repair pack: add tests

internal/repository/repack_test.go

@@ -62,7 +62,7 @@ func createRandomBlobs(t testing.TB, repo restic.Repository, blobs int, pData fl
 	}
 }
 
-func createRandomWrongBlob(t testing.TB, repo restic.Repository) {
+func createRandomWrongBlob(t testing.TB, repo restic.Repository) restic.BlobHandle {
 	length := randomSize(10*1024, 1024*1024) // 10KiB to 1MiB of data
 	buf := make([]byte, length)
 	rand.Read(buf)
@@ -80,6 +80,7 @@ func createRandomWrongBlob(t testing.TB, repo restic.Repository) {
 	if err := repo.Flush(context.Background()); err != nil {
 		t.Fatalf("repo.Flush() returned error %v", err)
 	}
+	return restic.BlobHandle{ID: id, Type: restic.DataBlob}
 }
 
 // selectBlobs splits the list of all blobs randomly into two lists. A blob

internal/repository/repair_pack_test.go

@@ -0,0 +1,130 @@
+package repository_test
+
+import (
+	"context"
+	"math/rand"
+	"testing"
+	"time"
+
+	"github.com/restic/restic/internal/backend"
+	"github.com/restic/restic/internal/index"
+	"github.com/restic/restic/internal/repository"
+	"github.com/restic/restic/internal/restic"
+	"github.com/restic/restic/internal/test"
+	rtest "github.com/restic/restic/internal/test"
+	"github.com/restic/restic/internal/ui/progress"
+)
+
+func listBlobs(repo restic.Repository) restic.BlobSet {
+	blobs := restic.NewBlobSet()
+	repo.Index().Each(context.TODO(), func(pb restic.PackedBlob) {
+		blobs.Insert(pb.BlobHandle)
+	})
+	return blobs
+}
+
+func replaceFile(t *testing.T, repo restic.Repository, h backend.Handle, damage func([]byte) []byte) {
+	buf, err := backend.LoadAll(context.TODO(), nil, repo.Backend(), h)
+	test.OK(t, err)
+	buf = damage(buf)
+	test.OK(t, repo.Backend().Remove(context.TODO(), h))
+	test.OK(t, repo.Backend().Save(context.TODO(), h, backend.NewByteReader(buf, repo.Backend().Hasher())))
+}
+
+func TestRepairBrokenPack(t *testing.T) {
+	repository.TestAllVersions(t, testRepairBrokenPack)
+}
+
+func testRepairBrokenPack(t *testing.T, version uint) {
+	tests := []struct {
+		name   string
+		damage func(repo restic.Repository, packsBefore restic.IDSet) (restic.IDSet, restic.BlobSet)
+	}{
+		{
+			"valid pack",
+			func(repo restic.Repository, packsBefore restic.IDSet) (restic.IDSet, restic.BlobSet) {
+				return packsBefore, restic.NewBlobSet()
+			},
+		},
+		{
+			"broken pack",
+			func(repo restic.Repository, packsBefore restic.IDSet) (restic.IDSet, restic.BlobSet) {
+				wrongBlob := createRandomWrongBlob(t, repo)
+				damagedPacks := findPacksForBlobs(t, repo, restic.NewBlobSet(wrongBlob))
+				return damagedPacks, restic.NewBlobSet(wrongBlob)
+			},
+		},
+		{
+			"partially broken pack",
+			func(repo restic.Repository, packsBefore restic.IDSet) (restic.IDSet, restic.BlobSet) {
+				// damage one of the pack files
+				damagedID := packsBefore.List()[0]
+				replaceFile(t, repo, backend.Handle{Type: backend.PackFile, Name: damagedID.String()},
+					func(buf []byte) []byte {
+						buf[0] ^= 0xff
+						return buf
+					})
+
+				// find blob that starts at offset 0
+				var damagedBlob restic.BlobHandle
+				for blobs := range repo.Index().ListPacks(context.TODO(), restic.NewIDSet(damagedID)) {
+					for _, blob := range blobs.Blobs {
+						if blob.Offset == 0 {
+							damagedBlob = blob.BlobHandle
+						}
+					}
+				}
+
+				return restic.NewIDSet(damagedID), restic.NewBlobSet(damagedBlob)
+			},
+		}, {
+			"truncated pack",
+			func(repo restic.Repository, packsBefore restic.IDSet) (restic.IDSet, restic.BlobSet) {
+				// damage one of the pack files
+				damagedID := packsBefore.List()[0]
+				replaceFile(t, repo, backend.Handle{Type: backend.PackFile, Name: damagedID.String()},
+					func(buf []byte) []byte {
+						buf = buf[0:10]
+						return buf
+					})
+
+				// all blobs in the file are broken
+				damagedBlobs := restic.NewBlobSet()
+				for blobs := range repo.Index().ListPacks(context.TODO(), restic.NewIDSet(damagedID)) {
+					for _, blob := range blobs.Blobs {
+						damagedBlobs.Insert(blob.BlobHandle)
+					}
+				}
+				return restic.NewIDSet(damagedID), damagedBlobs
+			},
+		},
+	}
+
+	for _, test := range tests {
+		t.Run(test.name, func(t *testing.T) {
+			repo := repository.TestRepositoryWithVersion(t, version)
+
+			seed := time.Now().UnixNano()
+			rand.Seed(seed)
+			t.Logf("rand seed is %v", seed)
+
+			createRandomBlobs(t, repo, 5, 0.7)
+			packsBefore := listPacks(t, repo)
+			blobsBefore := listBlobs(repo)
+
+			toRepair, damagedBlobs := test.damage(repo, packsBefore)
+
+			rtest.OK(t, repository.RepairPacks(context.TODO(), repo, toRepair, &progress.NoopPrinter{}))
+			// reload index
+			rtest.OK(t, repo.SetIndex(index.NewMasterIndex()))
+			repo.LoadIndex(context.TODO(), nil)
+
+			packsAfter := listPacks(t, repo)
+			blobsAfter := listBlobs(repo)
+
+			rtest.Assert(t, len(packsAfter.Intersect(toRepair)) == 0, "some damaged packs were not removed")
+			rtest.Assert(t, len(packsBefore.Sub(toRepair).Sub(packsAfter)) == 0, "not-damaged packs were removed")
+			rtest.Assert(t, blobsBefore.Sub(damagedBlobs).Equals(blobsAfter), "diverging blob lists")
+		})
+	}
+}