mirror of
https://github.com/octoleo/restic.git
synced 2024-11-15 17:47:21 +00:00
7132df529e
A compressed index is only about one third the size of an uncompressed one. Thus increase the number of entries in an index to avoid cluttering the repository with small indexes.
615 lines
14 KiB
Go
615 lines
14 KiB
Go
package repository_test
|
|
|
|
import (
|
|
"bytes"
|
|
"context"
|
|
"crypto/sha256"
|
|
"encoding/json"
|
|
"fmt"
|
|
"io"
|
|
"math/rand"
|
|
"os"
|
|
"path/filepath"
|
|
"strings"
|
|
"testing"
|
|
"time"
|
|
|
|
"github.com/google/go-cmp/cmp"
|
|
"github.com/restic/restic/internal/archiver"
|
|
"github.com/restic/restic/internal/crypto"
|
|
"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"
|
|
)
|
|
|
|
var testSizes = []int{5, 23, 2<<18 + 23, 1 << 20}
|
|
|
|
var rnd = rand.New(rand.NewSource(time.Now().UnixNano()))
|
|
|
|
func TestSave(t *testing.T) {
|
|
repo, cleanup := repository.TestRepository(t)
|
|
defer cleanup()
|
|
|
|
for _, size := range testSizes {
|
|
data := make([]byte, size)
|
|
_, err := io.ReadFull(rnd, data)
|
|
rtest.OK(t, err)
|
|
|
|
id := restic.Hash(data)
|
|
|
|
// save
|
|
sid, _, err := repo.SaveBlob(context.TODO(), restic.DataBlob, data, restic.ID{}, false)
|
|
rtest.OK(t, err)
|
|
|
|
rtest.Equals(t, id, sid)
|
|
|
|
rtest.OK(t, repo.Flush(context.Background()))
|
|
// rtest.OK(t, repo.SaveIndex())
|
|
|
|
// read back
|
|
buf, err := repo.LoadBlob(context.TODO(), restic.DataBlob, id, nil)
|
|
rtest.OK(t, err)
|
|
rtest.Equals(t, size, len(buf))
|
|
|
|
rtest.Assert(t, len(buf) == len(data),
|
|
"number of bytes read back does not match: expected %d, got %d",
|
|
len(data), len(buf))
|
|
|
|
rtest.Assert(t, bytes.Equal(buf, data),
|
|
"data does not match: expected %02x, got %02x",
|
|
data, buf)
|
|
}
|
|
}
|
|
|
|
func TestSaveFrom(t *testing.T) {
|
|
repo, cleanup := repository.TestRepository(t)
|
|
defer cleanup()
|
|
|
|
for _, size := range testSizes {
|
|
data := make([]byte, size)
|
|
_, err := io.ReadFull(rnd, data)
|
|
rtest.OK(t, err)
|
|
|
|
id := restic.Hash(data)
|
|
|
|
// save
|
|
id2, _, err := repo.SaveBlob(context.TODO(), restic.DataBlob, data, id, false)
|
|
rtest.OK(t, err)
|
|
rtest.Equals(t, id, id2)
|
|
|
|
rtest.OK(t, repo.Flush(context.Background()))
|
|
|
|
// read back
|
|
buf, err := repo.LoadBlob(context.TODO(), restic.DataBlob, id, nil)
|
|
rtest.OK(t, err)
|
|
rtest.Equals(t, size, len(buf))
|
|
|
|
rtest.Assert(t, len(buf) == len(data),
|
|
"number of bytes read back does not match: expected %d, got %d",
|
|
len(data), len(buf))
|
|
|
|
rtest.Assert(t, bytes.Equal(buf, data),
|
|
"data does not match: expected %02x, got %02x",
|
|
data, buf)
|
|
}
|
|
}
|
|
|
|
func BenchmarkSaveAndEncrypt(t *testing.B) {
|
|
repo, cleanup := repository.TestRepository(t)
|
|
defer cleanup()
|
|
|
|
size := 4 << 20 // 4MiB
|
|
|
|
data := make([]byte, size)
|
|
_, err := io.ReadFull(rnd, data)
