[WIP] testing

.github/workflows/checkpatch.yaml

@@ -1,7 +1,7 @@
 name: Check pull request commits
 on:
   push:
-  pull_request:
+  # pull_request:
 
 jobs:
   checkpatch:

.github/workflows/gotests.yml

@@ -1,8 +1,8 @@
 name: Tetragon Go Test
 on:
-  pull_request:
-    paths-ignore:
-      - 'docs/**'
+  # pull_request:
+  #   paths-ignore:
+  #     - 'docs/**'
   push:
     branches:
       - main

.github/workflows/lint-codeql.yml

@@ -1,11 +1,11 @@
 name: codeql
 
 on:
-  pull_request:
-    branches:
-      - main
-    paths-ignore:
-      - 'docs/**'
+  # pull_request:
+  #   branches:
+  #     - main
+  #   paths-ignore:
+  #     - 'docs/**'
   push:
     branches:
       - main

.github/workflows/packages-e2e-tests.yaml

@@ -1,10 +1,13 @@
 name: Packages e2e Tests
 
 on:
-  pull_request:
-    paths-ignore:
-    - "**.md"
-    - 'docs/**'
+  push:
+    branches:
+      - main
+  # pull_request:
+  #   paths-ignore:
+  #   - "**.md"
+  #   - 'docs/**'
 
 jobs:
   standalone-tarball-builds:

.github/workflows/podinfo-test.yaml

@@ -1,12 +1,12 @@
 name: PodInfo Integration Test
 on:
-  pull_request:
-    types:
-      - opened
-      - synchronize
-      - reopened
-    paths-ignore:
-      - 'docs/**'
+  # pull_request:
+  #   types:
+  #     - opened
+  #     - synchronize
+  #     - reopened
+  #   paths-ignore:
+  #     - 'docs/**'
   push:
     branches:
       - main

.github/workflows/run-e2e-tests.yaml

@@ -13,9 +13,6 @@ concurrency:
   group: ${{ github.ref }}
   cancel-in-progress: true
 
-env:
-  clusterName: tetragon-ci
-
 jobs:
   prepare:
     runs-on: ubuntu-22.04
@@ -74,7 +71,7 @@ jobs:
     strategy:
       fail-fast: false
       matrix:
-        os: [ ubuntu-22.04, ubuntu-22.04-arm64 ]
+        os: [ ubuntu-22.04 ]
         package: ${{fromJson(needs.prepare.outputs.packages)}}
     steps:
     - name: Checkout Code
@@ -117,8 +114,18 @@ jobs:
       env:
         GHA_OS: ${{matrix.os}}
       run: |
+        set -x
         cd go/src/github.com/cilium/tetragon
-        make e2e-test E2E_TESTS=${{matrix.package.f}} E2E_BUILD_IMAGES=0 E2E_AGENT=${{ needs.prepare.outputs.agentImage }} E2E_OPERATOR=${{ needs.prepare.outputs.operatorImage }} EXTRA_TESTFLAGS="-cluster-name=${{ env.clusterName }} -args -v=4"
+        make e2e-test \
+          E2E_TESTS=${{matrix.package.f}} \
+          E2E_BUILD_IMAGES=0 \
+          E2E_AGENT=${{ needs.prepare.outputs.agentImage }} \
+          E2E_OPERATOR=${{ needs.prepare.outputs.operatorImage }}
+
+    - name: ls logs
+      if: failure() || cancelled()
+      run: |
+        ls /tmp/tetragon.e2e.*
 
     - name: Upload Tetragon Logs
       if: failure() || cancelled()

.github/workflows/static-checks.yaml

@@ -10,10 +10,10 @@ on:
     paths-ignore:
       - 'docs/**'
       - '**.md'
-  pull_request:
-    paths-ignore:
-      - 'docs/**'
-      - '**.md'
+  # pull_request:
+  #   paths-ignore:
+  #     - 'docs/**'
+  #     - '**.md'
 
 permissions:
   # For golangci/golangci-lint to have read access to pull request for `only-new-issues` option.

