[Release/1.15.x] net 3914 gha consul container test no splitting

.github/workflows/test-integrations.yml

@@ -343,50 +343,16 @@ jobs:
           DD_ENV: ci
         run: datadog-ci junit upload --service "$GITHUB_REPOSITORY" $TEST_RESULTS_DIR/results.xml
 
-  generate-compatibility-job-matrices:
-    needs: [setup]
-    runs-on: ${{ fromJSON(needs.setup.outputs.compute-small) }}
-    name: Generate Compatibility Job Matrices
-    outputs:
-      compatibility-matrix: ${{ steps.set-matrix.outputs.compatibility-matrix }}
-    steps:
-      - uses: actions/checkout@24cb9080177205b6e8c946b17badbe402adc938f # v3.4.0
-      - name: Generate Compatibility Job Matrix
-        id: set-matrix
-        env:
-          TOTAL_RUNNERS: 6
-          JQ_SLICER: '[ inputs ] | [_nwise(length / $runnercount | floor)]'
-        run: |
-          cd ./test/integration/consul-container
-          NUM_RUNNERS=$TOTAL_RUNNERS
-          NUM_DIRS=$(find ./test -mindepth 1 -maxdepth 2 -type d | wc -l)
-
-          if [ "$NUM_DIRS" -lt "$NUM_RUNNERS" ]; then
-            echo "TOTAL_RUNNERS is larger than the number of tests/packages to split."
-            NUM_RUNNERS=$((NUM_DIRS-1))
-          fi
-          # fix issue where test splitting calculation generates 1 more split than TOTAL_RUNNERS.
-          NUM_RUNNERS=$((NUM_RUNNERS-1))
-          {
-            echo -n "compatibility-matrix="
-            find ./test -maxdepth 2 -type d -print0 | xargs -0 -n 1 \
-              | grep -v util | grep -v upgrade \
-              | jq --raw-input --argjson runnercount "$NUM_RUNNERS" "$JQ_SLICER" \
-              | jq --compact-output 'map(join(" "))'
-          } >> "$GITHUB_OUTPUT"
   compatibility-integration-test:
     runs-on: ${{ fromJSON(needs.setup.outputs.compute-xl) }}
     needs:
       - setup
       - dev-build
-      - generate-compatibility-job-matrices
     permissions:
       id-token: write # NOTE: this permission is explicitly required for Vault auth.
       contents: read
-    strategy:
-      fail-fast: false
-      matrix:
-        test-cases: ${{ fromJSON(needs.generate-compatibility-job-matrices.outputs.compatibility-matrix) }}
+    env:
+      ENVOY_VERSION: "1.24.6"
     steps:
       - uses: actions/checkout@24cb9080177205b6e8c946b17badbe402adc938f # v3.4.0
       - uses: actions/setup-go@6edd4406fa81c3da01a34fa6f6343087c207a568 # v3.5.0
@@ -402,7 +368,6 @@ jobs:
           path: .
       - name: restore mode+x
         run: chmod +x consul
-
       - name: Build consul:local image
         run: docker build -t ${{ env.CONSUL_LATEST_IMAGE_NAME }}:local -f ./build-support/docker/Consul-Dev.dockerfile .
       - name: Configure GH workaround for ipv6 loopback
@@ -416,21 +381,18 @@ jobs:
             mkdir -p "/tmp/test-results"
             cd ./test/integration/consul-container
             docker run --rm ${{ env.CONSUL_LATEST_IMAGE_NAME }}:local consul version
-            echo "Running $(sed 's,|, ,g' <<< "${{ matrix.test-cases }}" |wc -w) subtests"
-            # shellcheck disable=SC2001
-            sed 's, ,\n,g' <<< "${{ matrix.test-cases }}"
             go run gotest.tools/gotestsum@v${{env.GOTESTSUM_VERSION}} \
               --raw-command \
-              --format=short-verbose \
+              --format=standard-verbose \
               --debug \
               --rerun-fails=3 \
               -- \
               go test \
-              -p=4 \
+              -p=6 \
               -tags "${{ env.GOTAGS }}" \
               -timeout=30m \
               -json \
-              ${{ matrix.test-cases }} \
+              `go list ./... | grep -v upgrade` \
               --target-image ${{ env.CONSUL_LATEST_IMAGE_NAME }} \
               --target-version local \
               --latest-image docker.mirror.hashicorp.services/${{ env.CONSUL_LATEST_IMAGE_NAME }} \
@@ -474,53 +436,19 @@ jobs:
           DD_ENV: ci
         run: datadog-ci junit upload --service "$GITHUB_REPOSITORY" $TEST_RESULTS_DIR/results.xml
 