|
|
rtest.OK(t, err)
|
|
|
|
id := restic.ID(sha256.Sum256(data))
|
|
|
|
t.ReportAllocs()
|
|
t.ResetTimer()
|
|
t.SetBytes(int64(size))
|
|
|
|
for i := 0; i < t.N; i++ {
|
|
_, _, err = repo.SaveBlob(context.TODO(), restic.DataBlob, data, id, true)
|
|
rtest.OK(t, err)
|
|
}
|
|
}
|
|
|
|
func TestLoadTree(t *testing.T) {
|
|
repo, cleanup := repository.TestRepository(t)
|
|
defer cleanup()
|
|
|
|
if rtest.BenchArchiveDirectory == "" {
|
|
t.Skip("benchdir not set, skipping")
|
|
}
|
|
|
|
// archive a few files
|
|
sn := archiver.TestSnapshot(t, repo, rtest.BenchArchiveDirectory, nil)
|
|
rtest.OK(t, repo.Flush(context.Background()))
|
|
|
|
_, err := repo.LoadTree(context.TODO(), *sn.Tree)
|
|
rtest.OK(t, err)
|
|
}
|
|
|
|
func BenchmarkLoadTree(t *testing.B) {
|
|
repo, cleanup := repository.TestRepository(t)
|
|
defer cleanup()
|
|
|
|
if rtest.BenchArchiveDirectory == "" {
|
|
t.Skip("benchdir not set, skipping")
|
|
}
|
|
|
|
// archive a few files
|
|
sn := archiver.TestSnapshot(t, repo, rtest.BenchArchiveDirectory, nil)
|
|
rtest.OK(t, repo.Flush(context.Background()))
|
|
|
|
t.ResetTimer()
|
|
|
|
for i := 0; i < t.N; i++ {
|
|
_, err := repo.LoadTree(context.TODO(), *sn.Tree)
|
|
rtest.OK(t, err)
|
|
}
|
|
}
|
|
|
|
func TestLoadBlob(t *testing.T) {
|
|
repo, cleanup := repository.TestRepository(t)
|
|
defer cleanup()
|
|
|
|
length := 1000000
|
|
buf := restic.NewBlobBuffer(length)
|
|
_, err := io.ReadFull(rnd, buf)
|
|
rtest.OK(t, err)
|
|
|
|
id, _, err := repo.SaveBlob(context.TODO(), restic.DataBlob, buf, restic.ID{}, false)
|
|
rtest.OK(t, err)
|
|
rtest.OK(t, repo.Flush(context.Background()))
|
|
|
|
base := restic.CiphertextLength(length)
|
|
for _, testlength := range []int{0, base - 20, base - 1, base, base + 7, base + 15, base + 1000} {
|
|
buf = make([]byte, 0, testlength)
|
|
buf, err := repo.LoadBlob(context.TODO(), restic.DataBlob, id, buf)
|
|
if err != nil {
|
|
t.Errorf("LoadBlob() returned an error for buffer size %v: %v", testlength, err)
|
|
continue
|
|
}
|
|
|
|
if len(buf) != length {
|
|
t.Errorf("LoadBlob() returned the wrong number of bytes: want %v, got %v", length, len(buf))
|
|
continue
|
|
}
|
|
}
|
|
}
|
|
|
|
func BenchmarkLoadBlob(b *testing.B) {
|
|
repo, cleanup := repository.TestRepository(b)
|
|
defer cleanup()
|
|
|
|
length := 1000000
|
|
buf := restic.NewBlobBuffer(length)
|
|
_, err := io.ReadFull(rnd, buf)
|
|
rtest.OK(b, err)
|
|
|
|
id, _, err := repo.SaveBlob(context.TODO(), restic.DataBlob, buf, restic.ID{}, false)
|
|
rtest.OK(b, err)
|
|
rtest.OK(b, repo.Flush(context.Background()))
|
|
|
|
b.ResetTimer()
|
|
b.SetBytes(int64(length))
|
|
|
|
for i := 0; i < b.N; i++ {
|
|
var err error
|
|
buf, err = repo.LoadBlob(context.TODO(), restic.DataBlob, id, buf)
|
|
|
|
// Checking the SHA-256 with restic.Hash can make up 38% of the time
|
|
// spent in this loop, so pause the timer.
