forked from dgraph-io/badger
-
Notifications
You must be signed in to change notification settings - Fork 0
/
backup.go
288 lines (261 loc) · 8.02 KB
/
backup.go
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
/*
* Copyright 2017 Dgraph Labs, Inc. and Contributors
*
* Licensed under the Apache License, Version 2.0 (the "License");
* you may not use this file except in compliance with the License.
* You may obtain a copy of the License at
*
* http://www.apache.org/licenses/LICENSE-2.0
*
* Unless required by applicable law or agreed to in writing, software
* distributed under the License is distributed on an "AS IS" BASIS,
* WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
* See the License for the specific language governing permissions and
* limitations under the License.
*/
package badger
import (
"bufio"
"bytes"
"context"
"encoding/binary"
"io"
"github.com/dgraph-io/badger/v3/pb"
"github.com/dgraph-io/badger/v3/y"
"github.com/dgraph-io/ristretto/z"
"github.com/golang/protobuf/proto"
)
// flushThreshold determines when a buffer will be flushed. When performing a
// backup/restore, the entries will be batched up until the total size of batch
// is more than flushThreshold or entry size (without the value size) is more
// than the maxBatchSize.
const flushThreshold = 100 << 20
// Backup dumps a protobuf-encoded list of all entries in the database into the
// given writer, that are newer than or equal to the specified version. It
// returns a timestamp (version) indicating the version of last entry that is
// dumped, which after incrementing by 1 can be passed into later invocation to
// generate incremental backup of entries that have been added/modified since
// the last invocation of DB.Backup().
// DB.Backup is a wrapper function over Stream.Backup to generate full and
// incremental backups of the DB. For more control over how many goroutines are
// used to generate the backup, or if you wish to backup only a certain range
// of keys, use Stream.Backup directly.
func (db *DB) Backup(w io.Writer, since uint64) (uint64, error) {
stream := db.NewStream()
stream.LogPrefix = "DB.Backup"
return stream.Backup(w, since)
}
// Backup dumps a protobuf-encoded list of all entries in the database into the
// given writer, that are newer than or equal to the specified version. It returns a
// timestamp(version) indicating the version of last entry that was dumped, which
// after incrementing by 1 can be passed into a later invocation to generate an
// incremental dump of entries that have been added/modified since the last
// invocation of Stream.Backup().
//
// This can be used to backup the data in a database at a given point in time.
func (stream *Stream) Backup(w io.Writer, since uint64) (uint64, error) {
stream.KeyToList = func(key []byte, itr *Iterator) (*pb.KVList, error) {
list := &pb.KVList{}
a := itr.Alloc
for ; itr.Valid(); itr.Next() {
item := itr.Item()
if !bytes.Equal(item.Key(), key) {
return list, nil
}
if item.Version() < since {
// Ignore versions less than given timestamp, or skip older
// versions of the given key.
return list, nil
}
var valCopy []byte
if !item.IsDeletedOrExpired() {
// No need to copy value, if item is deleted or expired.
var err error
err = item.Value(func(val []byte) error {
valCopy = a.Copy(val)
return nil
})
if err != nil {
stream.db.opt.Errorf("Key [%x, %d]. Error while fetching value [%v]\n",
item.Key(), item.Version(), err)
return nil, err
}
}
// clear txn bits
meta := item.meta &^ (bitTxn | bitFinTxn)
kv := y.NewKV(a)
*kv = pb.KV{
Key: a.Copy(item.Key()),
Value: valCopy,
UserMeta: a.Copy([]byte{item.UserMeta()}),
Version: item.Version(),
ExpiresAt: item.ExpiresAt(),
Meta: a.Copy([]byte{meta}),
}
list.Kv = append(list.Kv, kv)
switch {
case item.DiscardEarlierVersions():
// If we need to discard earlier versions of this item, add a delete
// marker just below the current version.
list.Kv = append(list.Kv, &pb.KV{
Key: item.KeyCopy(nil),
Version: item.Version() - 1,
Meta: []byte{bitDelete},
})
return list, nil
case item.IsDeletedOrExpired():
return list, nil
}
}
return list, nil
}
var maxVersion uint64
stream.Send = func(buf *z.Buffer) error {
list, err := BufferToKVList(buf)
if err != nil {
return err
}
out := list.Kv[:0]
for _, kv := range list.Kv {
if maxVersion < kv.Version {
maxVersion = kv.Version
}
if !kv.StreamDone {
// Don't pick stream done changes.