-  generate-upgrade-job-matrices:
-    needs: [setup]
-    runs-on: ${{ fromJSON(needs.setup.outputs.compute-small) }}
-    name: Generate Upgrade Job Matrices
-    outputs:
-      upgrade-matrix: ${{ steps.set-matrix.outputs.upgrade-matrix }}
-    steps:
-      - uses: actions/checkout@24cb9080177205b6e8c946b17badbe402adc938f # v3.4.0
-      - uses: actions/setup-go@6edd4406fa81c3da01a34fa6f6343087c207a568 # v3.5.0
-        with:
-          go-version-file: 'go.mod'
-      - name: Generate Updgrade Job Matrix
-        id: set-matrix
-        env:
-          TOTAL_RUNNERS: 5
-          JQ_SLICER: '[ inputs ] | [_nwise(length / $runnercount | floor)]'
-        run: |
-          cd ./test/integration/consul-container/test/upgrade
-          NUM_RUNNERS=$TOTAL_RUNNERS
-          NUM_DIRS=$(go test ./...  -list=. -json | jq -r '.Output | select (. !=null) |  select(. | startswith("Test")) | gsub("[\\n\\t]"; "")' | wc -l)
-
-          if [ "$NUM_DIRS" -lt "$NUM_RUNNERS" ]; then
-            echo "TOTAL_RUNNERS is larger than the number of tests/packages to split."
-            NUM_RUNNERS=$((NUM_DIRS-1))
-          fi
-          # fix issue where test splitting calculation generates 1 more split than TOTAL_RUNNERS.
-          NUM_RUNNERS=$((NUM_RUNNERS-1))
-          {
-            echo -n "upgrade-matrix="
-            go test ./...  -list=. -json | jq -r '.Output | select (. !=null) |  select(. | startswith("Test")) | gsub("[\\n\\t]"; "")' \
-            | jq --raw-input --argjson runnercount "$NUM_RUNNERS" "$JQ_SLICER" \
-            | jq --compact-output 'map(join("|"))'
-          } >> "$GITHUB_OUTPUT"
 