|
|
b.StopTimer()
|
|
rtest.OK(b, err)
|
|
if len(buf) != length {
|
|
b.Errorf("wanted %d bytes, got %d", length, len(buf))
|
|
}
|
|
|
|
id2 := restic.Hash(buf)
|
|
if !id.Equal(id2) {
|
|
b.Errorf("wrong data returned, wanted %v, got %v", id.Str(), id2.Str())
|
|
}
|
|
b.StartTimer()
|
|
}
|
|
}
|
|
|
|
func BenchmarkLoadUnpacked(b *testing.B) {
|
|
repo, cleanup := repository.TestRepository(b)
|
|
defer cleanup()
|
|
|
|
length := 1000000
|
|
buf := restic.NewBlobBuffer(length)
|
|
_, err := io.ReadFull(rnd, buf)
|
|
rtest.OK(b, err)
|
|
|
|
dataID := restic.Hash(buf)
|
|
|
|
storageID, err := repo.SaveUnpacked(context.TODO(), restic.PackFile, buf)
|
|
rtest.OK(b, err)
|
|
// rtest.OK(b, repo.Flush())
|
|
|
|
b.ResetTimer()
|
|
b.SetBytes(int64(length))
|
|
|
|
for i := 0; i < b.N; i++ {
|
|
data, err := repo.LoadUnpacked(context.TODO(), nil, restic.PackFile, storageID)
|
|
rtest.OK(b, err)
|
|
|
|
// See comment in BenchmarkLoadBlob.
|
|
b.StopTimer()
|
|
if len(data) != length {
|
|
b.Errorf("wanted %d bytes, got %d", length, len(data))
|
|
}
|
|
|
|
id2 := restic.Hash(data)
|
|
if !dataID.Equal(id2) {
|
|
b.Errorf("wrong data returned, wanted %v, got %v", storageID.Str(), id2.Str())
|
|
}
|
|
b.StartTimer()
|
|
}
|
|
}
|
|
|
|
func TestLoadJSONUnpacked(t *testing.T) {
|
|
repo, cleanup := repository.TestRepository(t)
|
|
defer cleanup()
|
|
|
|
if rtest.BenchArchiveDirectory == "" {
|
|
t.Skip("benchdir not set, skipping")
|
|
}
|
|
|
|
// archive a snapshot
|
|
sn := restic.Snapshot{}
|
|
sn.Hostname = "foobar"
|
|
sn.Username = "test!"
|
|
|
|
id, err := repo.SaveJSONUnpacked(context.TODO(), restic.SnapshotFile, &sn)
|
|
rtest.OK(t, err)
|
|
|
|
var sn2 restic.Snapshot
|
|
|
|
// restore
|
|
err = repo.LoadJSONUnpacked(context.TODO(), restic.SnapshotFile, id, &sn2)
|
|
rtest.OK(t, err)
|
|
|
|
rtest.Equals(t, sn.Hostname, sn2.Hostname)
|
|
rtest.Equals(t, sn.Username, sn2.Username)
|
|
|
|
var cf restic.Config
|
|
|
|
// load and check Config
|
|
err = repo.LoadJSONUnpacked(context.TODO(), restic.ConfigFile, id, &cf)
|
|
rtest.OK(t, err)
|
|
|
|
rtest.Equals(t, cf.ChunkerPolynomial, repository.TestChunkerPol)
|
|
}
|
|
|
|
var repoFixture = filepath.Join("testdata", "test-repo.tar.gz")
|
|
|
|
func TestRepositoryLoadIndex(t *testing.T) {
|
|
repodir, cleanup := rtest.Env(t, repoFixture)
|
|
defer cleanup()
|
|
|
|
repo := repository.TestOpenLocal(t, repodir)
|
|
rtest.OK(t, repo.LoadIndex(context.TODO()))
|
|
}
|
|
|
|
// loadIndex loads the index id from backend and returns it.
