Add support for keys with ttl (#177)

* ttl support first commit

* imports fix

* put api args correction

* put options added

* upgrade method added

* upgrade log added

* v0 to v1 migration script added

* error assertion added

* temp migration dir fix

Co-authored-by: yash <yash.chandra@grabpay.com>

scripts/migrations/v0_to_v1.go

@@ -0,0 +1,159 @@
+package migrations
+
+import (
+	"encoding/binary"
+	"fmt"
+	"io"
+	"io/ioutil"
+	"os"
+	"path"
+	"path/filepath"
+
+	"github.com/prologic/bitcask/internal"
+)
+
+const (
+	keySize                 = 4
+	valueSize               = 8
+	checksumSize            = 4
+	ttlSize                 = 8
+	defaultDatafileFilename = "%09d.data"
+)
+
+func ApplyV0ToV1(dir string, maxDatafileSize int) error {
+	temp, err := prepare(dir)
+	if err != nil {
+		return err
+	}
+	defer os.RemoveAll(temp)
+	err = apply(dir, temp, maxDatafileSize)
+	if err != nil {
+		return err
+	}
+	return cleanup(dir, temp)
+}
+
+func prepare(dir string) (string, error) {
+	return ioutil.TempDir(dir, "migration")
+}
+
+func apply(dir, temp string, maxDatafileSize int) error {
+	datafilesPath, err := internal.GetDatafiles(dir)
+	if err != nil {
+		return err
+	}
+	var id, newOffset int
+	datafile, err := getNewDatafile(temp, id)
+	if err != nil {
+		return err
+	}
+	id++
+	for _, p := range datafilesPath {
+		df, err := os.Open(p)
+		if err != nil {
+			return err
+		}
+		var off int64
+		for {
+			entry, err := getSingleEntry(df, off)
+			if err == io.EOF {
+				break
+			}
+			if err != nil {
+				return err
+			}
+			if newOffset+len(entry) > maxDatafileSize {
+				err = datafile.Sync()
+				if err != nil {
+					return err
+				}
+				datafile, err = getNewDatafile(temp, id)
+				if err != nil {
+					return err
+				}
+				id++
+				newOffset = 0
+			}
+			newEntry := make([]byte, len(entry)+ttlSize)
+			copy(newEntry[:len(entry)], entry)
+			n, err := datafile.Write(newEntry)
+			if err != nil {
+				return err
+			}
+			newOffset += n
+			off += int64(len(entry))
+		}
+	}
+	return datafile.Sync()
+}
+
+func cleanup(dir, temp string) error {
+	files, err := ioutil.ReadDir(dir)
+	if err != nil {
+		return err
+	}
+	for _, file := range files {
+		if !file.IsDir() {
+			err := os.RemoveAll(path.Join([]string{dir, file.Name()}...))
+			if err != nil {
+				return err
+			}
+		}
+	}
+	files, err = ioutil.ReadDir(temp)
+	if err != nil {
+		return err
+	}
+	for _, file := range files {
+		err := os.Rename(
+			path.Join([]string{temp, file.Name()}...),
+			path.Join([]string{dir, file.Name()}...),
+		)
+		if err != nil {
+			return err
+		}
+	}
+	return nil
+}
+
+func getNewDatafile(path string, id int) (*os.File, error) {
+	fn := filepath.Join(path, fmt.Sprintf(defaultDatafileFilename, id))
+	return os.OpenFile(fn, os.O_WRONLY|os.O_APPEND|os.O_CREATE, 0640)
+}
+
+func getSingleEntry(f *os.File, offset int64) ([]byte, error) {
+	prefixBuf, err := readPrefix(f, offset)
+	if err != nil {
+		return nil, err
+	}
+	actualKeySize, actualValueSize := getKeyValueSize(prefixBuf)
+	entryBuf, err := read(f, uint64(actualKeySize)+actualValueSize+checksumSize, offset+keySize+valueSize)
+	if err != nil {
+		return nil, err
+	}
+	return append(prefixBuf, entryBuf...), nil
+}
+
+func readPrefix(f *os.File, offset int64) ([]byte, error) {
+	prefixBuf := make([]byte, keySize+valueSize)
+	_, err := f.ReadAt(prefixBuf, offset)
+	if err != nil {
+		return nil, err
+	}
+	return prefixBuf, nil
+}
+
+func read(f *os.File, bufSize uint64, offset int64) ([]byte, error) {
+	buf := make([]byte, bufSize)
+	_, err := f.ReadAt(buf, offset)
+	if err != nil {
+		return nil, err
+	}
+	return buf, nil
+}
+
+func getKeyValueSize(buf []byte) (uint32, uint64) {
+	actualKeySize := binary.BigEndian.Uint32(buf[:keySize])
+	actualValueSize := binary.BigEndian.Uint64(buf[keySize:])
+	return actualKeySize, actualValueSize
+}

