fix: max_locks_per_transaction requires PostgreSQL restart

- Fixed critical bug where ALTER SYSTEM + pg_reload_conf() was used
  but max_locks_per_transaction requires a full PostgreSQL restart
- Added automatic restart attempt (systemctl, service, pg_ctl)
- Added loud warnings if restart fails with manual fix instructions
- Updated preflight checks to warn about low max_locks_per_transaction
- This was causing 'out of shared memory' errors on BLOB-heavy restores

bin/README.md

@@ -4,8 +4,8 @@ This directory contains pre-compiled binaries for the DB Backup Tool across mult
 
 ## Build Information
 - **Version**: 3.42.34
-- **Build Time**: 2026-01-14_16:19:00_UTC
-- **Git Commit**: 7711a20
+- **Build Time**: 2026-01-15_14:33:12_UTC
+- **Git Commit**: 09a9177
 
 ## Recent Updates (v1.1.0)
 - ✅ Fixed TUI progress display with line-by-line output

internal/backup/engine.go

@@ -28,14 +28,22 @@ import (
 	"dbbackup/internal/swap"
 )
 
+// ProgressCallback is called with byte-level progress updates during backup operations
+type ProgressCallback func(current, total int64, description string)
+
+// DatabaseProgressCallback is called with database count progress during cluster backup
+type DatabaseProgressCallback func(done, total int, dbName string)
+
 // Engine handles backup operations
 type Engine struct {
-	cfg              *config.Config
-	log              logger.Logger
-	db               database.Database
-	progress         progress.Indicator
-	detailedReporter *progress.DetailedReporter
-	silent           bool // Silent mode for TUI
+	cfg                *config.Config
+	log                logger.Logger
+	db                 database.Database
+	progress           progress.Indicator
+	detailedReporter   *progress.DetailedReporter
+	silent             bool // Silent mode for TUI
+	progressCallback   ProgressCallback
+	dbProgressCallback DatabaseProgressCallback
 }
 
 // New creates a new backup engine
@@ -86,6 +94,30 @@ func NewSilent(cfg *config.Config, log logger.Logger, db database.Database, prog
 	}
 }
 
+// SetProgressCallback sets a callback for detailed progress reporting (for TUI mode)
+func (e *Engine) SetProgressCallback(cb ProgressCallback) {
+	e.progressCallback = cb
+}
+
+// SetDatabaseProgressCallback sets a callback for database count progress during cluster backup
+func (e *Engine) SetDatabaseProgressCallback(cb DatabaseProgressCallback) {
+	e.dbProgressCallback = cb
+}
+
+// reportProgress reports progress to the callback if set
+func (e *Engine) reportProgress(current, total int64, description string) {
+	if e.progressCallback != nil {
+		e.progressCallback(current, total, description)
+	}
+}
+
+// reportDatabaseProgress reports database count progress to the callback if set
+func (e *Engine) reportDatabaseProgress(done, total int, dbName string) {
+	if e.dbProgressCallback != nil {
+		e.dbProgressCallback(done, total, dbName)
+	}
+}
+
 // loggerAdapter adapts our logger to the progress.Logger interface
 type loggerAdapter struct {
 	logger logger.Logger
@@ -465,6 +497,8 @@ func (e *Engine) BackupCluster(ctx context.Context) error {
 			estimator.UpdateProgress(idx)
 			e.printf("   [%d/%d] Backing up database: %s\n", idx+1, len(databases), name)
 			quietProgress.Update(fmt.Sprintf("Backing up database %d/%d: %s", idx+1, len(databases), name))
+			// Report database progress to TUI callback
+			e.reportDatabaseProgress(idx+1, len(databases), name)
 			mu.Unlock()
 