out = append(out, kv)
}
}
list.Kv = out
return writeTo(list, w)
}
if err := stream.Orchestrate(context.Background()); err != nil {
return 0, err
}
return maxVersion, nil
}
func writeTo(list *pb.KVList, w io.Writer) error {
if err := binary.Write(w, binary.LittleEndian, uint64(proto.Size(list))); err != nil {
return err
}
buf, err := proto.Marshal(list)
if err != nil {
return err
}
_, err = w.Write(buf)
return err
}
// KVLoader is used to write KVList objects in to badger. It can be used to restore a backup.
type KVLoader struct {
db *DB
throttle *y.Throttle
entries []*Entry
entriesSize int64
totalSize int64
}
// NewKVLoader returns a new instance of KVLoader.
func (db *DB) NewKVLoader(maxPendingWrites int) *KVLoader {
return &KVLoader{
db: db,
throttle: y.NewThrottle(maxPendingWrites),
entries: make([]*Entry, 0, db.opt.maxBatchCount),
}
}
// Set writes the key-value pair to the database.
func (l *KVLoader) Set(kv *pb.KV) error {
var userMeta, meta byte
if len(kv.UserMeta) > 0 {
userMeta = kv.UserMeta[0]
}
if len(kv.Meta) > 0 {
meta = kv.Meta[0]
}
e := &Entry{
Key: y.KeyWithTs(kv.Key, kv.Version),
Value: kv.Value,
UserMeta: userMeta,
ExpiresAt: kv.ExpiresAt,
meta: meta,
}
estimatedSize := int64(e.estimateSize(l.db.opt.ValueThreshold))
// Flush entries if inserting the next entry would overflow the transactional limits.
if int64(len(l.entries))+1 >= l.db.opt.maxBatchCount ||
l.entriesSize+estimatedSize >= l.db.opt.maxBatchSize ||
l.totalSize >= flushThreshold {
if err := l.send(); err != nil {
return err
}
}
l.entries = append(l.entries, e)
l.entriesSize += estimatedSize
l.totalSize += estimatedSize + int64(len(e.Value))
return nil
}
func (l *KVLoader) send() error {
if err := l.throttle.Do(); err != nil {
return err
}
if err := l.db.batchSetAsync(l.entries, func(err error) {
l.throttle.Done(err)
}); err != nil {
return err
}
l.entries = make([]*Entry, 0, l.db.opt.maxBatchCount)
l.entriesSize = 0
l.totalSize = 0
return nil
}
// Finish is meant to be called after all the key-value pairs have been loaded.
func (l *KVLoader) Finish() error {
if len(l.entries) > 0 {
if err := l.send(); err != nil {
return err
}
}
return l.throttle.Finish()
}
// Load reads a protobuf-encoded list of all entries from a reader and writes
// them to the database. This can be used to restore the database from a backup
// made by calling DB.Backup(). If more complex logic is needed to restore a badger
// backup, the KVLoader interface should be used instead.
//
// DB.Load() should be called on a database that is not running any other
// concurrent transactions while it is running.
func (db *DB) Load(r io.Reader, maxPendingWrites int) error {
br := bufio.NewReaderSize(r, 16<<10)
unmarshalBuf := make([]byte, 1<<10)
ldr := db.NewKVLoader(maxPendingWrites)
for {
var sz uint64
err := binary.Read(br, binary.LittleEndian, &sz)
if err == io.EOF {
break
} else if err != nil {
return err
}
if cap(unmarshalBuf) < int(sz) {
unmarshalBuf = make([]byte, sz)
}
if _, err = io.ReadFull(br, unmarshalBuf[:sz]); err != nil {
return err
}
list := &pb.KVList{}
if err := proto.Unmarshal(unmarshalBuf[:sz], list); err != nil {
return err
}
for _, kv := range list.Kv {
if err := ldr.Set(kv); err != nil {
return err
}
// Update nextTxnTs, memtable stores this
// timestamp in badger head when flushed.
if kv.Version >= db.orc.nextTxnTs {
db.orc.nextTxnTs = kv.Version + 1
}
}
}
if err := ldr.Finish(); err != nil {
return err
}
db.orc.txnMark.Done(db.orc.nextTxnTs - 1)
return nil
}