feat: HA-Cluster v1.2.x — Split-Brain, TOTP, Enterprise-FW, Drift-Fix, VIP-Recovery

- keepalived: pg_role='standby' hat Vorrang vor role für BACKUP-Bestimmung
- keepalived-master.sh: gecrasht Dienste beim MASTER-Übergang starten (nicht nur reload)
- confighash: ip_addresses per Interface-Name hashen statt per FK (Cross-Node-Drift-Fix)
- TOTP/2FA: RFC 6238 — Setup-Flow, QR-Code, Admin-Disable; two-step Login
- Firewall-UI: Enterprise-Design — auto-Beschreibung, icon-only Actions, zero-hit Indikator
- fe80-Filter: Link-local IPv6 aus NTP/DNS Listen-Dropdowns entfernen
- VIP-Dashboard, Dual-Path VRRP, GW-Tracking (Migrations 0033/0034)
- Forward Proxy + DNS erweiterte Einstellungen (Migrations 0031/0032)
- unbound-control: edgeguard in unbound-Gruppe via postinst

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>

internal/handlers/auth.go

@@ -60,15 +60,22 @@ func (h *AuthHandler) WithClusterTLS(store *clustertls.Store) *AuthHandler {
 	return h
 }
 
+const totpPendingCookie = "edgeguard_totp_pending"
+
 // Register mounts /auth/login + /logout (public) and /auth/me
 // (gated by requireAuth, passed in as a per-route middleware).
 func (h *AuthHandler) Register(rg *gin.RouterGroup, requireAuth gin.HandlerFunc) {
 	g := rg.Group("/auth")
 	g.POST("/login", h.Login)
 	g.POST("/logout", h.Logout)
+	g.POST("/totp-verify", h.TOTPVerify)
 	g.GET("/me", requireAuth, h.Me)
 	g.POST("/reset-password", h.ResetPassword)
 	g.POST("/change-password", requireAuth, h.ChangePassword)
+	// TOTP self-service (authenticated user manages own 2FA)
+	g.POST("/totp/setup", requireAuth, h.TOTPSetup)
+	g.POST("/totp/confirm", requireAuth, h.TOTPConfirm)
+	g.DELETE("/totp", requireAuth, h.TOTPDisable)
 }
 
 type loginRequest struct {
@@ -77,9 +84,10 @@ type loginRequest struct {
 }
 
 type loginResponse struct {
-	Actor     string    `json:"actor"`
-	Role      string    `json:"role"`
-	ExpiresAt time.Time `json:"expires_at"`
+	Actor        string    `json:"actor"`
+	Role         string    `json:"role"`
+	ExpiresAt    time.Time `json:"expires_at"`
+	TOTPRequired bool      `json:"totp_required,omitempty"`
 }
 
 func (h *AuthHandler) Login(c *gin.Context) {
@@ -101,12 +109,13 @@ func (h *AuthHandler) Login(c *gin.Context) {
 	email := strings.TrimSpace(req.Email)
 	actor, role := "", "admin"
 	remote := c.ClientIP()
+	var totpEnabled bool
 
 	// 1. Try DB users table first.
 	if h.Users != nil {
-		u, hash, dbErr := h.Users.FindByEmail(c.Request.Context(), email)
+		ai, dbErr := h.Users.FindForAuth(c.Request.Context(), email)
 		if dbErr == nil {
-			if !u.Active {
+			if !ai.Active {
 				if h.Audit != nil {
 					_ = h.Audit.Log(c.Request.Context(), email, "auth.login.failed",
 						email, gin.H{"reason": "account_disabled", "remote": remote}, h.NodeID)
@@ -114,7 +123,7 @@ func (h *AuthHandler) Login(c *gin.Context) {
 				response.Unauthorized(c, errors.New("account_disabled"))
 				return
 			}
-			if !usersvc.VerifyPassword(hash, req.Password) {
+			if !usersvc.VerifyPassword(ai.PasswordHash, req.Password) {
 				if h.Audit != nil {
 					_ = h.Audit.Log(c.Request.Context(), email, "auth.login.failed",
 						email, gin.H{"reason": "invalid_credentials", "remote": remote}, h.NodeID)
@@ -122,9 +131,10 @@ func (h *AuthHandler) Login(c *gin.Context) {
 				response.Unauthorized(c, errors.New("invalid_credentials"))
 				return
 			}
-			actor = u.Email
-			role = u.Role
-			h.Users.RecordLogin(c.Request.Context(), u.ID)
+			actor = ai.Email
+			role = ai.Role
+			totpEnabled = ai.TOTPEnabled
+			h.Users.RecordLogin(c.Request.Context(), ai.ID)
 		}
 	}
 
