Re-enabled notifyFleetAuditUninstall on Windows (#6257) (#6289)

* Write integration test.

* Skip notifyFleetAuditUninstall on Windows.

* Re-enable notifyFleetAuditUninstall on Windows.

* Update test to be in Fleet group.

* Adjust OGC supported.

* Add changelog.

* Update changelog.

* Add check for connected to Fleet.

* Remove TestRepeatedInstallUninstall.

(cherry picked from commit dcf5b7a)

Co-authored-by: Blake Rouse <blake.rouse@elastic.co>

changelog/fragments/1733768621-Notify-Fleet-of-uninstall-on-Windows.yaml

@@ -0,0 +1,32 @@
+# Kind can be one of:
+# - breaking-change: a change to previously-documented behavior
+# - deprecation: functionality that is being removed in a later release
+# - bug-fix: fixes a problem in a previous version
+# - enhancement: extends functionality but does not break or fix existing behavior
+# - feature: new functionality
+# - known-issue: problems that we are aware of in a given version
+# - security: impacts on the security of a product or a user’s deployment.
+# - upgrade: important information for someone upgrading from a prior version
+# - other: does not fit into any of the other categories
+kind: bug-fix
+
+# Change summary; a 80ish characters long description of the change.
+summary: Notify Fleet of uninstall on Windows
+
+# Long description; in case the summary is not enough to describe the change
+# this field accommodate a description without length limits.
+# NOTE: This field will be rendered only for breaking-change and known-issue kinds at the moment.
+#description:
+
+# Affected component; usually one of "elastic-agent", "fleet-server", "filebeat", "metricbeat", "auditbeat", "all", etc.
+component:
+
+# PR URL; optional; the PR number that added the changeset.
+# If not present is automatically filled by the tooling finding the PR where this changelog fragment has been added.
+# NOTE: the tooling supports backports, so it's able to fill the original PR number instead of the backport PR number.
+# Please provide it if you are adding a fragment for a different PR.
+pr: https://github.com/elastic/elastic-agent/pull/6257
+
+# Issue URL; optional; the GitHub issue related to this changeset (either closes or is part of).
+# If not present is automatically filled by the tooling with the issue linked to the PR number.
+issue: https://github.com/elastic/elastic-agent/issues/5952

internal/pkg/agent/install/uninstall.go

@@ -144,9 +144,7 @@ func Uninstall(ctx context.Context, cfgFile, topPath, uninstallToken string, log
 	}
 	pt.Describe("Removed install directory")
 
-	// Skip on Windows because of https://github.com/elastic/elastic-agent/issues/5952
-	// Once the root-cause is identified then this can be re-enabled on Windows.
-	if notifyFleet && runtime.GOOS != "windows" {
+	if notifyFleet {
 		notifyFleetAuditUninstall(ctx, log, pt, cfg, ai) //nolint:errcheck // ignore the error as we can't act on it
 	}
 

pkg/testing/fixture.go

@@ -53,6 +53,7 @@ type Fixture struct {
 
 	srcPackage string
 	workDir    string
+	extractDir string
 
 	installed   bool
 	installOpts *InstallOpts
@@ -198,7 +199,7 @@ func (f *Fixture) Prepare(ctx context.Context, components ...UsableComponent) er
 	if err != nil {
 		return err
 	}
-	if f.workDir != "" {
+	if f.extractDir != "" {
 		// already prepared
 		return fmt.Errorf("already been prepared")
 	}
@@ -212,16 +213,17 @@ func (f *Fixture) Prepare(ctx context.Context, components ...UsableComponent) er
 	if err != nil {
 		return err
 	}
-	workDir := createTempDir(f.t)
-	finalDir := filepath.Join(workDir, name)
-	err = ExtractArtifact(f.t, src, workDir)
+	extractDir := createTempDir(f.t)
+	finalDir := filepath.Join(extractDir, name)
+	err = ExtractArtifact(f.t, src, extractDir)
 	if err != nil {
-		return fmt.Errorf("extracting artifact %q in %q: %w", src, workDir, err)
+		return fmt.Errorf("extracting artifact %q in %q: %w", src, extractDir, err)
 	}
 	err = f.prepareComponents(finalDir, components...)
 	if err != nil {
 		return err
 	}
+	f.extractDir = finalDir
 	f.workDir = finalDir
 	return nil
 }
@@ -823,7 +825,7 @@ func (f *Fixture) IsInstalled() bool {
 
 // EnsurePrepared ensures that the fixture has been prepared.
 func (f *Fixture) EnsurePrepared(ctx context.Context) error {
-	if f.workDir == "" {
+	if f.extractDir == "" {
 		return f.Prepare(ctx)
 	}
 	return nil

pkg/testing/fixture_install.go

@@ -631,6 +631,7 @@ func (f *Fixture) uninstallNoPkgManager(ctx context.Context, uninstallOpts *Unin
 		return out, fmt.Errorf("error running uninstall command: %w", err)
 	}
 	f.installed = false
+	f.workDir = f.extractDir
 
