deploy-felhom-compose/controller/internal/storage/migrate_drive.go

package storage

import (
	"bufio"
	"context"
	"fmt"
	"log"
	"os"
	"os/exec"
	"path/filepath"
	"strings"
	"sync"
	"time"

	"gitea.dooplex.hu/admin/felhom-controller/internal/backup"
	"gitea.dooplex.hu/admin/felhom-controller/internal/settings"
)

// StackProviderForMigration abstracts stack operations needed by drive migration.
type StackProviderForMigration interface {
	ListDeployedStacks() []StackSummaryForMigration
	GetStackHDDPath(name string) string
	StopStack(name string) error
	StartStack(name string) error
	UpdateStackHDDPath(name, newPath string) error
	StackExists(name string) bool
}

// StackSummaryForMigration holds minimal stack info for drive migration.
type StackSummaryForMigration struct {
	Name        string
	DisplayName string
}

// DriveMigrateRequest holds parameters for migrating all apps from one drive to another.
type DriveMigrateRequest struct {
	SourcePath string // e.g., "/mnt/hdd_1"
	DestPath   string // e.g., "/mnt/hdd_2"
}

// DriveMigrateProgress tracks drive migration state.
type DriveMigrateProgress struct {
	Step           string // "validating","stopping","copying","verifying","configuring","starting","backup","done","error","rolling_back"
	Message        string
	BytesCopied    int64
	BytesTotal     int64
	Percent        int
	Error          string
	ElapsedSeconds int
	Detail         string // sub-step detail (e.g., which app is being configured)
}

// DriveMigrator orchestrates full drive migration.
type DriveMigrator struct {
	Sett            *settings.Settings
	StackProvider   StackProviderForMigration
	BackupTrigger   BackupTrigger
	AlertRefresh    func()
	PushHubReport   func()
	PushInfraBackup func()
	SyncFBMounts    func()
	Logger          *log.Logger

	mu     sync.Mutex
	active bool // global migration lock
}

// IsActive returns whether a full drive migration is currently in progress.
func (dm *DriveMigrator) IsActive() bool {
	dm.mu.Lock()
	defer dm.mu.Unlock()
	return dm.active
}

// rollbackAction describes a reversible action in the migration transaction.
type rollbackAction struct {
	description string
	undo        func() error
}

// migrationTx is a transaction log that enables reverse-order rollback.
type migrationTx struct {
	actions []rollbackAction
	logger  *log.Logger
}

func (tx *migrationTx) add(desc string, undoFn func() error) {
	tx.actions = append(tx.actions, rollbackAction{description: desc, undo: undoFn})
}

func (tx *migrationTx) rollback() {
	for i := len(tx.actions) - 1; i >= 0; i-- {
		a := tx.actions[i]
		tx.logger.Printf("[INFO] [storage] Rollback: %s", a.description)
		if err := a.undo(); err != nil {
			tx.logger.Printf("[ERROR] [storage] Rollback failed: %s: %v", a.description, err)
		}
	}
}

// MigrateDrive performs a full drive migration, moving all apps from source to dest.
func (dm *DriveMigrator) MigrateDrive(ctx context.Context, req DriveMigrateRequest, progress chan<- DriveMigrateProgress) error {
	start := time.Now()
	// TODO: debug should be driven by a dedicated Debug field, not Logger presence.
	// Currently always true when Logger is set (which is always in practice).
	debug := dm.Logger != nil

	dbg := func(format string, args ...interface{}) {
		if debug {
			dm.Logger.Printf("[DEBUG] [storage] MigrateDrive: "+format, args...)
		}
	}
	_ = dbg // used below

	dbg("starting drive migration: source=%s dest=%s", req.SourcePath, req.DestPath)

	send := func(step, msg string, pct int) {
		progress <- DriveMigrateProgress{
			Step:           step,
			Message:        msg,
			Percent:        pct,
			ElapsedSeconds: int(time.Since(start).Seconds()),
		}
	}
	sendDetail := func(step, msg, detail string, pct int) {
		progress <- DriveMigrateProgress{
			Step:           step,
			Message:        msg,
			Detail:         detail,
			Percent:        pct,
			ElapsedSeconds: int(time.Since(start).Seconds()),
		}
	}
	fail := func(msg string, err error) error {
		errStr := ""
		if err != nil {
			errStr = err.Error()
		}
		progress <- DriveMigrateProgress{
			Step:           "error",
			Message:        msg,
			Error:          errStr,
			ElapsedSeconds: int(time.Since(start).Seconds()),
		}
		return fmt.Errorf("%s: %w", msg, err)
	}

