Merge pull request #1036 from restic/prune-remove-invalid-files

prune: Remove invalid files

CHANGELOG.md

@@ -40,6 +40,12 @@ Small changes
    https://github.com/restic/restic/issues/1023
    https://github.com/restic/restic/pull/1025
 
+ * The `prune` command has been improved and will now remove invalid pack
+   files, for example files that have not been uploaded completely because a
+   backup was interrupted.
+   https://github.com/restic/restic/issues/1029
+   https://github.com/restic/restic/pull/1036
+
 Important Changes in 0.6.1
 ==========================
 

src/cmds/restic/cmd_prune.go

@@ -106,11 +106,15 @@ func pruneRepository(gopts GlobalOptions, repo restic.Repository) error {
 	Verbosef("building new index for repo\n")
 
 	bar := newProgressMax(!gopts.Quiet, uint64(stats.packs), "packs")
-	idx, err := index.New(ctx, repo, restic.NewIDSet(), bar)
+	idx, invalidFiles, err := index.New(ctx, repo, restic.NewIDSet(), bar)
 	if err != nil {
 		return err
 	}
 
+	for _, id := range invalidFiles {
+		Warnf("incomplete pack file (will be removed): %v\n", id)
+	}
+
 	blobs := 0
 	for _, pack := range idx.Packs {
 		stats.bytes += pack.Size
@@ -196,6 +200,12 @@ func pruneRepository(gopts GlobalOptions, repo restic.Repository) error {
 
 	// find packs that are unneeded
 	removePacks := restic.NewIDSet()
+
+	Verbosef("will remove %d invalid files\n", len(invalidFiles))
+	for _, id := range invalidFiles {
+		removePacks.Insert(id)
+	}
+
 	for packID, p := range idx.Packs {
 
 		hasActiveBlob := false

src/cmds/restic/cmd_rebuild_index.go

@@ -50,7 +50,7 @@ func rebuildIndex(ctx context.Context, repo restic.Repository, ignorePacks resti
 	}
 
 	bar := newProgressMax(!globalOptions.Quiet, packs-uint64(len(ignorePacks)), "packs")
-	idx, err := index.New(ctx, repo, ignorePacks, bar)
+	idx, _, err := index.New(ctx, repo, ignorePacks, bar)
 	if err != nil {
 		return err
 	}

src/restic/index/index.go

@@ -8,6 +8,7 @@ import (
 	"restic"
 	"restic/debug"
 	"restic/list"
+	"restic/pack"
 	"restic/worker"
 
 	"restic/errors"
@@ -33,22 +34,29 @@ func newIndex() *Index {
 	}
 }
 
-// New creates a new index for repo from scratch.
-func New(ctx context.Context, repo restic.Repository, ignorePacks restic.IDSet, p *restic.Progress) (*Index, error) {
+// New creates a new index for repo from scratch. InvalidFiles contains all IDs
+// of files  that cannot be listed successfully.
+func New(ctx context.Context, repo restic.Repository, ignorePacks restic.IDSet, p *restic.Progress) (idx *Index, invalidFiles restic.IDs, err error) {
 	p.Start()
 	defer p.Done()
 
 	ch := make(chan worker.Job)
 	go list.AllPacks(ctx, repo, ignorePacks, ch)
 
-	idx := newIndex()
+	idx = newIndex()
 
 	for job := range ch {
 		p.Report(restic.Stat{Blobs: 1})
 
 		packID := job.Data.(restic.ID)
 		if job.Error != nil {
-			fmt.Fprintf(os.Stderr, "unable to list pack %v: %v\n", packID.Str(), job.Error)
+			cause := errors.Cause(job.Error)
+			if _, ok := cause.(pack.InvalidFileError); ok {
+				invalidFiles = append(invalidFiles, packID)
+				continue
+			}
+
+			fmt.Fprintf(os.Stderr, "pack file cannot be listed %v: %v\n", packID.Str(), job.Error)
 			continue
 		}
 
@@ -58,11 +66,11 @@ func New(ctx context.Context, repo restic.Repository, ignorePacks restic.IDSet,
 
 		err := idx.AddPack(packID, j.Size(), j.Entries())
 		if err != nil {
-			return nil, err
+			return nil, nil, err
 		}
 	}
 
-	return idx, nil
+	return idx, invalidFiles, nil
 }
 
 type packJSON struct {

src/restic/index/index_test.go

@@ -43,7 +43,7 @@ func TestIndexNew(t *testing.T) {
 	repo, cleanup := createFilledRepo(t, 3, 0)
 	defer cleanup()
 
-	idx, err := New(context.TODO(), repo, restic.NewIDSet(), nil)
+	idx, _, err := New(context.TODO(), repo, restic.NewIDSet(), nil)
 	if err != nil {
 		t.Fatalf("New() returned error %v", err)
 	}
@@ -70,7 +70,7 @@ func TestIndexLoad(t *testing.T) {
 
 	validateIndex(t, repo, loadIdx)
 
-	newIdx, err := New(context.TODO(), repo, restic.NewIDSet(), nil)
+	newIdx, _, err := New(context.TODO(), repo, restic.NewIDSet(), nil)
 	if err != nil {
 		t.Fatalf("New() returned error %v", err)
 	}
@@ -134,7 +134,7 @@ func BenchmarkIndexNew(b *testing.B) {
 	b.ResetTimer()
 
 	for i := 0; i < b.N; i++ {
-		idx, err := New(context.TODO(), repo, restic.NewIDSet(), nil)
+		idx, _, err := New(context.TODO(), repo, restic.NewIDSet(), nil)
 
 		if err != nil {
 			b.Fatalf("New() returned error %v", err)
@@ -151,7 +151,7 @@ func BenchmarkIndexSave(b *testing.B) {
 	repo, cleanup := repository.TestRepository(b)
 	defer cleanup()
 
-	idx, err := New(context.TODO(), repo, restic.NewIDSet(), nil)
+	idx, _, err := New(context.TODO(), repo, restic.NewIDSet(), nil)
 	test.OK(b, err)
 
 	for i := 0; i < 8000; i++ {
@@ -184,7 +184,7 @@ func TestIndexDuplicateBlobs(t *testing.T) {
 	repo, cleanup := createFilledRepo(t, 3, 0.01)
 	defer cleanup()
 
-	idx, err := New(context.TODO(), repo, restic.NewIDSet(), nil)
+	idx, _, err := New(context.TODO(), repo, restic.NewIDSet(), nil)
 	if err != nil {
 		t.Fatal(err)
 	}

src/restic/repository/repack_test.go

@@ -147,7 +147,7 @@ func saveIndex(t *testing.T, repo restic.Repository) {
 }
 
 func rebuildIndex(t *testing.T, repo restic.Repository) {
-	idx, err := index.New(context.TODO(), repo, restic.NewIDSet(), nil)
+	idx, _, err := index.New(context.TODO(), repo, restic.NewIDSet(), nil)
 	if err != nil {
 		t.Fatal(err)
 	}