|
|
func loadIndex(ctx context.Context, repo restic.Repository, id restic.ID) (*repository.Index, error) {
|
|
buf, err := repo.LoadUnpacked(ctx, nil, restic.IndexFile, id)
|
|
if err != nil {
|
|
return nil, err
|
|
}
|
|
|
|
idx, oldFormat, err := repository.DecodeIndex(buf, id)
|
|
if oldFormat {
|
|
fmt.Fprintf(os.Stderr, "index %v has old format\n", id.Str())
|
|
}
|
|
return idx, err
|
|
}
|
|
|
|
func BenchmarkLoadIndex(b *testing.B) {
|
|
repository.TestUseLowSecurityKDFParameters(b)
|
|
|
|
repo, cleanup := repository.TestRepository(b)
|
|
defer cleanup()
|
|
|
|
idx := repository.NewIndex()
|
|
|
|
for i := 0; i < 5000; i++ {
|
|
idx.Store(restic.PackedBlob{
|
|
Blob: restic.Blob{
|
|
BlobHandle: restic.NewRandomBlobHandle(),
|
|
Length: 1234,
|
|
Offset: 1235,
|
|
},
|
|
PackID: restic.NewRandomID(),
|
|
})
|
|
}
|
|
|
|
id, err := repository.SaveIndex(context.TODO(), repo, idx)
|
|
rtest.OK(b, err)
|
|
|
|
b.Logf("index saved as %v (%v entries)", id.Str(), idx.Count(restic.DataBlob))
|
|
fi, err := repo.Backend().Stat(context.TODO(), restic.Handle{Type: restic.IndexFile, Name: id.String()})
|
|
rtest.OK(b, err)
|
|
b.Logf("filesize is %v", fi.Size)
|
|
|
|
b.ResetTimer()
|
|
|
|
for i := 0; i < b.N; i++ {
|
|
_, err := loadIndex(context.TODO(), repo, id)
|
|
rtest.OK(b, err)
|
|
}
|
|
}
|
|
|
|
// saveRandomDataBlobs generates random data blobs and saves them to the repository.
|
|
func saveRandomDataBlobs(t testing.TB, repo restic.Repository, num int, sizeMax int) {
|
|
for i := 0; i < num; i++ {
|
|
size := rand.Int() % sizeMax
|
|
|
|
buf := make([]byte, size)
|
|
_, err := io.ReadFull(rnd, buf)
|
|
rtest.OK(t, err)
|
|
|
|
_, _, err = repo.SaveBlob(context.TODO(), restic.DataBlob, buf, restic.ID{}, false)
|
|
rtest.OK(t, err)
|
|
}
|
|
}
|
|
|
|
func TestRepositoryIncrementalIndex(t *testing.T) {
|
|
r, cleanup := repository.TestRepository(t)
|
|
defer cleanup()
|
|
|
|
repo := r.(*repository.Repository)
|
|
|
|
repository.IndexFull = func(*repository.Index, bool) bool { return true }
|
|
|
|
// add 15 packs
|
|
for j := 0; j < 5; j++ {
|
|
// add 3 packs, write intermediate index
|
|
for i := 0; i < 3; i++ {
|
|
saveRandomDataBlobs(t, repo, 5, 1<<15)
|
|
rtest.OK(t, repo.FlushPacks(context.Background()))
|
|
}
|
|
|
|
rtest.OK(t, repo.SaveFullIndex(context.TODO()))
|
|
}
|
|
|
|
// add another 5 packs
|
|
for i := 0; i < 5; i++ {
|
|
saveRandomDataBlobs(t, repo, 5, 1<<15)
|
|
rtest.OK(t, repo.FlushPacks(context.Background()))
|
|
}
|
|
|
|
// save final index
|
|
rtest.OK(t, repo.SaveIndex(context.TODO()))
|
|
|
|
packEntries := make(map[restic.ID]map[restic.ID]struct{})
|
|
|
|
err := repo.List(context.TODO(), restic.IndexFile, func(id restic.ID, size int64) error {
|
|
idx, err := loadIndex(context.TODO(), repo, id)
|
|
rtest.OK(t, err)
|
|
|
|
for pb := range idx.Each(context.TODO()) {
|
|
if _, ok := packEntries[pb.PackID]; !ok {
|
|
packEntries[pb.PackID] = make(map[restic.ID]struct{})
|
|
}
|
|
|
|
packEntries[pb.PackID][id] = struct{}{}
|
|
}
|
|
return nil
|
|
})
|
|
if err != nil {
|
|
t.Fatal(err)
|
|
}
|
|
|
|
for packID, ids := range packEntries {
|
|
if len(ids) > 1 {
|
|
t.Errorf("pack %v listed in %d indexes\n", packID, len(ids))
|
|
}
|
|
}
|
|
|
|
}
|
|
|
|
// buildPackfileWithoutHeader returns a manually built pack file without a header.