scripts/migrations/v0_to_v1_test.go

@@ -0,0 +1,58 @@
+package migrations
+
+import (
+	"encoding/binary"
+	"encoding/hex"
+	"io"
+	"io/ioutil"
+	"os"
+	"path/filepath"
+	"testing"
+
+	"github.com/stretchr/testify/assert"
+)
+
+func Test_ApplyV0ToV1(t *testing.T) {
+	assert := assert.New(t)
+	testdir, err := ioutil.TempDir("/tmp", "bitcask")
+	assert.NoError(err)
+	defer os.RemoveAll(testdir)
+	w0, err := os.OpenFile(filepath.Join(testdir, "000000000.data"), os.O_CREATE|os.O_APPEND|os.O_WRONLY, 0640)
+	assert.NoError(err)
+	w1, err := os.OpenFile(filepath.Join(testdir, "000000001.data"), os.O_CREATE|os.O_APPEND|os.O_WRONLY, 0640)
+	assert.NoError(err)
+	w2, err := os.OpenFile(filepath.Join(testdir, "000000002.data"), os.O_CREATE|os.O_APPEND|os.O_WRONLY, 0640)
+	assert.NoError(err)
+	defer w0.Close()
+	defer w1.Close()
+	defer w2.Close()
+	buf := make([]byte, 104)
+	binary.BigEndian.PutUint32(buf[:4], 5)
+	binary.BigEndian.PutUint64(buf[4:12], 7)
+	copy(buf[12:28], "mykeymyvalue0AAA")
+	binary.BigEndian.PutUint32(buf[28:32], 3)
+	binary.BigEndian.PutUint64(buf[32:40], 5)
+	copy(buf[40:52], "keyvalue0BBB")
+	_, err = w0.Write(buf[:52])
+	assert.NoError(err)
+	_, err = w1.Write(buf[:52])
+	assert.NoError(err)
+	_, err = w2.Write(buf[:52])
+	assert.NoError(err)
+	err = ApplyV0ToV1(testdir, 104)
+	assert.NoError(err)
+	r0, err := os.Open(filepath.Join(testdir, "000000000.data"))
+	assert.NoError(err)
+	defer r0.Close()
+	n, err := io.ReadFull(r0, buf)
+	assert.NoError(err)
+	assert.Equal(104, n)
+	assert.Equal("0000000500000000000000076d796b65796d7976616c75653041414100000000000000000000000300000000000000056b657976616c75653042424200000000000000000000000500000000000000076d796b65796d7976616c7565304141410000000000000000", hex.EncodeToString(buf))
+	r1, err := os.Open(filepath.Join(testdir, "000000001.data"))
+	assert.NoError(err)
+	defer r1.Close()
+	n, err = io.ReadFull(r1, buf[:100])
+	assert.NoError(err)
+	assert.Equal(100, n)
+	assert.Equal("0000000300000000000000056b657976616c75653042424200000000000000000000000500000000000000076d796b65796d7976616c75653041414100000000000000000000000300000000000000056b657976616c7565304242420000000000000000", hex.EncodeToString(buf[:100]))
+}