	// Acquire global migration lock
	dm.mu.Lock()
	if dm.active {
		dm.mu.Unlock()
		return fail("Egy másik meghajtó-migráció folyamatban van", fmt.Errorf("migration already active"))
	}
	dm.active = true
	dm.mu.Unlock()
	defer func() {
		dm.mu.Lock()
		dm.active = false
		dm.mu.Unlock()
	}()

	// --- Pre-validation ---
	send("validating", "Ellenőrzés...", 1)

	srcLabel := dm.Sett.GetStorageLabel(req.SourcePath)
	dstLabel := dm.Sett.GetStorageLabel(req.DestPath)

	if dm.Sett.IsDisconnected(req.SourcePath) {
		return fail("A forrás meghajtó le van választva", fmt.Errorf("source disconnected"))
	}
	if dm.Sett.IsDecommissioned(req.SourcePath) {
		return fail("A forrás meghajtó már kiváltott", fmt.Errorf("source decommissioned"))
	}
	if dm.Sett.IsDisconnected(req.DestPath) {
		return fail("A cél meghajtó le van választva", fmt.Errorf("dest disconnected"))
	}
	if dm.Sett.IsDecommissioned(req.DestPath) {
		return fail("A cél meghajtó már kiváltott", fmt.Errorf("dest decommissioned"))
	}

	// Find apps on source drive
	var appsToMigrate []StackSummaryForMigration
	for _, stack := range dm.StackProvider.ListDeployedStacks() {
		hddPath := dm.StackProvider.GetStackHDDPath(stack.Name)
		if hddPath == req.SourcePath {
			appsToMigrate = append(appsToMigrate, stack)
		}
	}

	dbg("found %d apps on source drive: %v", len(appsToMigrate), func() []string {
		names := make([]string, len(appsToMigrate))
		for i, a := range appsToMigrate {
			names[i] = a.Name
		}
		return names
	}())

	if len(appsToMigrate) == 0 {
		return fail("A forrás meghajtón nincs telepített alkalmazás", fmt.Errorf("no apps on source"))
	}

	// Check for conflicts on destination
	for _, app := range appsToMigrate {
		destAppData := backup.AppDataDir(req.DestPath, app.Name)
		if info, err := os.Stat(destAppData); err == nil && info.IsDir() {
			entries, _ := os.ReadDir(destAppData)
			if len(entries) > 0 {
				return fail(
					fmt.Sprintf("A cél meghajtón már létezik adat: %s/%s", req.DestPath, app.Name),
					fmt.Errorf("conflict: %s already exists on destination", app.Name),
				)
			}
		}
	}

	// Estimate total size (exclude restic repos inside felhom-data/backups/)
	var totalBytes int64
	entries, _ := os.ReadDir(req.SourcePath)
	for _, entry := range entries {
		if !entry.IsDir() {
			continue
		}
		entryPath := filepath.Join(req.SourcePath, entry.Name())
		if entry.Name() == backup.FelhomDataDir {
			// Scan inside namespace dir, excluding restic repos from estimate
			subEntries, _ := os.ReadDir(entryPath)
			for _, sub := range subEntries {
				if !sub.IsDir() {
					continue
				}
				subPath := filepath.Join(entryPath, sub.Name())
				if sub.Name() == "backups" {
					totalBytes += dirSizeExcluding(subPath, "restic")
				} else {
					totalBytes += dirSize(subPath)
				}
			}
		} else {
			totalBytes += dirSize(entryPath)
		}
	}