@@ -133,16 +143,13 @@ func (h *AuthHandler) Login(c *gin.Context) {
 		if strings.EqualFold(st.AdminEmail, email) && st.VerifyAdminPassword(req.Password) {
 			actor = st.AdminEmail
 			role = "admin"
-			// Auto-migrate: insert the setup-store admin into the DB so it
-			// shows up in user management from this point on.
 			if h.Users != nil {
 				_, _ = h.Users.Upsert(c.Request.Context(), st.AdminEmail, req.Password, "admin", true)
 			}
 		}
 	}
 
-	// 3. Auth federation: cluster nodes forward failed auth to the primary
-	// via mTLS so users can log in with their primary credentials on any node.
+	// 3. Auth federation: cluster nodes forward failed auth to the primary.
 	if actor == "" && st.IsClusterNode && st.PrimaryFQDN != "" && h.ClusterTLS != nil {
 		if a, r, err := h.checkWithPrimary(c.Request.Context(), st.PrimaryFQDN, email, req.Password); err == nil {
 			actor = a
@@ -161,6 +168,21 @@ func (h *AuthHandler) Login(c *gin.Context) {
 		return
 	}
 
+	// TOTP gate: password OK but 2FA required → issue a short-lived pending
+	// cookie and tell the UI to show the TOTP input.
+	if totpEnabled {
+		pending, ptok, err := h.Signer.IssueWithRoleTTL(actor, "totp_pending", 2*time.Minute)
+		if err != nil {
+			response.Internal(c, err)
+			return
+		}
+		c.SetSameSite(http.SameSiteStrictMode)
+		c.SetCookie(totpPendingCookie, pending, int(2*time.Minute/time.Second), "/", "", true, true)
+		_ = ptok
+		response.OK(c, loginResponse{TOTPRequired: true})
+		return
+	}
+
 	raw, tok, err := h.Signer.IssueWithRole(actor, role)
 	if err != nil {
 		response.Internal(c, err)
@@ -179,6 +201,146 @@ func (h *AuthHandler) Login(c *gin.Context) {
 	})
 }
 