.github/workflows/vmtests.yml

@@ -1,8 +1,8 @@
 name: vmtests
 on:
-  pull_request:
-    paths-ignore:
-      - 'docs/**'
+  # pull_request:
+  #   paths-ignore:
+  #     - 'docs/**'
   push:
     branches:
       - main

Makefile

@@ -263,7 +263,10 @@ copy-golangci-lint:
 
 .PHONY: test
 test: tester-progs tetragon-bpf ## Run Go tests.
-	$(GO) test -exec "$(SUDO)" -p 1 -parallel 1 $(GOFLAGS) -gcflags=$(GO_BUILD_GCFLAGS) -timeout $(GO_TEST_TIMEOUT) -failfast -cover ./pkg/... ./cmd/... ./operator/... ${EXTRA_TESTFLAGS}
+	$(GO) test -exec "$(SUDO)" -p 1 -parallel 1 \
+	$(GOFLAGS) -gcflags=$(GO_BUILD_GCFLAGS) -timeout $(GO_TEST_TIMEOUT) -failfast -cover \
+	./pkg/... ./cmd/... ./operator/... \
+	${EXTRA_TESTFLAGS}
 
 .PHONY: tester-progs
 tester-progs: ## Compile helper programs for unit testing.
@@ -285,7 +288,10 @@ alignchecker: ## Run alignchecker.
 
 .PHONY: bench
 bench: ## Run Go benchmarks.
-	$(GO) test -exec "$(SUDO)" -p 1 -parallel 1 -run ^$$ $(GOFLAGS) -gcflags=$(GO_BUILD_GCFLAGS) -timeout $(GO_TEST_TIMEOUT) -failfast -cover ./pkg/... ./cmd/... ./operator/... -bench=. ${EXTRA_TESTFLAGS}
+	$(GO) test -exec "$(SUDO)" -p 1 -parallel 1 -run ^$$ \
+	$(GOFLAGS) -gcflags=$(GO_BUILD_GCFLAGS) -timeout $(GO_TEST_TIMEOUT) -failfast -cover \
+	./pkg/... ./cmd/... ./operator/... \
+	-bench=. ${EXTRA_TESTFLAGS}
 
 TEST_COMPILE ?= ./...
 .PHONY: test-compile
@@ -335,7 +341,13 @@ e2e-test: image image-operator
 else
 e2e-test:
 endif
-	$(GO) list $(E2E_TESTS) | xargs -Ipkg $(GO) test $(GOFLAGS) -gcflags=$(GO_BUILD_GCFLAGS) -timeout $(E2E_TEST_TIMEOUT) -failfast -cover pkg ${EXTRA_TESTFLAGS} -fail-fast -tetragon.helm.set tetragon.image.override="$(E2E_AGENT)" -tetragon.helm.set tetragonOperator.image.override="$(E2E_OPERATOR)" -tetragon.helm.url="" -tetragon.helm.chart="$(realpath ./install/kubernetes/tetragon)" $(E2E_BTF_FLAGS)
+	$(GO) list $(E2E_TESTS) | xargs -Ipkg \
+	$(GO) test $(GOFLAGS) -gcflags=$(GO_BUILD_GCFLAGS) -timeout $(E2E_TEST_TIMEOUT) -failfast -cover pkg \
+	${EXTRA_TESTFLAGS} \
+	-tetragon.helm.set tetragon.image.override="$(E2E_AGENT)" \
+	-tetragon.helm.set tetragonOperator.image.override="$(E2E_OPERATOR)" \
+	-tetragon.helm.url="" -tetragon.helm.chart="$(realpath ./install/kubernetes/tetragon)" \
+	$(E2E_BTF_FLAGS)
 
 ##@ Development
 

tests/e2e/checker/rpcchecker.go

@@ -334,7 +334,7 @@ func (rc *RPCChecker) updateContextEventCheckers(ctx context.Context) context.Co
 func getExportDir(ctx context.Context) (string, error) {
 	exportDir, ok := ctx.Value(state.ExportDir).(string)
 	if !ok {
-		return "", fmt.Errorf("export dir has not been created. Call helpers.CreateExportDir() first")
+		return "", fmt.Errorf("export dir has not been created. Call runner.SetupExport() first")
 	}
 	return exportDir, nil
 }

tests/e2e/helpers/cluster.go

@@ -42,22 +42,14 @@ nodes:
 `
 
 var (
-	clusterName  string
 	clusterImage string
 )
 
 func init() {
-	flag.StringVar(&clusterName, "cluster-name", "tetragon-ci", "Set the name of the k8s cluster being used")
 	// renovate: datasource=docker
 	flag.StringVar(&clusterImage, "cluster-image", "kindest/node:v1.32.3", "Set the node image for the kind cluster")
 }
 