   upgrade-integration-test:
     runs-on: ${{ fromJSON(needs.setup.outputs.compute-xl) }}
     needs:
       - setup
       - dev-build
-      - generate-upgrade-job-matrices
     strategy:
       fail-fast: false
       matrix:
         consul-version: [ "1.14", "1.15"]
-        test-cases: ${{ fromJSON(needs.generate-upgrade-job-matrices.outputs.upgrade-matrix) }}
     env:
-      CONSUL_VERSION: ${{ matrix.consul-version }}
+      CONSUL_LATEST_VERSION: ${{ matrix.consul-version }}
+      ENVOY_VERSION: "1.24.6"
     steps:
       - uses: actions/checkout@24cb9080177205b6e8c946b17badbe402adc938f # v3.4.0
       - uses: actions/setup-go@6edd4406fa81c3da01a34fa6f6343087c207a568 # v3.5.0
@@ -549,9 +477,6 @@ jobs:
           mkdir -p "${{ env.TEST_RESULTS_DIR }}"
           cd ./test/integration/consul-container/test/upgrade
           docker run --rm ${{ env.CONSUL_LATEST_IMAGE_NAME }}:local consul version
-          echo "Running $(sed 's,|, ,g' <<< "${{ matrix.test-cases }}" |wc -w) subtests"
-          # shellcheck disable=SC2001
-          sed 's,|,\n,g' <<< "${{ matrix.test-cases }}"
           go run gotest.tools/gotestsum@v${{env.GOTESTSUM_VERSION}} \
             --raw-command \
             --format=short-verbose \
@@ -564,11 +489,10 @@ jobs:
             -tags "${{ env.GOTAGS }}" \
             -timeout=30m \
             -json ./... \
-            -run "${{ matrix.test-cases }}" \
             --target-image ${{ env.CONSUL_LATEST_IMAGE_NAME }} \
             --target-version local \
-            --latest-image ${{ env.CONSUL_LATEST_IMAGE_NAME }} \
-            --latest-version "${{ env.CONSUL_VERSION }}"
+            --latest-image docker.mirror.hashicorp.services/${{ env.CONSUL_LATEST_IMAGE_NAME }} \
+            --latest-version "${{ env.CONSUL_LATEST_VERSION }}"
           ls -lrt
         env:
           # this is needed because of incompatibility between RYUK container and GHA
@@ -590,9 +514,7 @@ jobs:
     - vault-integration-test
     - generate-envoy-job-matrices
     - envoy-integration-test
-    - generate-compatibility-job-matrices
     - compatibility-integration-test
-    - generate-upgrade-job-matrices
     - upgrade-integration-test
     runs-on: ${{ fromJSON(needs.setup.outputs.compute-small) }}
     if: ${{ always() }}

test/integration/consul-container/libs/service/connect.go

@@ -182,7 +182,7 @@ func NewConnectService(ctx context.Context, sidecarCfg SidecarConfig, serviceBin
 
 	req := testcontainers.ContainerRequest{
 		FromDockerfile: dockerfileCtx,
-		WaitingFor:     wait.ForLog("").WithStartupTimeout(10 * time.Second),
+		WaitingFor:     wait.ForLog("").WithStartupTimeout(100 * time.Second),
 		AutoRemove:     false,
 		Name:           containerName,
 		Cmd: []string{

test/integration/consul-container/libs/service/examples.go

@@ -150,7 +150,7 @@ func NewExampleService(ctx context.Context, name string, httpPort int, grpcPort
 
 	req := testcontainers.ContainerRequest{
 		Image:      hashicorpDockerProxy + "/fortio/fortio",
-		WaitingFor: wait.ForLog("").WithStartupTimeout(10 * time.Second),
+		WaitingFor: wait.ForLog("").WithStartupTimeout(100 * time.Second),
 		AutoRemove: false,
 		Name:       containerName,
 		Cmd:        command,

test/integration/consul-container/libs/service/gateway.go

@@ -174,7 +174,7 @@ func NewGatewayService(ctx context.Context, gwCfg GatewayConfig, node libcluster
 
 	req := testcontainers.ContainerRequest{
 		FromDockerfile: dockerfileCtx,
-		WaitingFor:     wait.ForLog("").WithStartupTimeout(10 * time.Second),
+		WaitingFor:     wait.ForLog("").WithStartupTimeout(100 * time.Second),
 		AutoRemove:     false,
 		Name:           containerName,
 		Cmd: []string{

