cmd/gc: fix the issue that pending delete objects are misclassified a…

…s leaked (#5198)
juicedata · Sep 27, 2024 · 014a68d · 014a68d
1 parent 730a347
commit 014a68d
Show file tree

Hide file tree

Showing 7 changed files with 81 additions and 17 deletions.
diff --git a/cmd/fsck.go b/cmd/fsck.go
@@ -142,7 +142,7 @@ func fsck(ctx *cli.Context) error {
 	// List all slices in metadata engine
 	sliceCSpin := progress.AddCountSpinner("Listed slices")
 	slices := make(map[meta.Ino][]meta.Slice)
-	r := m.ListSlices(c, slices, false, sliceCSpin.Increment)
+	r := m.ListSlices(c, slices, false, false, sliceCSpin.Increment)
 	if r != 0 {
 		logger.Fatalf("list all slices: %s", r)
 	}

diff --git a/cmd/gc.go b/cmd/gc.go
@@ -210,7 +210,7 @@ func gc(ctx *cli.Context) error {
 
 	// List all slices in metadata engine
 	slices := make(map[meta.Ino][]meta.Slice)
-	r := m.ListSlices(c, slices, delete, sliceCSpin.Increment)
+	r := m.ListSlices(c, slices, true, delete, sliceCSpin.Increment)
 	if r != 0 {
 		logger.Fatalf("list all slices: %s", r)
 	}
@@ -247,9 +247,16 @@ func gc(ctx *cli.Context) error {
 		logger.Fatalf("list all blocks: %s", err)
 	}
 	vkeys := make(map[uint64]uint32)
+	pkeys := make(map[uint64]uint32)
 	ckeys := make(map[uint64]uint32)
 	var total int64
 	var totalBytes uint64
+	for _, s := range slices[0] {
+		pkeys[s.Id] = s.Size
+		total += int64(int(s.Size-1)/chunkConf.BlockSize) + 1
+		totalBytes += uint64(s.Size)
+	}
+	slices[0] = nil
 	for _, s := range slices[1] {
 		ckeys[s.Id] = s.Size
 		total += int64(int(s.Size-1)/chunkConf.BlockSize) + 1
@@ -269,6 +276,7 @@ func gc(ctx *cli.Context) error {
 
 	bar := progress.AddCountBar("Scanned objects", total)
 	valid := progress.AddDoubleSpinnerTwo("Valid objects", "Valid data")
+	pending := progress.AddDoubleSpinnerTwo("Pending delete objects", "Pending delete data")
 	compacted := progress.AddDoubleSpinnerTwo("Compacted objects", "Compacted data")
 	leaked := progress.AddDoubleSpinnerTwo("Leaked objects", "Leaked data")
 	skipped := progress.AddDoubleSpinnerTwo("Skipped objects", "Skipped data")
@@ -321,10 +329,12 @@ func gc(ctx *cli.Context) error {
 		bar.Increment()
 		cid, _ := strconv.Atoi(parts[0])
 		size := vkeys[uint64(cid)]
-		var cobj bool
+		var pobj, cobj bool
+		if size == 0 {
+			size, pobj = pkeys[uint64(cid)]
+		}
 		if size == 0 {
-			size = ckeys[uint64(cid)]
-			cobj = true
+			size, cobj = ckeys[uint64(cid)]
 		}
 		if size == 0 {
 			logger.Debugf("find leaked object: %s, size: %d", obj.Key(), obj.Size())
@@ -337,6 +347,8 @@ func gc(ctx *cli.Context) error {
 			if (indx+1)*csize > int(size) {
 				logger.Warnf("size of slice %d is larger than expected: %d > %d", cid, indx*chunkConf.BlockSize+csize, size)
 				foundLeaked(obj)
+			} else if pobj {
+				pending.IncrInt64(obj.Size())
 			} else if cobj {
 				compacted.IncrInt64(obj.Size())
 			} else {
@@ -346,6 +358,8 @@ func gc(ctx *cli.Context) error {
 			if indx*chunkConf.BlockSize+csize != int(size) {
 				logger.Warnf("size of slice %d is %d, but expect %d", cid, indx*chunkConf.BlockSize+csize, size)
 				foundLeaked(obj)
+			} else if pobj {
+				pending.IncrInt64(obj.Size())
 			} else if cobj {
 				compacted.IncrInt64(obj.Size())
 			} else {
@@ -371,13 +385,14 @@ func gc(ctx *cli.Context) error {
 	progress.Done()
 