-// GetClusterName fetches the cluster name configured with -cluster-name or the temporary
-// kind cluster name.
-func GetClusterName() string {
-	return clusterName
-}
-
 func SetMinKernelVersion() env.Func {
 	return func(ctx context.Context, cfg *envconf.Config) (context.Context, error) {
 		client, err := cfg.NewClient()
@@ -141,7 +133,6 @@ func MaybeCreateTempKindCluster(testenv env.Environment, namePrefix string) env.
 	return func(ctx context.Context, cfg *envconf.Config) (context.Context, error) {
 		if cfg.KubeconfigFile() == "" {
 			name := envconf.RandomName(namePrefix, 16)
-			clusterName = name
 			klog.Infof("No kubeconfig specified, creating temporary kind cluster %s", name)
 			var err error
 			err = writeKindConfig()

tests/e2e/helpers/dumpinfo.go

@@ -29,6 +29,7 @@ import (
 
 var (
 	TetragonContainerName = "tetragon"
+	OperatorContainerName = "tetragon-operator"
 	TetragonJsonPathname  = "/var/run/cilium/tetragon/tetragon.log"
 )
 
@@ -52,63 +53,86 @@ func DumpInfo(ctx context.Context, cfg *envconf.Config) (context.Context, error)
 	if err != nil {
 		return ctx, err
 	}
-	r := client.Resources(opts.Namespace)
 
-	podList := &corev1.PodList{}
-	if err = r.List(
-		ctx,
-		podList,
-		resources.WithLabelSelector(fmt.Sprintf("app.kubernetes.io/instance=%s", opts.DaemonSetName)),
-	); err != nil {
+	if err := dumpPodSummary("pods.txt", exportDir); err != nil {
+		klog.ErrorS(err, "Failed to dump pod summary")
+	}
+	err = dumpAgentInfo(ctx, exportDir, opts, client)
+	if err != nil {
+		return ctx, err
+	}
+	err = dumpOperatorInfo(ctx, exportDir, opts, client)
+	if err != nil {
 		return ctx, err
 	}
 
+	return ctx, nil
+}
+
+func GetExportDir(ctx context.Context) (string, error) {
+	exportDir, ok := ctx.Value(state.ExportDir).(string)
+	if !ok {
+		return "", fmt.Errorf("export dir has not been created. Call runner.SetupExport() first")
+	}
+
+	return exportDir, nil
+}
+
+func dumpAgentInfo(ctx context.Context, exportDir string, opts *flags.HelmOptions, client klient.Client) error {
+	podList, err := getPods(ctx, opts.DaemonSetName, opts, client)
+	if err != nil {
+		return err
+	}
+
 	for _, pod := range podList.Items {
 		if err := extractJson(&pod, exportDir); err != nil {
 			klog.ErrorS(err, "Failed to extract json events")
 		}
-		if err := extractLogs(&pod, exportDir, true); err != nil {
+		if err := extractLogs(&pod, exportDir, TetragonContainerName, true); err != nil {
 			klog.ErrorS(err, "Failed to extract previous tetragon logs")
 		}
-		if err := extractLogs(&pod, exportDir, false); err != nil {
+		if err := extractLogs(&pod, exportDir, TetragonContainerName, false); err != nil {
 			klog.ErrorS(err, "Failed to extract tetragon logs")
 		}
-		if err := describeTetragonPod(&pod, exportDir); err != nil {
+		if err := describePod(&pod, opts.DaemonSetName, exportDir); err != nil {
 			klog.ErrorS(err, "Failed to describe tetragon pods")
 		}
-		if err := dumpPodSummary("pods.txt", exportDir); err != nil {
-			klog.ErrorS(err, "Failed to dump pod summary")
-		}
 		dumpBpftool(ctx, client, exportDir, pod.Namespace, pod.Name, TetragonContainerName)
 	}
-
-	return ctx, nil
+	return nil
 }
 