test/integration/consul-container/libs/topology/peering_topology.go

@@ -29,6 +29,13 @@ type BuiltCluster struct {
 	Gateway   libservice.Service
 }
 
+type PeeringClusterSize struct {
+	AcceptingNumServers int
+	AcceptingNumClients int
+	DialingNumServers   int
+	DialingNumClients   int
+}
+
 // BasicPeeringTwoClustersSetup sets up a scenario for testing peering, which consists of
 //
 //   - an accepting cluster with 3 servers and 1 client agent. The client should be used to
@@ -43,11 +50,12 @@ func BasicPeeringTwoClustersSetup(
 	t *testing.T,
 	consulImage string,
 	consulVersion string,
+	pcs PeeringClusterSize,
 	peeringThroughMeshgateway bool,
 ) (*BuiltCluster, *BuiltCluster) {
 	acceptingCluster, acceptingCtx, acceptingClient := NewCluster(t, &ClusterConfig{
-		NumServers: 3,
-		NumClients: 1,
+		NumServers: pcs.AcceptingNumServers,
+		NumClients: pcs.AcceptingNumClients,
 		BuildOpts: &libcluster.BuildOptions{
 			Datacenter:           "dc1",
 			ConsulImageName:      consulImage,
@@ -58,8 +66,8 @@ func BasicPeeringTwoClustersSetup(
 	})
 
 	dialingCluster, dialingCtx, dialingClient := NewCluster(t, &ClusterConfig{
-		NumServers: 1,
-		NumClients: 1,
+		NumServers: pcs.DialingNumServers,
+		NumClients: pcs.DialingNumClients,
 		BuildOpts: &libcluster.BuildOptions{
 			Datacenter:           "dc2",
 			ConsulImageName:      consulImage,

test/integration/consul-container/libs/utils/version.go

@@ -2,6 +2,7 @@ package utils
 
 import (
 	"flag"
+	"strings"
 
 	"github.com/hashicorp/go-version"
 )
@@ -43,7 +44,7 @@ func GetLatestImageName() string {
 
 func DockerImage(image, version string) string {
 	v := image + ":" + version
-	if image == DefaultImageNameENT && isSemVer(version) {
+	if strings.Contains(image, DefaultImageNameENT) && isSemVer(version) {
 		// Enterprise versions get a suffix.
 		v += ImageVersionSuffixENT
 	}

test/integration/consul-container/test/gateways/http_route_test.go

@@ -504,8 +504,6 @@ func TestHTTPRouteParentRefChange(t *testing.T) {
 		t.Skip("too slow for testing.Short")
 	}
 
-	t.Parallel()
-
 	// infrastructure set up
 	address := "localhost"
 

test/integration/consul-container/test/peering/rotate_server_and_ca_then_fail_test.go

@@ -50,7 +50,13 @@ import (
 func TestPeering_RotateServerAndCAThenFail_(t *testing.T) {
 	t.Parallel()
 
-	accepting, dialing := libtopology.BasicPeeringTwoClustersSetup(t, utils.GetTargetImageName(), utils.TargetVersion, false)
+	accepting, dialing := libtopology.BasicPeeringTwoClustersSetup(t, utils.GetTargetImageName(), utils.TargetVersion,
+		libtopology.PeeringClusterSize{
+			AcceptingNumServers: 3,
+			AcceptingNumClients: 1,
+			DialingNumServers:   1,
+			DialingNumClients:   1,
+		}, false)
 	var (
 		acceptingCluster     = accepting.Cluster
 		dialingCluster       = dialing.Cluster