 	vc, _ := valid.Current()
+	pc, pb := pending.Current()
 	cc, cb := compacted.Current()
 	lc, lb := leaked.Current()
 	sc, sb := skipped.Current()
 	dsc, dsb := cleanedSliceSpin.Current()
 	fc, fb := cleanedFileSpin.Current()
-	logger.Infof("scanned %d objects, %d valid, %d compacted (%d bytes), %d leaked (%d bytes), %d delslices (%d bytes), %d delfiles (%d bytes), %d skipped (%d bytes)",
-		bar.Current(), vc, cc, cb, lc, lb, dsc, dsb, fc, fb, sc, sb)
+	logger.Infof("scanned %d objects, %d valid, %d pending delete (%d bytes), %d compacted (%d bytes), %d leaked (%d bytes), %d delslices (%d bytes), %d delfiles (%d bytes), %d skipped (%d bytes)",
+		bar.Current(), vc, pc, pb, cc, cb, lc, lb, dsc, dsb, fc, fb, sc, sb)
 	if lc > 0 && !delete {
 		logger.Infof("Please add `--delete` to clean leaked objects")
 	}

diff --git a/pkg/meta/base_test.go b/pkg/meta/base_test.go
@@ -1486,7 +1486,7 @@ func testCompaction(t *testing.T, m Meta, trash bool) {
 	}
 	p.Done()
 	sliceMap := make(map[Ino][]Slice)
-	if st := m.ListSlices(ctx, sliceMap, false, nil); st != 0 {
+	if st := m.ListSlices(ctx, sliceMap, false, false, nil); st != 0 {
 		t.Fatalf("list all slices: %s", st)
 	}
 
@@ -1694,7 +1694,7 @@ func testTruncateAndDelete(t *testing.T, m Meta) {
 	}
 	var total int64
 	slices := make(map[Ino][]Slice)
-	m.ListSlices(ctx, slices, false, func() { total++ })
+	m.ListSlices(ctx, slices, false, false, func() { total++ })
 	var totalSlices int
 	for _, ss := range slices {
 		totalSlices += len(ss)
@@ -1709,7 +1709,7 @@ func testTruncateAndDelete(t *testing.T, m Meta) {
 
 	time.Sleep(time.Millisecond * 100)
 	slices = make(map[Ino][]Slice)
-	m.ListSlices(ctx, slices, false, nil)
+	m.ListSlices(ctx, slices, false, false, nil)
 	totalSlices = 0
 	for _, ss := range slices {
 		totalSlices += len(ss)

diff --git a/pkg/meta/interface.go b/pkg/meta/interface.go
@@ -418,7 +418,7 @@ type Meta interface {
 	Compact(ctx Context, inode Ino, concurrency int, preFunc, postFunc func()) syscall.Errno
 
 	// ListSlices returns all slices used by all files.
-	ListSlices(ctx Context, slices map[Ino][]Slice, delete bool, showProgress func()) syscall.Errno
+	ListSlices(ctx Context, slices map[Ino][]Slice, scanPending, delete bool, showProgress func()) syscall.Errno
 	// Remove all files and directories recursively.
 	// count represents the number of attempted deletions of entries (even if failed).
 	Remove(ctx Context, parent Ino, name string, count *uint64) syscall.Errno

diff --git a/pkg/meta/redis.go b/pkg/meta/redis.go
@@ -3041,7 +3041,7 @@ func (m *redisMeta) hscan(ctx context.Context, key string, f func([]string) erro
 	return nil
 }
 
-func (m *redisMeta) ListSlices(ctx Context, slices map[Ino][]Slice, delete bool, showProgress func()) syscall.Errno {
+func (m *redisMeta) ListSlices(ctx Context, slices map[Ino][]Slice, scanPending, delete bool, showProgress func()) syscall.Errno {
 	m.cleanupLeakedInodes(delete)
 	m.cleanupLeakedChunks(delete)
 	m.cleanupOldSliceRefs(delete)
@@ -3083,10 +3083,33 @@ func (m *redisMeta) ListSlices(ctx Context, slices map[Ino][]Slice, delete bool,
 		}
 		return nil
 	})
-	if err != nil || m.getFormat().TrashDays == 0 {
+	if err != nil {
+		logger.Warnf("scan chunks: %s", err)
 		return errno(err)
 	}
 
