for interrupt signals
 // and call cleanup function when it catches it.
 func SetupInterruptHandler(cleanup func()) {
 	c := make(chan os.Signal, 1)