|
|
func buildPackfileWithoutHeader(t testing.TB, blobSizes []int, key *crypto.Key) (blobs []restic.Blob, packfile []byte) {
|
|
var offset uint
|
|
for i, size := range blobSizes {
|
|
plaintext := test.Random(800+i, size)
|
|
|
|
// we use a deterministic nonce here so the whole process is
|
|
// deterministic, last byte is the blob index
|
|
var nonce = []byte{
|
|
0x15, 0x98, 0xc0, 0xf7, 0xb9, 0x65, 0x97, 0x74,
|
|
0x12, 0xdc, 0xd3, 0x62, 0xa9, 0x6e, 0x20, byte(i),
|
|
}
|
|
|
|
before := len(packfile)
|
|
packfile = append(packfile, nonce...)
|
|
packfile = key.Seal(packfile, nonce, plaintext, nil)
|
|
after := len(packfile)
|
|
|
|
ciphertextLength := after - before
|
|
|
|
blobs = append(blobs, restic.Blob{
|
|
BlobHandle: restic.BlobHandle{
|
|
ID: restic.Hash(plaintext),
|
|
Type: restic.DataBlob,
|
|
},
|
|
Length: uint(ciphertextLength),
|
|
Offset: offset,
|
|
})
|
|
|
|
offset = uint(len(packfile))
|
|
}
|
|
|
|
return blobs, packfile
|
|
}
|
|
|
|
func TestStreamPack(t *testing.T) {
|
|
// always use the same key for deterministic output
|
|
const jsonKey = `{"mac":{"k":"eQenuI8adktfzZMuC8rwdA==","r":"k8cfAly2qQSky48CQK7SBA=="},"encrypt":"MKO9gZnRiQFl8mDUurSDa9NMjiu9MUifUrODTHS05wo="}`
|
|
|
|
var key crypto.Key
|
|
err := json.Unmarshal([]byte(jsonKey), &key)
|
|
if err != nil {
|
|
t.Fatal(err)
|
|
}
|
|
|
|
blobSizes := []int{
|
|
10,
|
|
5231,
|
|
18812,
|
|
123123,
|
|
12301,
|
|
892242,
|
|
28616,
|
|
13351,
|
|
252287,
|
|
188883,
|
|
2522811,
|
|
18883,
|
|
}
|
|
|
|
packfileBlobs, packfile := buildPackfileWithoutHeader(t, blobSizes, &key)
|
|
|
|
load := func(ctx context.Context, h restic.Handle, length int, offset int64, fn func(rd io.Reader) error) error {
|
|
data := packfile
|
|
|
|
if offset > int64(len(data)) {
|
|
offset = 0
|
|
length = 0
|
|
}
|
|
data = data[offset:]
|
|
|
|
if length > len(data) {
|
|
length = len(data)
|
|
}
|
|
|
|
data = data[:length]
|
|
|
|
return fn(bytes.NewReader(data))
|
|
|
|
}
|
|
|
|
// first, test regular usage
|
|
t.Run("regular", func(t *testing.T) {
|
|
tests := []struct {
|
|
blobs []restic.Blob
|
|
}{
|
|
{packfileBlobs[1:2]},
|
|
{packfileBlobs[2:5]},
|
|
{packfileBlobs[2:8]},
|
|
{[]restic.Blob{
|
|
packfileBlobs[0],
|
|
packfileBlobs[8],
|
|
packfileBlobs[4],
|
|