+type totpVerifyRequest struct {
+	Code string `json:"code" binding:"required"`
+}
+
+// TOTPVerify completes the two-step login: verifies the TOTP code from the
+// pending cookie and, on success, issues a full session JWT.
+func (h *AuthHandler) TOTPVerify(c *gin.Context) {
+	var req totpVerifyRequest
+	if err := c.ShouldBindJSON(&req); err != nil {
+		response.BadRequest(c, err)
+		return
+	}
+	pendingRaw, err := c.Cookie(totpPendingCookie)
+	if err != nil || pendingRaw == "" {
+		response.Unauthorized(c, errors.New("no_pending_totp"))
+		return
+	}
+	ptok, err := h.Signer.Verify(pendingRaw)
+	if err != nil || ptok.Role != "totp_pending" {
+		response.Unauthorized(c, errors.New("invalid_pending_token"))
+		return
+	}
+
+	if h.Users == nil {
+		response.Internal(c, errors.New("users repo unavailable"))
+		return
+	}
+	ai, err := h.Users.FindForAuth(c.Request.Context(), ptok.Actor)
+	if err != nil || !ai.TOTPEnabled || ai.TOTPSecret == nil {
+		response.Unauthorized(c, errors.New("totp_not_configured"))
+		return
+	}
+	if !usersvc.VerifyTOTP(*ai.TOTPSecret, req.Code) {
+		if h.Audit != nil {
+			_ = h.Audit.Log(c.Request.Context(), ptok.Actor, "auth.totp.failed",
+				ptok.Actor, gin.H{"remote": c.ClientIP()}, h.NodeID)
+		}
+		response.Unauthorized(c, errors.New("invalid_totp_code"))
+		return
+	}
+
+	// Clear pending cookie, issue full session.
+	c.SetSameSite(http.SameSiteStrictMode)
+	c.SetCookie(totpPendingCookie, "", -1, "/", "", true, true)
+
+	raw, tok, err := h.Signer.IssueWithRole(ptok.Actor, ai.Role)
+	if err != nil {
+		response.Internal(c, err)
+		return
+	}
+	setSessionCookie(c, raw, tok.Exp)
+
+	if h.Audit != nil {
+		_ = h.Audit.Log(c.Request.Context(), ptok.Actor, "auth.login.success",
+			ptok.Actor, gin.H{"role": ai.Role, "remote": c.ClientIP(), "totp": true}, h.NodeID)
+	}
+	response.OK(c, loginResponse{
+		Actor:     tok.Actor,
+		Role:      tok.Role,
+		ExpiresAt: time.Unix(tok.Exp, 0).UTC(),
+	})
+}
+
+// TOTPSetup generates a new TOTP secret for the authenticated user and returns
+// the provisioning URI (renders as QR code in the UI). Secret is not saved yet.
+func (h *AuthHandler) TOTPSetup(c *gin.Context) {
+	tok := CurrentToken(c)
+	if tok == nil {
+		response.Unauthorized(c, nil)
+		return
+	}
+	secret, uri, err := usersvc.GenerateTOTPSecret(tok.Actor)
+	if err != nil {
+		response.Internal(c, err)
+		return
+	}
+	response.OK(c, gin.H{"secret": secret, "uri": uri})
+}
+
+type totpConfirmRequest struct {
+	Secret string `json:"secret" binding:"required"`
+	Code   string `json:"code"   binding:"required"`
+}
+
+// TOTPConfirm verifies the code against the provisioned secret and, on success,
+// enables TOTP for the user.
+func (h *AuthHandler) TOTPConfirm(c *gin.Context) {
+	var req totpConfirmRequest
+	if err := c.ShouldBindJSON(&req); err != nil {
+		response.BadRequest(c, err)
+		return
+	}
+	tok := CurrentToken(c)
+	if tok == nil || h.Users == nil {
+		response.Unauthorized(c, nil)
+		return
+	}
+	u, _, err := h.Users.FindByEmail(c.Request.Context(), tok.Actor)
+	if err != nil {
+		response.Internal(c, err)
+		return
+	}
+	if err := h.Users.ConfirmTOTP(c.Request.Context(), u.ID, req.Secret, req.Code); err != nil {
+		if err.Error() == "invalid_totp_code" {
+			response.Err(c, http.StatusUnprocessableEntity, err)
+			return
+		}
+		response.Internal(c, err)
+		return
+	}
+	if h.Audit != nil {
+		_ = h.Audit.Log(c.Request.Context(), tok.Actor, "auth.totp.enabled",
+			tok.Actor, nil, h.NodeID)
+	}
+	response.OK(c, gin.H{"ok": true})
+}
+
+// TOTPDisable disables TOTP for the authenticated user.
+func (h *AuthHandler) TOTPDisable(c *gin.Context) {
+	tok := CurrentToken(c)
+	if tok == nil || h.Users == nil {
+		response.Unauthorized(c, nil)
+		return
+	}
+	u, _, err := h.Users.FindByEmail(c.Request.Context(), tok.Actor)
+	if err != nil {
+		response.Internal(c, err)
+		return
+	}
+	if err := h.Users.DisableTOTP(c.Request.Context(), u.ID); err != nil {
+		response.Internal(c, err)
+		return
+	}
+	if h.Audit != nil {
+		_ = h.Audit.Log(c.Request.Context(), tok.Actor, "auth.totp.disabled",
+			tok.Actor, nil, h.NodeID)
+	}
+	response.OK(c, gin.H{"ok": true})
+}
+
 func (h *AuthHandler) Logout(c *gin.Context) {
 	clearSessionCookie(c)
 	response.OK(c, gin.H{"logged_out": true})