	// Check free space on destination
	freeBytes := getFreeBytes(req.DestPath)
	if freeBytes > 0 && totalBytes > 0 && int64(float64(totalBytes)*1.05) > freeBytes {
		return fail(
			fmt.Sprintf("Nincs elég szabad hely: szükséges ~%s, szabad %s",
				bytesHuman(totalBytes), bytesHuman(freeBytes)),
			fmt.Errorf("insufficient disk space"),
		)
	}

	dbg("estimated data: %s (%d bytes), free on dest: %s (%d bytes)", bytesHuman(totalBytes), totalBytes, bytesHuman(freeBytes), freeBytes)
	dm.Logger.Printf("[INFO] [storage] Drive migration: %s (%s) → %s (%s), %d apps, ~%s data",
		req.SourcePath, srcLabel, req.DestPath, dstLabel, len(appsToMigrate), bytesHuman(totalBytes))

	tx := &migrationTx{logger: dm.Logger}

	// --- Step 1: Stop all affected apps ---
	send("stopping", fmt.Sprintf("Alkalmazások leállítása (%d db)...", len(appsToMigrate)), 5)

	var stoppedApps []string
	for _, app := range appsToMigrate {
		sendDetail("stopping", "Leállítás: "+app.DisplayName, app.Name, 5)
		if err := dm.StackProvider.StopStack(app.Name); err != nil {
			dm.Logger.Printf("[ERROR] [storage] Drive migration: failed to stop %s: %v", app.Name, err)
			// Rollback: restart already stopped apps
			send("rolling_back", "Hiba a leállításnál, visszagörgetés...", 0)
			for _, name := range stoppedApps {
				_ = dm.StackProvider.StartStack(name)
			}
			return fail("Alkalmazás leállítása sikertelen: "+app.DisplayName, err)
		}
		stoppedApps = append(stoppedApps, app.Name)
	}
	tx.add("Restart all stopped apps", func() error {
		for _, name := range stoppedApps {
			if err := dm.StackProvider.StartStack(name); err != nil {
				dm.Logger.Printf("[WARN] [storage] Rollback: failed to restart %s: %v", name, err)
			}
		}
		return nil
	})

	// --- Step 2: rsync entire drive (excluding restic repos) ---
	send("copying", "Adatok másolása...", 10)

	rsyncCmd := exec.CommandContext(ctx, "rsync", "-a", "--info=progress2",
		"--exclude=felhom-data/backups/primary/restic/",
		"--exclude=felhom-data/backups/secondary/restic/",
		req.SourcePath+"/", req.DestPath+"/",
	)

	stdout, err := rsyncCmd.StdoutPipe()
	if err != nil {
		send("rolling_back", "rsync indítása sikertelen, visszagörgetés...", 0)
		tx.rollback()
		return fail("rsync pipe hiba", err)
	}
	stderr, err := rsyncCmd.StderrPipe()
	if err != nil {
		send("rolling_back", "rsync indítása sikertelen, visszagörgetés...", 0)
		tx.rollback()
		return fail("rsync stderr pipe hiba", err)
	}

	if err := rsyncCmd.Start(); err != nil {
		send("rolling_back", "rsync indítása sikertelen, visszagörgetés...", 0)
		tx.rollback()
		return fail("rsync indítás sikertelen", err)
	}

	// Parse rsync progress
	go func() {
		scanner := bufio.NewScanner(stdout)
		for scanner.Scan() {
			line := scanner.Text()
			if b, pct, ok := parseRsyncProgress(line); ok {
				scaledPct := 10 + pct*50/100 // scale to 10-60%
				if scaledPct > 60 {
					scaledPct = 60
				}
				progress <- DriveMigrateProgress{
					Step:           "copying",
					Message:        fmt.Sprintf("Adatok másolása... %s / %s", bytesHuman(b), bytesHuman(totalBytes)),
					BytesCopied:    b,
					BytesTotal:     totalBytes,
					Percent:        scaledPct,
					ElapsedSeconds: int(time.Since(start).Seconds()),
				}
			}
		}
	}()

