v5.4.5: Fix disk space estimation for cluster archives

- Use 1.2x multiplier for cluster .tar.gz (pre-compressed dumps)
- Use 5x multiplier for single .sql.gz files (was 7x)
- New CheckSystemMemoryWithType() for archive-aware estimation
- 119GB archive now estimates ~143GB instead of ~833GB

CHANGELOG.md

@@ -5,6 +5,35 @@ All notable changes to dbbackup will be documented in this file.
 The format is based on [Keep a Changelog](https://keepachangelog.com/en/1.0.0/),
 and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0.html).
 
+## [5.4.5] - 2026-02-02
+
+### Fixed
+- **Accurate Disk Space Estimation for Cluster Archives**
+  - Fixed WARNING showing 836GB for 119GB archive - was using wrong compression multiplier
+  - Cluster archives (.tar.gz) contain pre-compressed .dump files → now uses 1.2x multiplier
+  - Single SQL files (.sql.gz) still use 5x multiplier (was 7x, slightly optimized)
+  - New `CheckSystemMemoryWithType(size, isClusterArchive)` method for accurate estimates
+  - 119GB cluster archive now correctly estimates ~143GB instead of ~833GB
+
+## [5.4.4] - 2026-02-02
+
+### Fixed
+- **TUI Header Separator Fix** - Capped separator length at 40 chars to prevent line overflow on wide terminals
+
+## [5.4.3] - 2026-02-02
+
+### Fixed
+- **Bulletproof SIGINT Handling** - Zero zombie processes guaranteed
+  - All external commands now use `cleanup.SafeCommand()` with process group isolation
+  - `KillCommandGroup()` sends signals to entire process group (-pgid)
+  - No more orphaned pg_restore/pg_dump/psql/pigz processes on Ctrl+C
+  - 16 files updated with proper signal handling
+
+- **Eliminated External gzip Process** - The `zgrep` command was spawning `gzip -cdfq`
+  - Replaced with in-process pgzip decompression in `preflight.go`
+  - `estimateBlobsInSQL()` now uses pure Go pgzip.NewReader
+  - Zero external gzip processes during restore
+
 ## [5.1.22] - 2026-02-01
 
 ### Added

internal/restore/engine.go

@@ -1371,7 +1371,7 @@ func (e *Engine) RestoreCluster(ctx context.Context, archivePath string, preExtr
 	if statErr == nil && archiveStats != nil {
 		backupSizeBytes = archiveStats.Size()
 	}
-	memCheck := guard.CheckSystemMemory(backupSizeBytes)
+	memCheck := guard.CheckSystemMemoryWithType(backupSizeBytes, true) // true = cluster archive with pre-compressed dumps
 	if memCheck != nil {
 		if memCheck.Critical {
 			e.log.Error("🚨 CRITICAL MEMORY WARNING", "error", memCheck.Recommendation)

internal/restore/large_db_guard.go

@@ -358,6 +358,14 @@ func (g *LargeDBGuard) WarnUser(strategy *RestoreStrategy, silentMode bool) {
 
 // CheckSystemMemory validates system has enough memory for restore
 func (g *LargeDBGuard) CheckSystemMemory(backupSizeBytes int64) *MemoryCheck {
+	return g.CheckSystemMemoryWithType(backupSizeBytes, false)
+}
+
+// CheckSystemMemoryWithType validates system memory with archive type awareness
+// isClusterArchive: true for .tar.gz cluster backups (contain pre-compressed .dump files)
+//
+//	false for single .sql.gz files (compressed SQL that expands significantly)
+func (g *LargeDBGuard) CheckSystemMemoryWithType(backupSizeBytes int64, isClusterArchive bool) *MemoryCheck {
 	check := &MemoryCheck{
 		BackupSizeGB: float64(backupSizeBytes) / (1024 * 1024 * 1024),
 	}
@@ -374,8 +382,18 @@ func (g *LargeDBGuard) CheckSystemMemory(backupSizeBytes int64) *MemoryCheck {
 	check.SwapTotalGB = float64(memInfo.SwapTotal) / (1024 * 1024 * 1024)
 	check.SwapFreeGB = float64(memInfo.SwapFree) / (1024 * 1024 * 1024)
 
-	// Estimate uncompressed size (typical compression ratio 5:1 to 10:1)
-	estimatedUncompressedGB := check.BackupSizeGB * 7 // Conservative estimate
+	// Estimate uncompressed size based on archive type:
+	// - Cluster archives (.tar.gz): contain pre-compressed .dump files, ratio ~1.2x
+	// - Single SQL files (.sql.gz): compressed SQL expands significantly, ratio ~5-7x
+	var compressionMultiplier float64
+	if isClusterArchive {
+		compressionMultiplier = 1.2 // tar.gz with already-compressed .dump files
+		g.log.Debug("Using cluster archive compression ratio", "multiplier", compressionMultiplier)
+	} else {
+		compressionMultiplier = 5.0 // Conservative for gzipped SQL (was 7, reduced to 5)
+		g.log.Debug("Using single file compression ratio", "multiplier", compressionMultiplier)
+	}
+	estimatedUncompressedGB := check.BackupSizeGB * compressionMultiplier
 
 	// Memory requirements
 	// - PostgreSQL needs ~2-4GB for shared_buffers

internal/tui/rich_cluster_progress.go

@@ -93,7 +93,12 @@ func (v *RichClusterProgressView) renderHeader(snapshot *progress.ProgressSnapsh
 	}
 
 	title := "Cluster Restore Progress"
-	separator := strings.Repeat("━", maxInt(0, v.width-len(title)-4))
+	// Cap separator at 40 chars to avoid long lines on wide terminals
+	sepLen := maxInt(0, v.width-len(title)-4)
+	if sepLen > 40 {
+		sepLen = 40
+	}
+	separator := strings.Repeat("━", sepLen)
 
 	return fmt.Sprintf("%s %s\n  Elapsed: %s | %s",
 		title, separator,

main.go

@@ -16,7 +16,7 @@ import (
 
 // Build information (set by ldflags)
 var (
-	version   = "5.4.3"
+	version   = "5.4.5"
 	buildTime = "unknown"
 	gitCommit = "unknown"
 )