 			// Check database size and warn if very large

internal/restore/engine.go

@@ -1937,6 +1937,8 @@ type OriginalSettings struct {
 }
 
 // boostPostgreSQLSettings boosts multiple PostgreSQL settings for large restores
+// NOTE: max_locks_per_transaction requires a PostgreSQL RESTART to take effect!
+// maintenance_work_mem can be changed with pg_reload_conf().
 func (e *Engine) boostPostgreSQLSettings(ctx context.Context, lockBoostValue int) (*OriginalSettings, error) {
 	connStr := e.buildConnString()
 	db, err := sql.Open("pgx", connStr)
@@ -1956,30 +1958,113 @@ func (e *Engine) boostPostgreSQLSettings(ctx context.Context, lockBoostValue int
 	// Get current maintenance_work_mem
 	db.QueryRowContext(ctx, "SHOW maintenance_work_mem").Scan(&original.MaintenanceWorkMem)
 
-	// Boost max_locks_per_transaction (if not already high enough)
+	// CRITICAL: max_locks_per_transaction requires a PostgreSQL RESTART!
+	// pg_reload_conf() is NOT sufficient for this parameter.
+	needsRestart := false
 	if original.MaxLocks < lockBoostValue {
 		_, err = db.ExecContext(ctx, fmt.Sprintf("ALTER SYSTEM SET max_locks_per_transaction = %d", lockBoostValue))
 		if err != nil {
-			e.log.Warn("Could not boost max_locks_per_transaction", "error", err)
+			e.log.Warn("Could not set max_locks_per_transaction", "error", err)
+		} else {
+			needsRestart = true
+			e.log.Warn("max_locks_per_transaction requires PostgreSQL restart to take effect",
+				"current", original.MaxLocks,
+				"target", lockBoostValue)
 		}
 	}
 
 	// Boost maintenance_work_mem to 2GB for faster index creation
+	// (this one CAN be applied via pg_reload_conf)
 	_, err = db.ExecContext(ctx, "ALTER SYSTEM SET maintenance_work_mem = '2GB'")
 	if err != nil {
 		e.log.Warn("Could not boost maintenance_work_mem", "error", err)
 	}
 
-	// Reload config to apply changes (no restart needed for these settings)
+	// Reload config to apply maintenance_work_mem
 	_, err = db.ExecContext(ctx, "SELECT pg_reload_conf()")
 	if err != nil {
 		return original, fmt.Errorf("failed to reload config: %w", err)
 	}
 
+	// If max_locks_per_transaction needs a restart, try to do it
+	if needsRestart {
+		if restarted := e.tryRestartPostgreSQL(ctx); restarted {
+			e.log.Info("PostgreSQL restarted successfully - max_locks_per_transaction now active")
+			// Wait for PostgreSQL to be ready
+			time.Sleep(3 * time.Second)
+		} else {
+			// Cannot restart - warn user loudly
+			e.log.Error("=" + strings.Repeat("=", 70))
+			e.log.Error("WARNING: max_locks_per_transaction change requires PostgreSQL restart!")
+			e.log.Error("Current value: " + strconv.Itoa(original.MaxLocks) + ", needed: " + strconv.Itoa(lockBoostValue))
+			e.log.Error("Restore may fail with 'out of shared memory' error on BLOB-heavy databases.")
+			e.log.Error("")
+			e.log.Error("To fix manually:")
+			e.log.Error("  1. sudo systemctl restart postgresql")
+			e.log.Error("  2. Or: sudo -u postgres pg_ctl restart -D $PGDATA")
+			e.log.Error("  3. Then re-run the restore")
+			e.log.Error("=" + strings.Repeat("=", 70))
+			// Continue anyway - might work for small restores
+		}
+	}
+
 	return original, nil
 }
 