}},
|
|
{[]restic.Blob{
|
|
packfileBlobs[0],
|
|
packfileBlobs[len(packfileBlobs)-1],
|
|
}},
|
|
}
|
|
|
|
for _, test := range tests {
|
|
t.Run("", func(t *testing.T) {
|
|
ctx, cancel := context.WithCancel(context.Background())
|
|
defer cancel()
|
|
|
|
gotBlobs := make(map[restic.ID]int)
|
|
|
|
handleBlob := func(blob restic.BlobHandle, buf []byte, err error) error {
|
|
gotBlobs[blob.ID]++
|
|
|
|
id := restic.Hash(buf)
|
|
if !id.Equal(blob.ID) {
|
|
t.Fatalf("wrong id %v for blob %s returned", id, blob.ID)
|
|
}
|
|
|
|
return err
|
|
}
|
|
|
|
wantBlobs := make(map[restic.ID]int)
|
|
for _, blob := range test.blobs {
|
|
wantBlobs[blob.ID] = 1
|
|
}
|
|
|
|
err = repository.StreamPack(ctx, load, &key, restic.ID{}, test.blobs, handleBlob)
|
|
if err != nil {
|
|
t.Fatal(err)
|
|
}
|
|
|
|
if !cmp.Equal(wantBlobs, gotBlobs) {
|
|
t.Fatal(cmp.Diff(wantBlobs, gotBlobs))
|
|
}
|
|
})
|
|
}
|
|
})
|
|
|
|
// next, test invalid uses, which should return an error
|
|
t.Run("invalid", func(t *testing.T) {
|
|
tests := []struct {
|
|
blobs []restic.Blob
|
|
err string
|
|
}{
|
|
{
|
|
// pass one blob several times
|
|
blobs: []restic.Blob{
|
|
packfileBlobs[3],
|
|
packfileBlobs[8],
|
|
packfileBlobs[3],
|
|
packfileBlobs[4],
|
|
},
|
|
err: "overlapping blobs in pack",
|
|
},
|
|
|
|
{
|
|
// pass something that's not a valid blob in the current pack file
|
|
blobs: []restic.Blob{
|
|
{
|
|
Offset: 123,
|
|
Length: 20000,
|
|
},
|
|
},
|
|
err: "ciphertext verification failed",
|
|
},
|
|
|
|
{
|
|
// pass a blob that's too small
|
|
blobs: []restic.Blob{
|
|
{
|
|
Offset: 123,
|
|
Length: 10,
|
|
},
|
|
},
|
|
err: "invalid blob length",
|
|
},
|
|
}
|
|
|
|
for _, test := range tests {
|
|
t.Run("", func(t *testing.T) {
|
|
ctx, cancel := context.WithCancel(context.Background())
|
|
defer cancel()
|
|
|
|
handleBlob := func(blob restic.BlobHandle, buf []byte, err error) error {
|
|
return err
|
|
}
|
|
|
|
err = repository.StreamPack(ctx, load, &key, restic.ID{}, test.blobs, handleBlob)
|
|
if err == nil {
|
|
t.Fatalf("wanted error %v, got nil", test.err)
|
|
}
|
|
|
|
if !strings.Contains(err.Error(), test.err) {
|
|
t.Fatalf("wrong error returned, it should contain %q but was %q", test.err, err)
|
|
}
|
|
})
|
|
}
|
|
})
|
|
}
|