 	// Check that Elastic Agent files are actually removed
 	basePath := f.installOpts.BasePath

testing/integration/install_test.go

@@ -27,6 +27,7 @@ import (
 	"github.com/elastic/elastic-agent/internal/pkg/agent/install"
 	atesting "github.com/elastic/elastic-agent/pkg/testing"
 	"github.com/elastic/elastic-agent/pkg/testing/define"
+	"github.com/elastic/elastic-agent/pkg/testing/tools/check"
 	"github.com/elastic/elastic-agent/pkg/testing/tools/fleettools"
 	"github.com/elastic/elastic-agent/pkg/testing/tools/testcontext"
 	"github.com/elastic/elastic-agent/testing/installtest"
@@ -420,13 +421,14 @@ func TestInstallUninstallAudit(t *testing.T) {
 	require.Equal(t, "uninstall", res.Source.AuditUnenrolledReason)
 }
 
-// TestRepeatedInstallUninstall will install then uninstall the agent
-// repeatedly.  This test exists because of a number of race
-// conditions that have occurred in the uninstall process.  Current
-// testing shows each iteration takes around 16 seconds.
-func TestRepeatedInstallUninstall(t *testing.T) {
-	define.Require(t, define.Requirements{
-		Group: Default,
+// TestRepeatedInstallUninstallFleet will install then uninstall the agent
+// repeatedly with it enrolled into Fleet.  This test exists because of a number
+// of race conditions that have occurred in the uninstall process when enrolled
+// into Fleet. Current testing shows each iteration takes around 16 seconds.
+func TestRepeatedInstallUninstallFleet(t *testing.T) {
+	info := define.Require(t, define.Requirements{
+		Group: Fleet,
+		Stack: &define.Stack{}, // needs a fleet-server.
 		// We require sudo for this test to run
 		// `elastic-agent install` (even though it will
 		// be installed as non-root).
@@ -437,37 +439,63 @@ func TestRepeatedInstallUninstall(t *testing.T) {
 		Local: false,
 	})
 
+	prepareCtx, prepareCancel := testcontext.WithDeadline(t, context.Background(), time.Now().Add(5*time.Minute))
+	defer prepareCancel()
+
+	policyResp, enrollmentTokenResp := createPolicyAndEnrollmentToken(prepareCtx, t, info.KibanaClient, createBasicPolicy())
+	t.Logf("Created policy %+v", policyResp.AgentPolicy)
+
+	t.Log("Getting default Fleet Server URL...")
+	fleetServerURL, err := fleettools.DefaultURL(prepareCtx, info.KibanaClient)
+	require.NoError(t, err, "failed getting Fleet Server URL")
+
+	// Get path to Elastic Agent executable
+	fixture, err := define.NewFixtureFromLocalBuild(t, define.Version())
+	require.NoError(t, err)
+
+	// Prepare the Elastic Agent so the binary is extracted and ready to use.
+	err = fixture.Prepare(prepareCtx)
+	require.NoError(t, err)
+
 	maxRunTime := 2 * time.Minute
 	iterations := 100
 	for i := 0; i < iterations; i++ {
-		t.Run(fmt.Sprintf("%s-%d", t.Name(), i), func(t *testing.T) {
-
-			// Get path to Elastic Agent executable
-			fixture, err := define.NewFixtureFromLocalBuild(t, define.Version())
-			require.NoError(t, err)
-
+		successful := t.Run(fmt.Sprintf("%s-%d", t.Name(), i), func(t *testing.T) {
 			ctx, cancel := testcontext.WithDeadline(t, context.Background(), time.Now().Add(maxRunTime))
 			defer cancel()
 
-			// Prepare the Elastic Agent so the binary is extracted and ready to use.
-			err = fixture.Prepare(ctx)
-			require.NoError(t, err)
-
 			// Run `elastic-agent install`.  We use `--force` to prevent interactive
 			// execution.
-			opts := &atesting.InstallOpts{Force: true}
+			opts := &atesting.InstallOpts{
+				Force: true,
+				EnrollOpts: atesting.EnrollOpts{
+					URL:             fleetServerURL,
+					EnrollmentToken: enrollmentTokenResp.APIKey,
+				},
+			}
 			out, err := fixture.Install(ctx, opts)
 			if err != nil {
 				t.Logf("install output: %s", out)
-				require.NoError(t, err)
+				require.NoErrorf(t, err, "install failed: %s", err)
 			}
 
-			// Check that Agent was installed in default base path
+			// Check that Agent was installed in successfully
 			require.NoError(t, installtest.CheckSuccess(ctx, fixture, opts.BasePath, &installtest.CheckOpts{Privileged: opts.Privileged}))
-			t.Run("check agent package version", testAgentPackageVersion(ctx, fixture, true))
+
+			// Check connected to Fleet.
+			check.ConnectedToFleet(ctx, t, fixture, 5*time.Minute)
+
+			// Perform uninstall.
 			out, err = fixture.Uninstall(ctx, &atesting.UninstallOpts{Force: true})
-			require.NoErrorf(t, err, "uninstall failed: %s", err)
+			if err != nil {
+				t.Logf("uninstall output: %s", out)
+				require.NoErrorf(t, err, "uninstall failed: %s", err)
+			}
 		})
+		if !successful {
+			// quit now, another test run will continue to fail now
+			return
+		}
 	}
 }