+// tryRestartPostgreSQL attempts to restart PostgreSQL using various methods
+// Returns true if restart was successful
+func (e *Engine) tryRestartPostgreSQL(ctx context.Context) bool {
+	e.progress.Update("Attempting PostgreSQL restart for lock settings...")
+
+	// Method 1: systemctl (most common on modern Linux)
+	cmd := exec.CommandContext(ctx, "sudo", "systemctl", "restart", "postgresql")
+	if err := cmd.Run(); err == nil {
+		return true
+	}
+
+	// Method 2: systemctl with version suffix (e.g., postgresql-15)
+	for _, ver := range []string{"17", "16", "15", "14", "13", "12"} {
+		cmd = exec.CommandContext(ctx, "sudo", "systemctl", "restart", "postgresql-"+ver)
+		if err := cmd.Run(); err == nil {
+			return true
+		}
+	}
+
+	// Method 3: service command (older systems)
+	cmd = exec.CommandContext(ctx, "sudo", "service", "postgresql", "restart")
+	if err := cmd.Run(); err == nil {
+		return true
+	}
+
+	// Method 4: pg_ctl as postgres user
+	cmd = exec.CommandContext(ctx, "sudo", "-u", "postgres", "pg_ctl", "restart", "-D", "/var/lib/postgresql/data", "-m", "fast")
+	if err := cmd.Run(); err == nil {
+		return true
+	}
+
+	// Method 5: Try common PGDATA paths
+	pgdataPaths := []string{
+		"/var/lib/pgsql/data",
+		"/var/lib/pgsql/17/data",
+		"/var/lib/pgsql/16/data",
+		"/var/lib/pgsql/15/data",
+		"/var/lib/postgresql/17/main",
+		"/var/lib/postgresql/16/main",
+		"/var/lib/postgresql/15/main",
+	}
+	for _, pgdata := range pgdataPaths {
+		cmd = exec.CommandContext(ctx, "sudo", "-u", "postgres", "pg_ctl", "restart", "-D", pgdata, "-m", "fast")
+		if err := cmd.Run(); err == nil {
+			return true
+		}
+	}
+
+	return false
+}
+
 // resetPostgreSQLSettings restores original PostgreSQL settings
+// NOTE: max_locks_per_transaction changes are written but require restart to take effect.
+// We don't restart here since we're done with the restore.
 func (e *Engine) resetPostgreSQLSettings(ctx context.Context, original *OriginalSettings) error {
 	connStr := e.buildConnString()
 	db, err := sql.Open("pgx", connStr)
@@ -1988,25 +2073,28 @@ func (e *Engine) resetPostgreSQLSettings(ctx context.Context, original *Original
 	}
 	defer db.Close()
 
-	// Reset max_locks_per_transaction
+	// Reset max_locks_per_transaction (will take effect on next restart)
 	if original.MaxLocks == 64 { // Default
 		db.ExecContext(ctx, "ALTER SYSTEM RESET max_locks_per_transaction")
 	} else if original.MaxLocks > 0 {
 		db.ExecContext(ctx, fmt.Sprintf("ALTER SYSTEM SET max_locks_per_transaction = %d", original.MaxLocks))
 	}
 
-	// Reset maintenance_work_mem
+	// Reset maintenance_work_mem (takes effect immediately with reload)
 	if original.MaintenanceWorkMem == "64MB" { // Default
 		db.ExecContext(ctx, "ALTER SYSTEM RESET maintenance_work_mem")
 	} else if original.MaintenanceWorkMem != "" {
 		db.ExecContext(ctx, fmt.Sprintf("ALTER SYSTEM SET maintenance_work_mem = '%s'", original.MaintenanceWorkMem))
 	}
 
-	// Reload config
+	// Reload config (only maintenance_work_mem will take effect immediately)
 	_, err = db.ExecContext(ctx, "SELECT pg_reload_conf()")
 	if err != nil {
 		return fmt.Errorf("failed to reload config: %w", err)
 	}
 
+	e.log.Info("PostgreSQL settings reset queued",
+		"note", "max_locks_per_transaction will revert on next PostgreSQL restart")
+
 	return nil
 }

internal/restore/preflight.go

@@ -201,10 +201,19 @@ func (e *Engine) checkPostgreSQL(ctx context.Context, result *PreflightResult) {
 		result.PostgreSQL.IsSuperuser = isSuperuser
 	}
 
-	// Add info/warnings
+	// CRITICAL: max_locks_per_transaction requires PostgreSQL RESTART to change!
+	// Warn users loudly about this - it's the #1 cause of "out of shared memory" errors
 	if result.PostgreSQL.MaxLocksPerTransaction < 256 {
-		e.log.Info("PostgreSQL max_locks_per_transaction is low - will auto-boost",
-			"current", result.PostgreSQL.MaxLocksPerTransaction)
+		e.log.Warn("PostgreSQL max_locks_per_transaction is LOW",
+			"current", result.PostgreSQL.MaxLocksPerTransaction,
+			"recommended", "256+",
+			"note", "REQUIRES PostgreSQL restart to change!")
+
+		result.Warnings = append(result.Warnings,
+			fmt.Sprintf("max_locks_per_transaction=%d is low (recommend 256+). "+
+				"This setting requires PostgreSQL RESTART to change. "+
+				"BLOB-heavy databases may fail with 'out of shared memory' error.",
+				result.PostgreSQL.MaxLocksPerTransaction))
 	}
 