+	if scanPending {
+		_ = m.hscan(Background, m.sliceRefs(), func(keys []string) error {
+			for i := 0; i < len(keys); i += 2 {
+				key, val := keys[i], keys[i+1]
+				if strings.HasPrefix(val, "-") { // < 0
+					ps := strings.Split(key, "_")
+					if len(ps) == 2 {
+						id, _ := strconv.ParseUint(ps[0][1:], 10, 64)
+						size, _ := strconv.ParseUint(ps[1], 10, 32)
+						if id > 0 && size > 0 {
+							slices[0] = append(slices[0], Slice{Id: id, Size: uint32(size)})
+						}
+					}
+				}
+			}
+			return nil
+		})
+	}
+
+	if m.getFormat().TrashDays == 0 {
+		return 0
+	}
 	return errno(m.scanTrashSlices(ctx, func(ss []Slice, _ int64) (bool, error) {
 		slices[1] = append(slices[1], ss...)
 		if showProgress != nil {

diff --git a/pkg/meta/sql.go b/pkg/meta/sql.go
@@ -2851,7 +2851,7 @@ func (m *dbMeta) scanAllChunks(ctx Context, ch chan<- cchunk, bar *utils.Bar) er
 	})
 }
 
-func (m *dbMeta) ListSlices(ctx Context, slices map[Ino][]Slice, delete bool, showProgress func()) syscall.Errno {
+func (m *dbMeta) ListSlices(ctx Context, slices map[Ino][]Slice, scanPending, delete bool, showProgress func()) syscall.Errno {
 	if delete {
 		m.doCleanupSlices()
 	}
@@ -2881,10 +2881,24 @@ func (m *dbMeta) ListSlices(ctx Context, slices map[Ino][]Slice, delete bool, sh
 	if err != nil {
 		return errno(err)
 	}
+
+	if scanPending {
+		_ = m.roTxn(func(s *xorm.Session) error {
+			var cks []sliceRef
+			err := s.Where("refs <= 0").Find(&cks)
+			if err != nil {
+				return err
+			}
+			for _, ck := range cks {
+				slices[0] = append(slices[0], Slice{Id: ck.Id, Size: ck.Size})
+			}
+			return nil
+		})
+	}
+
 	if m.getFormat().TrashDays == 0 {
 		return 0
 	}
-
 	return errno(m.scanTrashSlices(ctx, func(ss []Slice, _ int64) (bool, error) {
 		slices[1] = append(slices[1], ss...)
 		if showProgress != nil {

diff --git a/pkg/meta/tkv.go b/pkg/meta/tkv.go
@@ -2392,7 +2392,7 @@ func (m *kvMeta) scanAllChunks(ctx Context, ch chan<- cchunk, bar *utils.Bar) er
 	})
 }
 
-func (m *kvMeta) ListSlices(ctx Context, slices map[Ino][]Slice, delete bool, showProgress func()) syscall.Errno {
+func (m *kvMeta) ListSlices(ctx Context, slices map[Ino][]Slice, scanPending, delete bool, showProgress func()) syscall.Errno {
 	if delete {
 		m.doCleanupSlices()
 	}
@@ -2421,10 +2421,22 @@ func (m *kvMeta) ListSlices(ctx Context, slices map[Ino][]Slice, delete bool, sh
 			}
 		}
 	}
+
+	if scanPending {
+		// slice refs: Kccccccccnnnn
+		klen = 1 + 8 + 4
+		result, _ = m.scanValues(m.fmtKey("K"), -1, func(k, v []byte) bool {
+			return len(k) == klen && len(v) == 8 && parseCounter(v) < 0
+		})
+		for k := range result {
+			rb := utils.FromBuffer([]byte(k)[1:])
+			slices[0] = append(slices[0], Slice{Id: rb.Get64(), Size: rb.Get32()})
+		}
+	}
+
 	if m.getFormat().TrashDays == 0 {
 		return 0
 	}
-
 	return errno(m.scanTrashSlices(ctx, func(ss []Slice, _ int64) (bool, error) {
 		slices[1] = append(slices[1], ss...)
 		if showProgress != nil {