Fix bugs from BUGHUNT.md: restore race conditions, infra backup, DR wiring, docker-setup.sh, restore.html

controller/internal/web/handler_restore.go

@@ -11,21 +11,32 @@ import (
 
 // restorePageHandler renders the full-page DR restore UI.
 func (s *Server) restorePageHandler(w http.ResponseWriter, r *http.Request) {
-	if s.restorePlan == nil {
+	s.restoreMu.RLock()
+	plan := s.restorePlan
+	if plan == nil {
+		s.restoreMu.RUnlock()
 		http.Redirect(w, r, "/", http.StatusFound)
 		return
 	}
+	// Snapshot all needed fields under lock before rendering
+	customerID := plan.CustomerID
+	timestamp := plan.Timestamp
+	apps := plan.GetApps()
+	drives := make([]backup.DriveInfo, len(plan.Drives))
+	copy(drives, plan.Drives)
+	status := plan.GetStatus()
+	s.restoreMu.RUnlock()
 
 	data := map[string]interface{}{
 		"Title":        "Katasztrófa utáni visszaállítás",
 		"CustomerName": s.cfg.Customer.Name,
 		"Domain":       s.cfg.Customer.Domain,
 		"Version":      s.version,
-		"CustomerID":   s.restorePlan.CustomerID,
-		"Timestamp":    s.restorePlan.Timestamp,
-		"Apps":         s.restorePlan.GetApps(),
-		"Drives":       s.restorePlan.Drives,
-		"PlanStatus":   s.restorePlan.Status,
+		"CustomerID":   customerID,
+		"Timestamp":    timestamp,
+		"Apps":         apps,
+		"Drives":       drives,
+		"PlanStatus":   status,
 	}
 
 	s.render(w, "restore", data)
@@ -33,27 +44,33 @@ func (s *Server) restorePageHandler(w http.ResponseWriter, r *http.Request) {
 
 // apiRestoreStatus returns the current restore plan status as JSON.
 func (s *Server) apiRestoreStatus(w http.ResponseWriter, r *http.Request) {
-	if s.restorePlan == nil {
+	s.restoreMu.RLock()
+	plan := s.restorePlan
+	if plan == nil {
+		s.restoreMu.RUnlock()
 		jsonError(w, "not in restore mode", http.StatusBadRequest)
 		return
 	}
+	snapshot := plan.Snapshot()
+	s.restoreMu.RUnlock()
 
 	w.Header().Set("Content-Type", "application/json; charset=utf-8")
-	json.NewEncoder(w).Encode(s.restorePlan.Snapshot())
+	json.NewEncoder(w).Encode(snapshot)
 }
 
 // apiRestoreAll starts restoring all pending apps sequentially.
 func (s *Server) apiRestoreAll(w http.ResponseWriter, r *http.Request) {
-	if s.restorePlan == nil {
+	s.restoreMu.RLock()
+	plan := s.restorePlan
+	s.restoreMu.RUnlock()
+	if plan == nil {
 		jsonError(w, "not in restore mode", http.StatusBadRequest)
 		return
 	}
-	if s.restorePlan.Status == "restoring" {
+	if !plan.TryStartRestore() {
 		jsonError(w, "restore already in progress", http.StatusConflict)
 		return
 	}
-
-	s.restorePlan.Status = "restoring"
 	go s.executeAllRestores()
 
 	jsonResponse(w, map[string]interface{}{
@@ -64,7 +81,10 @@ func (s *Server) apiRestoreAll(w http.ResponseWriter, r *http.Request) {
 
 // apiRestoreSkip exits restore mode without restoring.
 func (s *Server) apiRestoreSkip(w http.ResponseWriter, r *http.Request) {
-	if s.restorePlan == nil {
+	s.restoreMu.RLock()
+	plan := s.restorePlan
+	s.restoreMu.RUnlock()
+	if plan == nil {
 		jsonError(w, "not in restore mode", http.StatusBadRequest)
 		return
 	}
@@ -82,13 +102,21 @@ func (s *Server) apiRestoreSkip(w http.ResponseWriter, r *http.Request) {
 func (s *Server) executeAllRestores() {
 	s.logger.Println("[INFO] Starting DR restore for all apps")
 
-	for i := range s.restorePlan.Apps {
-		app := &s.restorePlan.Apps[i]
+	s.restoreMu.RLock()
+	plan := s.restorePlan
+	s.restoreMu.RUnlock()
+	if plan == nil {
+		s.logger.Println("[WARN] Restore plan cleared before execution could start")
+		return
+	}
+
+	for i := range plan.Apps {
+		app := &plan.Apps[i]
 		if app.Status != "pending" {
 			continue
 		}
 
-		s.restorePlan.UpdateApp(app.Name, "restoring", "")
+		plan.UpdateApp(app.Name, "restoring", "")
 		s.logger.Printf("[INFO] Restoring app %s (%s)", app.Name, app.DisplayName)
 
 		ctx, cancel := context.WithTimeout(context.Background(), 10*time.Minute)
@@ -96,15 +124,15 @@ func (s *Server) executeAllRestores() {
 		cancel()
 
 		if err != nil {
-			s.restorePlan.UpdateApp(app.Name, "failed", err.Error())
+			plan.UpdateApp(app.Name, "failed", err.Error())
 			s.logger.Printf("[ERROR] Restore failed for %s: %v", app.Name, err)
 		} else {
-			s.restorePlan.UpdateApp(app.Name, "done", "")
+			plan.UpdateApp(app.Name, "done", "")
 			s.logger.Printf("[INFO] Restore completed for %s", app.Name)
 		}
 	}
 
-	s.restorePlan.Status = "done"
+	plan.SetStatus("done")
 	s.logger.Println("[INFO] All app restores completed")
 
 	// Re-scan stacks so dashboard picks up restored apps
@@ -113,15 +141,6 @@ func (s *Server) executeAllRestores() {
 			s.logger.Printf("[WARN] Post-restore stack scan failed: %v", err)
 		}
 	}
-
-	// Auto-clear restore mode after a brief delay so user can see final status
-	go func() {
-		time.Sleep(5 * time.Second)
-		// Only auto-clear if user hasn't already navigated away
-		if s.restorePlan != nil && s.restorePlan.AllDone() {
-			// Keep plan visible — user clicks "continue to dashboard" to clear
-		}
-	}()
 }
 
 // clearRestoreMode exits restore mode and returns to normal operation.