 	// Parse shared_buffers and warn if very low

internal/tui/backup_exec.go

@@ -4,6 +4,7 @@ import (
 	"context"
 	"fmt"
 	"strings"
+	"sync"
 	"time"
 
 	tea "github.com/charmbracelet/bubbletea"
@@ -33,6 +34,56 @@ type BackupExecutionModel struct {
 	startTime    time.Time
 	details      []string
 	spinnerFrame int
+
+	// Database count progress (for cluster backup)
+	dbTotal int
+	dbDone  int
+	dbName  string // Current database being backed up
+}
+
+// sharedBackupProgressState holds progress state that can be safely accessed from callbacks
+type sharedBackupProgressState struct {
+	mu        sync.Mutex
+	dbTotal   int
+	dbDone    int
+	dbName    string
+	hasUpdate bool
+}
+
+// Package-level shared progress state for backup operations
+var (
+	currentBackupProgressMu    sync.Mutex
+	currentBackupProgressState *sharedBackupProgressState
+)
+
+func setCurrentBackupProgress(state *sharedBackupProgressState) {
+	currentBackupProgressMu.Lock()
+	defer currentBackupProgressMu.Unlock()
+	currentBackupProgressState = state
+}
+
+func clearCurrentBackupProgress() {
+	currentBackupProgressMu.Lock()
+	defer currentBackupProgressMu.Unlock()
+	currentBackupProgressState = nil
+}
+
+func getCurrentBackupProgress() (dbTotal, dbDone int, dbName string, hasUpdate bool) {
+	currentBackupProgressMu.Lock()
+	defer currentBackupProgressMu.Unlock()
+
+	if currentBackupProgressState == nil {
+		return 0, 0, "", false
+	}
+
+	currentBackupProgressState.mu.Lock()
+	defer currentBackupProgressState.mu.Unlock()
+
+	hasUpdate = currentBackupProgressState.hasUpdate
+	currentBackupProgressState.hasUpdate = false
+
+	return currentBackupProgressState.dbTotal, currentBackupProgressState.dbDone,
+		currentBackupProgressState.dbName, hasUpdate
 }
 
 func NewBackupExecution(cfg *config.Config, log logger.Logger, parent tea.Model, ctx context.Context, backupType, dbName string, ratio int) BackupExecutionModel {
@@ -55,7 +106,6 @@ func NewBackupExecution(cfg *config.Config, log logger.Logger, parent tea.Model,
 }
 
 func (m BackupExecutionModel) Init() tea.Cmd {
-	// TUI handles all display through View() - no progress callbacks needed
 	return tea.Batch(
 		executeBackupWithTUIProgress(m.ctx, m.config, m.logger, m.backupType, m.databaseName, m.ratio),
 		backupTickCmd(),
@@ -91,6 +141,11 @@ func executeBackupWithTUIProgress(parentCtx context.Context, cfg *config.Config,
 
 		start := time.Now()
 
+		// Setup shared progress state for TUI polling
+		progressState := &sharedBackupProgressState{}
+		setCurrentBackupProgress(progressState)
+		defer clearCurrentBackupProgress()
+
 		dbClient, err := database.New(cfg, log)
 		if err != nil {
 			return backupCompleteMsg{
@@ -110,6 +165,16 @@ func executeBackupWithTUIProgress(parentCtx context.Context, cfg *config.Config,
 		// Pass nil as indicator - TUI itself handles all display, no stdout printing
 		engine := backup.NewSilent(cfg, log, dbClient, nil)
 