test/integration/consul-container/test/upgrade/common.go

@@ -0,0 +1,81 @@
+package upgrade
+
+import (
+	"context"
+	"fmt"
+
+	"github.com/hashicorp/consul/api"
+	"github.com/hashicorp/consul/test/integration/consul-container/libs/cluster"
+	libservice "github.com/hashicorp/consul/test/integration/consul-container/libs/service"
+	"github.com/hashicorp/consul/test/integration/consul-container/libs/utils"
+)
+
+// CreateAndRegisterStaticClientSidecarWith2Upstreams creates a static-client that
+// has two upstreams connecting to destinationNames: local bind addresses are 5000
+// and 5001.
+// - crossCluster: true if upstream is in another cluster
+func CreateAndRegisterStaticClientSidecarWith2Upstreams(c *cluster.Cluster, destinationNames []string, crossCluster bool) (*libservice.ConnectContainer, error) {
+	// Do some trickery to ensure that partial completion is correctly torn
+	// down, but successful execution is not.
+	var deferClean utils.ResettableDefer
+	defer deferClean.Execute()
+
+	node := c.Servers()[0]
+	mgwMode := api.MeshGatewayModeLocal
+
+	// Register the static-client service and sidecar first to prevent race with sidecar
+	// trying to get xDS before it's ready
+	req := &api.AgentServiceRegistration{
+		Name: libservice.StaticClientServiceName,
+		Port: 8080,
+		Connect: &api.AgentServiceConnect{
+			SidecarService: &api.AgentServiceRegistration{
+				Proxy: &api.AgentServiceConnectProxyConfig{
+					Upstreams: []api.Upstream{
+						{
+							DestinationName:  destinationNames[0],
+							LocalBindAddress: "0.0.0.0",
+							LocalBindPort:    cluster.ServiceUpstreamLocalBindPort,
+						},
+						{
+							DestinationName:  destinationNames[1],
+							LocalBindAddress: "0.0.0.0",
+							LocalBindPort:    cluster.ServiceUpstreamLocalBindPort2,
+						},
+					},
+				},
+			},
+		},
+	}
+
+	if crossCluster {
+		for _, upstream := range req.Connect.SidecarService.Proxy.Upstreams {
+			upstream.MeshGateway = api.MeshGatewayConfig{
+				Mode: mgwMode,
+			}
+		}
+	}
+
+	if err := node.GetClient().Agent().ServiceRegister(req); err != nil {
+		return nil, err
+	}
+
+	// Create a service and proxy instance
+	sidecarCfg := libservice.SidecarConfig{
+		Name:      fmt.Sprintf("%s-sidecar", libservice.StaticClientServiceName),
+		ServiceID: libservice.StaticClientServiceName,
+	}
+
+	clientConnectProxy, err := libservice.NewConnectService(context.Background(), sidecarCfg, []int{cluster.ServiceUpstreamLocalBindPort, cluster.ServiceUpstreamLocalBindPort2}, node)
+	if err != nil {
+		return nil, err
+	}
+	deferClean.Add(func() {
+		_ = clientConnectProxy.Terminate()
+	})
+
+	// disable cleanup functions now that we have an object with a Terminate() function
+	deferClean.Reset()
+
+	return clientConnectProxy, nil
+}

test/integration/consul-container/test/upgrade/fullstopupgrade_test.go

@@ -16,7 +16,6 @@ import (
 
 // Test upgrade a cluster of latest version to the target version
 func TestStandardUpgradeToTarget_fromLatest(t *testing.T) {
-	t.Parallel()
 
 	type testcase struct {
 		oldversion    string

test/integration/consul-container/test/upgrade/peering_control_plane_mgw_test.go → test/integration/consul-container/test/upgrade/peering/peering_control_plane_mgw_test.go

@@ -1,4 +1,4 @@
-package upgrade
+package peering
 
 import (
 	"context"
@@ -42,7 +42,14 @@ func TestPeering_Upgrade_ControlPlane_MGW(t *testing.T) {
 	}
 
 	run := func(t *testing.T, tc testcase) {
-		accepting, dialing := libtopology.BasicPeeringTwoClustersSetup(t, utils.GetLatestImageName(), tc.oldversion, true)
+		accepting, dialing := libtopology.BasicPeeringTwoClustersSetup(t, utils.GetLatestImageName(), utils.LatestVersion,
+			libtopology.PeeringClusterSize{
+				AcceptingNumServers: 1,
+				AcceptingNumClients: 1,
+				DialingNumServers:   1,
+				DialingNumClients:   1,
+			},
+			true)
 		var (
 			acceptingCluster = accepting.Cluster
 			dialingCluster   = dialing.Cluster