	var stderrBuf strings.Builder
	var stderrWg sync.WaitGroup
	stderrWg.Add(1)
	go func() {
		defer stderrWg.Done()
		buf := make([]byte, 4096)
		for {
			n, err := stderr.Read(buf)
			if n > 0 {
				stderrBuf.Write(buf[:n])
			}
			if err != nil {
				break
			}
		}
	}()

	if err := rsyncCmd.Wait(); err != nil {
		stderrWg.Wait()
		dbg("rsync failed after %s: %v — stderr: %s", time.Since(start).Round(time.Second), err, stderrBuf.String())
		send("rolling_back", "rsync sikertelen, visszagörgetés...", 0)
		tx.rollback()
		return fail("Adatmásolás sikertelen", fmt.Errorf("rsync failed: %w — %s", err, stderrBuf.String()))
	}
	stderrWg.Wait()
	dbg("rsync completed in %s", time.Since(start).Round(time.Second))

	// --- Step 3: Verify copy ---
	send("verifying", "Másolat ellenőrzése...", 62)

	for _, app := range appsToMigrate {
		destAppData := backup.AppDataDir(req.DestPath, app.Name)
		if _, err := os.Stat(destAppData); os.IsNotExist(err) {
			// appdata might not exist for all apps (SSD-only apps that share the drive)
			// Only warn, don't fail
			dm.Logger.Printf("[WARN] [storage] Drive migration: %s not found on destination (may be SSD-only)", destAppData)
		}
	}

	// --- Step 4: Update all app configs ---
	send("configuring", "Konfiguráció frissítése...", 65)

	dbg("updating HDD_PATH for %d apps", len(appsToMigrate))
	var configuredApps []string
	for i, app := range appsToMigrate {
		// Guard: verify app still exists
		if !dm.StackProvider.StackExists(app.Name) {
			dm.Logger.Printf("[WARN] [storage] Drive migration: app %s no longer exists, skipping config update", app.Name)
			continue
		}

		pct := 65 + (i * 10 / len(appsToMigrate))
		sendDetail("configuring", "Konfiguráció: "+app.DisplayName, app.Name, pct)

		oldPath := dm.StackProvider.GetStackHDDPath(app.Name)
		if err := dm.StackProvider.UpdateStackHDDPath(app.Name, req.DestPath); err != nil {
			dm.Logger.Printf("[ERROR] [storage] Drive migration: failed to update HDD_PATH for %s: %v", app.Name, err)
			send("rolling_back", "Konfiguráció frissítése sikertelen, visszagörgetés...", 0)
			// Rollback config changes
			for _, name := range configuredApps {
				_ = dm.StackProvider.UpdateStackHDDPath(name, req.SourcePath)
			}
			tx.rollback()
			return fail("HDD_PATH frissítés sikertelen: "+app.DisplayName, err)
		}
		configuredApps = append(configuredApps, app.Name)
		tx.add("Revert HDD_PATH for "+app.Name, func() error {
			return dm.StackProvider.UpdateStackHDDPath(app.Name, oldPath)
		})
	}

	// --- Step 5: Update storage registry ---
	send("configuring", "Tárolóregiszter frissítése...", 76)

	// Transfer IsDefault
	allPaths := dm.Sett.GetStoragePaths()
	var srcWasDefault bool
	var srcWasSchedulable bool
	for _, sp := range allPaths {
		if sp.Path == req.SourcePath {
			srcWasDefault = sp.IsDefault
			srcWasSchedulable = sp.Schedulable
		}
	}

	if srcWasDefault {
		_ = dm.Sett.SetDefaultStoragePath(req.DestPath)
	}
	if srcWasSchedulable {
		_ = dm.Sett.SetSchedulable(req.DestPath, true)
	}

	// Mark source as decommissioned
	if err := dm.Sett.SetDecommissioned(req.SourcePath, req.DestPath); err != nil {
		dm.Logger.Printf("[WARN] [storage] Drive migration: failed to mark source as decommissioned: %v", err)
	}
	tx.add("Clear decommissioned on source", func() error {
		return dm.Sett.ClearDecommissioned(req.SourcePath)
	})