+		// Set database progress callback for cluster backups
+		engine.SetDatabaseProgressCallback(func(done, total int, currentDB string) {
+			progressState.mu.Lock()
+			progressState.dbDone = done
+			progressState.dbTotal = total
+			progressState.dbName = currentDB
+			progressState.hasUpdate = true
+			progressState.mu.Unlock()
+		})
+
 		var backupErr error
 		switch backupType {
 		case "single":
@@ -157,10 +222,21 @@ func (m BackupExecutionModel) Update(msg tea.Msg) (tea.Model, tea.Cmd) {
 			// Increment spinner frame for smooth animation
 			m.spinnerFrame = (m.spinnerFrame + 1) % len(spinnerFrames)
 
-			// Update status based on elapsed time to show progress
+			// Poll for database progress updates from callbacks
+			dbTotal, dbDone, dbName, hasUpdate := getCurrentBackupProgress()
+			if hasUpdate {
+				m.dbTotal = dbTotal
+				m.dbDone = dbDone
+				m.dbName = dbName
+			}
+
+			// Update status based on progress and elapsed time
 			elapsedSec := int(time.Since(m.startTime).Seconds())
 
-			if elapsedSec < 2 {
+			if m.dbTotal > 0 && m.dbDone > 0 {
+				// We have real progress from cluster backup
+				m.status = fmt.Sprintf("Backing up database: %s", m.dbName)
+			} else if elapsedSec < 2 {
 				m.status = "Initializing backup..."
 			} else if elapsedSec < 5 {
 				if m.backupType == "cluster" {
@@ -234,6 +310,34 @@ func (m BackupExecutionModel) Update(msg tea.Msg) (tea.Model, tea.Cmd) {
 	return m, nil
 }
 
+// renderDatabaseProgressBar renders a progress bar for database count progress
+func renderBackupDatabaseProgressBar(done, total int, dbName string, width int) string {
+	if total == 0 {
+		return ""
+	}
+
+	// Calculate progress percentage
+	percent := float64(done) / float64(total)
+	if percent > 1.0 {
+		percent = 1.0
+	}
+
+	// Calculate filled width
+	barWidth := width - 20 // Leave room for label and percentage
+	if barWidth < 10 {
+		barWidth = 10
+	}
+	filled := int(float64(barWidth) * percent)
+	if filled > barWidth {
+		filled = barWidth
+	}
+
+	// Build progress bar
+	bar := strings.Repeat("█", filled) + strings.Repeat("░", barWidth-filled)
+
+	return fmt.Sprintf("  Database: [%s] %d/%d", bar, done, total)
+}
+
 func (m BackupExecutionModel) View() string {
 	var s strings.Builder
 	s.Grow(512) // Pre-allocate estimated capacity for better performance
@@ -255,12 +359,24 @@ func (m BackupExecutionModel) View() string {
 	s.WriteString(fmt.Sprintf("  %-10s %s\n", "Duration:", time.Since(m.startTime).Round(time.Second)))
 	s.WriteString("\n")
 
-	// Status with spinner
+	// Status display
 	if !m.done {
-		if m.cancelling {
-			s.WriteString(fmt.Sprintf("  %s %s\n", spinnerFrames[m.spinnerFrame], m.status))
+		// Show database progress bar if we have progress data (cluster backup)
+		if m.dbTotal > 0 && m.dbDone > 0 {
+			// Show progress bar instead of spinner when we have real progress
+			progressBar := renderBackupDatabaseProgressBar(m.dbDone, m.dbTotal, m.dbName, 50)
+			s.WriteString(progressBar + "\n")
+			s.WriteString(fmt.Sprintf("  %s\n", m.status))
 		} else {
-			s.WriteString(fmt.Sprintf("  %s %s\n", spinnerFrames[m.spinnerFrame], m.status))
+			// Show spinner during initial phases
+			if m.cancelling {
+				s.WriteString(fmt.Sprintf("  %s %s\n", spinnerFrames[m.spinnerFrame], m.status))
+			} else {
+				s.WriteString(fmt.Sprintf("  %s %s\n", spinnerFrames[m.spinnerFrame], m.status))
+			}
+		}
+
+		if !m.cancelling {
 			s.WriteString("\n  [KEY]  Press Ctrl+C or ESC to cancel\n")
 		}
 	} else {