🐛 Unblock e2e (#1683)

* Remove ginkgo from bingo in favor of go.mod version

Signed-off-by: Per Goncalves da Silva <[email protected]>

* fix(e2e): wait for leader election

TestClusterExtensionAfterOLMUpgrade was failing due to increased leader
election timeouts, causing reconciliation checks to run before leadership
was acquired.

This fix ensures the test explicitly waits for leader election logs
(`"successfully acquired lease"`) before verifying reconciliation.

* Add leader election wait fit for catalogd upgrade-e2e

Signed-off-by: Per Goncalves da Silva <[email protected]>

---------

Signed-off-by: Per Goncalves da Silva <[email protected]>
Co-authored-by: Per Goncalves da Silva <[email protected]>
Co-authored-by: Camila Macedo <[email protected]>

Author: 3 people

.bingo/Variables.mk

@@ -41,12 +41,6 @@ $(CRD_REF_DOCS): $(BINGO_DIR)/crd-ref-docs.mod
 	@echo "(re)installing $(GOBIN)/crd-ref-docs-v0.1.0"
 	@cd $(BINGO_DIR) && GOWORK=off $(GO) build -mod=mod -modfile=crd-ref-docs.mod -o=$(GOBIN)/crd-ref-docs-v0.1.0 "github.com/elastic/crd-ref-docs"
 
-GINKGO := $(GOBIN)/ginkgo-v2.22.2
-$(GINKGO): $(BINGO_DIR)/ginkgo.mod
-	@# Install binary/ries using Go 1.14+ build command. This is using bwplotka/bingo-controlled, separate go module with pinned dependencies.
-	@echo "(re)installing $(GOBIN)/ginkgo-v2.22.2"
-	@cd $(BINGO_DIR) && GOWORK=off $(GO) build -mod=mod -modfile=ginkgo.mod -o=$(GOBIN)/ginkgo-v2.22.2 "github.com/onsi/ginkgo/v2/ginkgo"
-
 GOLANGCI_LINT := $(GOBIN)/golangci-lint-v1.63.4
 $(GOLANGCI_LINT): $(BINGO_DIR)/golangci-lint.mod
 	@# Install binary/ries using Go 1.14+ build command. This is using bwplotka/bingo-controlled, separate go module with pinned dependencies.

.bingo/ginkgo.mod

@@ -16,8 +16,6 @@ CRD_DIFF="${GOBIN}/crd-diff-v0.1.0"
 
 CRD_REF_DOCS="${GOBIN}/crd-ref-docs-v0.1.0"
 
-GINKGO="${GOBIN}/ginkgo-v2.22.2"
-
 GOLANGCI_LINT="${GOBIN}/golangci-lint-v1.63.4"
 
 GORELEASER="${GOBIN}/goreleaser-v1.26.2"

.bingo/ginkgo.sum

@@ -44,6 +44,8 @@ TESTDATA_DIR := testdata
 CATALOGD_NAMESPACE := olmv1-system
 KIND_CLUSTER_IMAGE := kindest/node:v1.30.0@sha256:047357ac0cfea04663786a612ba1eaba9702bef25227a794b52890dd8bcd692e
 
+GINKGO := go run github.com/onsi/ginkgo/v2/ginkgo
+
 ##@ General
 
 # The help target prints out all targets with their descriptions organized
@@ -76,7 +78,7 @@ FOCUS := $(if $(TEST),-v -focus "$(TEST)")
 ifeq ($(origin E2E_FLAGS), undefined)
 E2E_FLAGS :=
 endif
-test-e2e: $(GINKGO) ## Run the e2e tests on existing cluster
+test-e2e: ## Run the e2e tests on existing cluster
 	$(GINKGO) $(E2E_FLAGS) -trace -vv $(FOCUS) test/e2e
 
 e2e: KIND_CLUSTER_NAME := catalogd-e2e
@@ -105,7 +107,7 @@ run-latest-release:
 	cd ..; curl -L -s https://github.com/operator-framework/operator-controller/releases/latest/download/install.sh | bash -s
 
 .PHONY: post-upgrade-checks
-post-upgrade-checks: $(GINKGO)
+post-upgrade-checks:
 	$(GINKGO) $(E2E_FLAGS) -trace -vv $(FOCUS) test/upgrade
 
 ##@ Build

.bingo/variables.env

@@ -54,6 +54,17 @@ var _ = Describe("ClusterCatalog Unpacking", func() {
 				managerPod = managerPods.Items[0]
 			}).Should(Succeed())
 
+			By("Waiting for acquired leader election")
+			// Average case is under 1 minute but in the worst case: (previous leader crashed)
+			// we could have LeaseDuration (137s) + RetryPeriod (26s) +/- 163s
+			leaderCtx, leaderCancel := context.WithTimeout(ctx, 3*time.Minute)
+			defer leaderCancel()
+
+			leaderSubstrings := []string{"successfully acquired lease"}
+			leaderElected, err := watchPodLogsForSubstring(leaderCtx, &managerPod, "manager", leaderSubstrings...)
+			Expect(err).To(Succeed())
+			Expect(leaderElected).To(BeTrue())
+
 			By("Reading logs to make sure that ClusterCatalog was reconciled by catalogdv1")
 			logCtx, cancel := context.WithTimeout(ctx, time.Minute)
 			defer cancel()

catalogd/Makefile

@@ -40,6 +40,17 @@ func TestClusterExtensionAfterOLMUpgrade(t *testing.T) {
 	t.Log("Wait for operator-controller deployment to be ready")
 	managerPod := waitForDeployment(t, ctx, "operator-controller-controller-manager")
 
+	t.Log("Wait for acquired leader election")
+	// Average case is under 1 minute but in the worst case: (previous leader crashed)
+	// we could have LeaseDuration (137s) + RetryPeriod (26s) +/- 163s
+	leaderCtx, leaderCancel := context.WithTimeout(ctx, 3*time.Minute)
+	defer leaderCancel()
+
+	leaderSubstrings := []string{"successfully acquired lease"}
+	leaderElected, err := watchPodLogsForSubstring(leaderCtx, managerPod, "manager", leaderSubstrings...)
+	require.NoError(t, err)
+	require.True(t, leaderElected)
+
 	t.Log("Reading logs to make sure that ClusterExtension was reconciled by operator-controller before we update it")
 	// Make sure that after we upgrade OLM itself we can still reconcile old objects without any changes
 	logCtx, cancel := context.WithTimeout(ctx, time.Minute)