	// --- Step 6: Update Tier 2 cross-drive configs ---
	send("configuring", "Mentési beállítások frissítése...", 78)

	allCrossConfigs := dm.Sett.GetAllCrossDriveConfigs()
	for name, cfg := range allCrossConfigs {
		if cfg == nil {
			continue
		}
		// Apps that moved (source→dest) with Tier 2 pointing to dest: clear (no redundancy)
		appHDD := dm.StackProvider.GetStackHDDPath(name)
		if appHDD == req.DestPath && cfg.DestinationPath == req.DestPath {
			dm.Logger.Printf("[INFO] [storage] Drive migration: clearing Tier 2 for %s (dest same as app drive)", name)
			_ = dm.Sett.SetCrossDriveConfig(name, nil)
			continue
		}
		// Apps on OTHER drives with Tier 2 pointing to source: redirect to dest
		if cfg.DestinationPath == req.SourcePath {
			dm.Logger.Printf("[INFO] [storage] Drive migration: redirecting Tier 2 for %s from %s to %s", name, req.SourcePath, req.DestPath)
			cfg.DestinationPath = req.DestPath
			_ = dm.Sett.SetCrossDriveConfig(name, cfg)
		}
	}

	// --- Step 7: Start all migrated apps ---
	send("starting", "Alkalmazások indítása...", 80)

	for i, app := range appsToMigrate {
		if !dm.StackProvider.StackExists(app.Name) {
			continue
		}
		pct := 80 + (i * 8 / len(appsToMigrate))
		sendDetail("starting", "Indítás: "+app.DisplayName, app.Name, pct)

		if err := dm.StackProvider.StartStack(app.Name); err != nil {
			dm.Logger.Printf("[WARN] [storage] Drive migration: failed to start %s after migration: %v", app.Name, err)
			// Non-fatal — log but continue
		}
	}

	// At this point, migration is considered successful — no more rollback.

	// --- Step 8: Trigger immediate backup ---
	send("backup", "Biztonsági mentés indítása...", 90)

	if dm.BackupTrigger != nil {
		if err := dm.BackupTrigger.TryRunDriveBackup(ctx, req.DestPath); err != nil {
			dm.Logger.Printf("[WARN] [storage] Drive migration: post-migration backup failed: %v", err)
		}
	}

	// --- Step 9: Post-migration notifications ---
	send("configuring", "Befejező lépések...", 95)

	if dm.SyncFBMounts != nil {
		dm.SyncFBMounts()
	}
	if dm.AlertRefresh != nil {
		dm.AlertRefresh()
	}
	if dm.PushHubReport != nil {
		dm.PushHubReport()
	}
	if dm.PushInfraBackup != nil {
		dm.PushInfraBackup()
	}

	elapsed := time.Since(start)
	dm.Logger.Printf("[INFO] [storage] Drive migration complete: %s → %s, %d apps, %s elapsed",
		req.SourcePath, req.DestPath, len(appsToMigrate), elapsed.Round(time.Second))

	// --- Step 10: Done ---
	appNames := make([]string, len(appsToMigrate))
	for i, app := range appsToMigrate {
		appNames[i] = app.DisplayName
	}

	progress <- DriveMigrateProgress{
		Step: "done",
		Message: fmt.Sprintf("A %s meghajtó sikeresen kiváltva! %d alkalmazás átköltöztetve ide: %s (%s). Idő: %s",
			srcLabel, len(appsToMigrate), dstLabel, req.DestPath, elapsed.Round(time.Second)),
		Percent:        100,
		ElapsedSeconds: int(elapsed.Seconds()),
		Detail:         strings.Join(appNames, ", "),
	}

	return nil
}

// dirSizeExcluding returns the total bytes in a directory, excluding subdirectories named excludeName.
func dirSizeExcluding(path, excludeName string) int64 {
	var total int64
	filepath.Walk(path, func(p string, info os.FileInfo, err error) error {
		if err != nil {
			return nil
		}
		if info.IsDir() && info.Name() == excludeName {
			return filepath.SkipDir
		}
		if !info.IsDir() {
			total += info.Size()
		}
		return nil
	})
	return total
}