restic/archiver_test.go
Philipp Serr 3d7f72311a Provoke unreferenced packs using fewer goroutines
TestParallelSaveWithDuplication has been reworked to provoke
unreferenced packs using fewer goroutines than before and create
only one bytes.Reader per blob. This reduces memory usage
significantly.

The following actions have been taken to keep the chance of provoking
unreferenced packs due to #358 high:
 * Interweaved processing of subsequent chunks
 * Delaying each goroutine by a few pseudo-randomly chosen nanoseconds
   (depending on the platform this will most probably only make the os
   yield execution to another thread): together with the interweaved
   processing of subsequent chunks, this ensures a minimalistic delay
   between processing of (some) duplicated chunks
 * Repeating the test 5 times with different seeds

On my test machine, the modified test provoked unreferenced packs 60
times in a row.
2015-12-28 18:33:26 +01:00

342 lines
7.8 KiB
Go

package restic_test
import (
"bytes"
"crypto/sha256"
"io"
"testing"
"time"
"github.com/restic/chunker"
"github.com/restic/restic"
"github.com/restic/restic/backend"
"github.com/restic/restic/checker"
"github.com/restic/restic/crypto"
"github.com/restic/restic/pack"
"github.com/restic/restic/repository"
. "github.com/restic/restic/test"
)
var testPol = chunker.Pol(0x3DA3358B4DC173)
type Rdr interface {
io.ReadSeeker
io.ReaderAt
}
type chunkedData struct {
buf []byte
chunks []*chunker.Chunk
}
func benchmarkChunkEncrypt(b testing.TB, buf, buf2 []byte, rd Rdr, key *crypto.Key) {
rd.Seek(0, 0)
ch := chunker.New(rd, testPol, sha256.New())
for {
chunk, err := ch.Next()
if err == io.EOF {
break
}
OK(b, err)
// reduce length of buf
buf = buf[:chunk.Length]
n, err := io.ReadFull(chunk.Reader(rd), buf)
OK(b, err)
Assert(b, uint(n) == chunk.Length, "invalid length: got %d, expected %d", n, chunk.Length)
_, err = crypto.Encrypt(key, buf2, buf)
OK(b, err)
}
}
func BenchmarkChunkEncrypt(b *testing.B) {
repo := SetupRepo()
defer TeardownRepo(repo)
data := Random(23, 10<<20) // 10MiB
rd := bytes.NewReader(data)
buf := make([]byte, chunker.MaxSize)
buf2 := make([]byte, chunker.MaxSize)
b.ResetTimer()
b.SetBytes(int64(len(data)))
for i := 0; i < b.N; i++ {
benchmarkChunkEncrypt(b, buf, buf2, rd, repo.Key())
}
}
func benchmarkChunkEncryptP(b *testing.PB, buf []byte, rd Rdr, key *crypto.Key) {
ch := chunker.New(rd, testPol, sha256.New())
for {
chunk, err := ch.Next()
if err == io.EOF {
break
}
// reduce length of chunkBuf
buf = buf[:chunk.Length]
io.ReadFull(chunk.Reader(rd), buf)
crypto.Encrypt(key, buf, buf)
}
}
func BenchmarkChunkEncryptParallel(b *testing.B) {
repo := SetupRepo()
defer TeardownRepo(repo)
data := Random(23, 10<<20) // 10MiB
buf := make([]byte, chunker.MaxSize)
b.ResetTimer()
b.SetBytes(int64(len(data)))
b.RunParallel(func(pb *testing.PB) {
for pb.Next() {
rd := bytes.NewReader(data)
benchmarkChunkEncryptP(pb, buf, rd, repo.Key())
}
})
}
func archiveDirectory(b testing.TB) {
repo := SetupRepo()
defer TeardownRepo(repo)
arch := restic.NewArchiver(repo)
_, id, err := arch.Snapshot(nil, []string{BenchArchiveDirectory}, nil)
OK(b, err)
b.Logf("snapshot archived as %v", id)
}
func TestArchiveDirectory(t *testing.T) {
if BenchArchiveDirectory == "" {
t.Skip("benchdir not set, skipping TestArchiveDirectory")
}
archiveDirectory(t)
}
func BenchmarkArchiveDirectory(b *testing.B) {
if BenchArchiveDirectory == "" {
b.Skip("benchdir not set, skipping BenchmarkArchiveDirectory")
}
for i := 0; i < b.N; i++ {
archiveDirectory(b)
}
}
func archiveWithDedup(t testing.TB) {
repo := SetupRepo()
defer TeardownRepo(repo)
if BenchArchiveDirectory == "" {
t.Skip("benchdir not set, skipping TestArchiverDedup")
}
var cnt struct {
before, after, after2 struct {
packs, dataBlobs, treeBlobs uint
}
}
// archive a few files
sn := SnapshotDir(t, repo, BenchArchiveDirectory, nil)
t.Logf("archived snapshot %v", sn.ID().Str())
// get archive stats
cnt.before.packs = repo.Count(backend.Data)
cnt.before.dataBlobs = repo.Index().Count(pack.Data)
cnt.before.treeBlobs = repo.Index().Count(pack.Tree)
t.Logf("packs %v, data blobs %v, tree blobs %v",
cnt.before.packs, cnt.before.dataBlobs, cnt.before.treeBlobs)
// archive the same files again, without parent snapshot