-func CreateExportDir(ctx context.Context, t *testing.T) (context.Context, error) {
-	dir, err := GetExportDir(ctx)
-	if err == nil {
-		klog.V(2).InfoS("export dir already exists, skipping creation", "test", t.Name(), "dir", dir)
-		return ctx, nil
-	}
-
-	dir, err = os.MkdirTemp("", fmt.Sprintf("tetragon.e2e.%s.*", t.Name()))
+func dumpOperatorInfo(ctx context.Context, exportDir string, opts *flags.HelmOptions, client klient.Client) error {
+	name := opts.DaemonSetName + "-operator"
+	podList, err := getPods(ctx, name, opts, client)
 	if err != nil {
-		return ctx, err
+		return err
 	}
 
-	klog.InfoS("created export dir for test", "test", t.Name(), "dir", dir)
-
-	return context.WithValue(ctx, state.ExportDir, dir), nil
+	for _, pod := range podList.Items {
+		if err := extractLogs(&pod, exportDir, OperatorContainerName, true); err != nil {
+			klog.ErrorS(err, "Failed to extract previous operator logs")
+		}
+		if err := extractLogs(&pod, exportDir, OperatorContainerName, false); err != nil {
+			klog.ErrorS(err, "Failed to extract operator logs")
+		}
+		if err := describePod(&pod, name, exportDir); err != nil {
+			klog.ErrorS(err, "Failed to describe operator pods")
+		}
+	}
+	return nil
 }
 
-func GetExportDir(ctx context.Context) (string, error) {
-	exportDir, ok := ctx.Value(state.ExportDir).(string)
-	if !ok {
-		return "", fmt.Errorf("export dir has not been created. Call runner.CreateExportDir() first")
+func getPods(ctx context.Context, nameLabel string, opts *flags.HelmOptions, client klient.Client) (*corev1.PodList, error) {
+	r := client.Resources(opts.Namespace)
+	podList := &corev1.PodList{}
+	if err := r.List(
+		ctx, podList,
+		resources.WithLabelSelector(fmt.Sprintf("app.kubernetes.io/name=%s", nameLabel)),
+	); err != nil {
+		return nil, err
 	}
-
-	return exportDir, nil
+	return podList, nil
 }
 
 func extractJson(pod *corev1.Pod, exportDir string) error {
@@ -119,22 +143,21 @@ func extractJson(pod *corev1.Pod, exportDir string) error {
 		filepath.Join(exportDir, fmt.Sprintf("tetragon.%s.json", pod.Name)))
 }
 
-func extractLogs(pod *corev1.Pod, exportDir string, prev bool) error {
+func extractLogs(pod *corev1.Pod, exportDir string, container string, prev bool) error {
 	var fname string
 	if prev {
-		fname = fmt.Sprintf("tetragon.%s.prev.log", pod.Name)
+		fname = fmt.Sprintf("%s.%s.prev.log", container, pod.Name)
 	} else {
-		fname = fmt.Sprintf("tetragon.%s.log", pod.Name)
+		fname = fmt.Sprintf("%s.%s.log", container, pod.Name)
 	}
-	return kubectlLogs(filepath.Join(exportDir, fname),
-		pod.Namespace,
-		pod.Name,
-		TetragonContainerName,
-		prev)
+	return kubectlLogs(
+		filepath.Join(exportDir, fname),
+		pod.Namespace, pod.Name, container, prev,
+	)
 }
 
-func describeTetragonPod(pod *corev1.Pod, exportDir string) error {
-	fname := fmt.Sprintf("tetragon.%s.describe", pod.Name)
+func describePod(pod *corev1.Pod, workload string, exportDir string) error {
+	fname := fmt.Sprintf("%s.%s.describe", workload, pod.Name)
 	return kubectlDescribe(filepath.Join(exportDir, fname),
 		pod.Namespace,
 		pod.Name)