sn2 := SnapshotDir(t, repo, BenchArchiveDirectory, nil)
t.Logf("archived snapshot %v", sn2.ID().Str())
// get archive stats again
cnt.after.packs = repo.Count(backend.Data)
cnt.after.dataBlobs = repo.Index().Count(pack.Data)
cnt.after.treeBlobs = repo.Index().Count(pack.Tree)
t.Logf("packs %v, data blobs %v, tree blobs %v",
cnt.after.packs, cnt.after.dataBlobs, cnt.after.treeBlobs)
// if there are more data blobs, something is wrong
if cnt.after.dataBlobs > cnt.before.dataBlobs {
t.Fatalf("TestArchiverDedup: too many data blobs in repository: before %d, after %d",
cnt.before.dataBlobs, cnt.after.dataBlobs)
}
// archive the same files again, with a parent snapshot
sn3 := SnapshotDir(t, repo, BenchArchiveDirectory, sn2.ID())
t.Logf("archived snapshot %v, parent %v", sn3.ID().Str(), sn2.ID().Str())
// get archive stats again
cnt.after2.packs = repo.Count(backend.Data)
cnt.after2.dataBlobs = repo.Index().Count(pack.Data)
cnt.after2.treeBlobs = repo.Index().Count(pack.Tree)
t.Logf("packs %v, data blobs %v, tree blobs %v",
cnt.after2.packs, cnt.after2.dataBlobs, cnt.after2.treeBlobs)
// if there are more data blobs, something is wrong
if cnt.after2.dataBlobs > cnt.before.dataBlobs {
t.Fatalf("TestArchiverDedup: too many data blobs in repository: before %d, after %d",
cnt.before.dataBlobs, cnt.after2.dataBlobs)
}
}
func TestArchiveDedup(t *testing.T) {
archiveWithDedup(t)
}
func BenchmarkLoadTree(t *testing.B) {
repo := SetupRepo()
defer TeardownRepo(repo)
if BenchArchiveDirectory == "" {
t.Skip("benchdir not set, skipping TestArchiverDedup")
}
// archive a few files
arch := restic.NewArchiver(repo)
sn, _, err := arch.Snapshot(nil, []string{BenchArchiveDirectory}, nil)
OK(t, err)
t.Logf("archived snapshot %v", sn.ID())
list := make([]backend.ID, 0, 10)
done := make(chan struct{})
for _, idx := range repo.Index().All() {
for blob := range idx.Each(done) {
if blob.Type != pack.Tree {
continue
}
list = append(list, blob.ID)
if len(list) == cap(list) {
close(done)
break
}
}
}
// start benchmark
t.ResetTimer()
for i := 0; i < t.N; i++ {
for _, id := range list {
_, err := restic.LoadTree(repo, id)
OK(t, err)
}
}
}
// Saves several identical chunks concurrently and later checks that there are no
// unreferenced packs in the repository. See also #292 and #358.
func TestParallelSaveWithDuplication(t *testing.T) {
for seed := 0; seed < 5; seed++ {
testParallelSaveWithDuplication(t, seed)
}
}
func testParallelSaveWithDuplication(t *testing.T, seed int) {
repo := SetupRepo()
defer TeardownRepo(repo)
dataSizeMb := 92
duplication := 7
arch := restic.NewArchiver(repo)
data, chunks := getRandomData(seed, dataSizeMb*1024*1024)
reader := bytes.NewReader(data)
errChannels := [](<-chan error){}
// interweaved processing of subsequent chunks
maxParallel := 2*duplication - 1
barrier := make(chan struct{}, maxParallel)
for _, c := range chunks {
for dupIdx := 0; dupIdx < duplication; dupIdx++ {
errChan := make(chan error)
errChannels = append(errChannels, errChan)
go func(reader *bytes.Reader, c *chunker.Chunk, errChan chan<- error) {
barrier <- struct{}{}
hash := c.Digest
id := backend.ID{}
copy(id[:], hash)
time.Sleep(time.Duration(hash[0]))
err := arch.Save(pack.Data, id, c.Length, c.Reader(reader))
<-barrier
errChan <- err
}(reader, c, errChan)
}
}
for _, errChan := range errChannels {
OK(t, <-errChan)
}
OK(t, repo.Flush())
OK(t, repo.SaveIndex())
chkr := createAndInitChecker(t, repo)
assertNoUnreferencedPacks(t, chkr)
}
func getRandomData(seed int, size int) ([]byte, []*chunker.Chunk) {
buf := Random(seed, size)
chunks := []*chunker.Chunk{}
chunker := chunker.New(bytes.NewReader(buf), testPol, sha256.New())
for {
c, err := chunker.Next()
if err == io.EOF {
break
}
chunks = append(chunks, c)
}
return buf, chunks
}
func createAndInitChecker(t *testing.T, repo *repository.Repository) *checker.Checker {
chkr := checker.New(repo)
hints, errs := chkr.LoadIndex()
if len(errs) > 0 {
t.Fatalf("expected no errors, got %v: %v", len(errs), errs)
}
if len(hints) > 0 {
t.Errorf("expected no hints, got %v: %v", len(hints), hints)
}
return chkr
}
func assertNoUnreferencedPacks(t *testing.T, chkr *checker.Checker) {
done := make(chan struct{})
defer close(done)
errChan := make(chan error)
go chkr.Packs(errChan, done)
for err := range errChan {
OK(t, err)
}
}