forked from gochain/gochain
/
table.go
463 lines (396 loc) · 11 KB
/
table.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
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
package ethdb
import (
"bytes"
"context"
"fmt"
"os"
"path/filepath"
"sort"
"sync"
"time"
"github.com/gochain-io/gochain/common"
"github.com/gochain-io/gochain/log"
)
// Table represents key/value storage for a particular data type.
// Contains zero or more segments that are separated by partitioner.
type Table struct {
mu sync.RWMutex
active string // active segment name
ldbSegments map[string]*LDBSegment // writable segments
segments *SegmentSet // all segments
Name string
Path string
Partitioner Partitioner
MinMutableSegmentCount int
MinCompactionAge time.Duration
// Maximum number of segments that can be opened at once.
MaxOpenSegmentCount int
SegmentOpener SegmentOpener
SegmentCompactor SegmentCompactor
}
// NewTable returns a new instance of Table.
func NewTable(name, path string, partitioner Partitioner) *Table {
return &Table{
Name: name,
Path: path,
Partitioner: partitioner,
MinMutableSegmentCount: DefaultMinMutableSegmentCount,
MinCompactionAge: DefaultMinCompactionAge,
MaxOpenSegmentCount: DefaultMaxOpenSegmentCount,
SegmentOpener: NewFileSegmentOpener(),
SegmentCompactor: NewFileSegmentCompactor(),
}
}
// Open initializes the table and all existing segments.
func (t *Table) Open() error {
t.ldbSegments = make(map[string]*LDBSegment)
t.segments = NewSegmentSet(t.MaxOpenSegmentCount)
if err := os.MkdirAll(t.Path, 0777); err != nil {
return err
}
names, err := t.SegmentOpener.ListSegmentNames(t.Path, t.Name)
if err != nil {
log.Error("Cannot list segment names", "path", t.Path, "name", t.Name, "err", err)
return err
}
for _, name := range names {
path := filepath.Join(t.Path, name)
// Determine the segment file type.
typ, err := SegmentFileType(path)
if err == ErrInvalidSegmentType {
log.Warn("Invalid segment type, skipping", "path", path, "name", name, "err", err)
continue
} else if err != nil && !os.IsNotExist(err) {
return err
}
// Open appropriate segment type.
switch typ {
case SegmentLDB1:
ldbSegment := NewLDBSegment(name, path)
if err := ldbSegment.Open(); err != nil {
log.Error("Cannot open ldb segment", "path", path, "name", name, "err", err)
t.Close()
return err
}
t.ldbSegments[name] = ldbSegment
default:
segment, err := t.SegmentOpener.OpenSegment(t.Name, name, path)
if err == ErrSegmentTypeUnknown {
log.Info("unknown segment type, skipping", "filename", name)
continue
} else if err != nil {
log.Error("Cannot open ethdb segment", "path", path, "table", t.Name, "name", name, "err", err)
return err
}
t.segments.Add(segment)
}
// Set as active if it has the highest lexicographical name.
if name > t.active {
t.active = name
}
}
return nil
}
// Close closes all segments within the table.
func (t *Table) Close() error {
for _, segment := range t.ldbSegments {
if err := segment.Close(); err != nil {
return err
}
}
if t.segments != nil {
for _, segment := range t.segments.Slice() {
if err := segment.Close(); err != nil {
return err
}
t.segments.Remove(segment.Name())
}
}
return nil
}
// ActiveSegmentName the name of the current active segment.
func (t *Table) ActiveSegmentName() string {
t.mu.RLock()
defer t.mu.RUnlock()
return t.active
}
// ActiveSegment returns the active segment.
func (t *Table) ActiveSegment() MutableSegment {
t.mu.RLock()
defer t.mu.RUnlock()
return t.ldbSegments[t.active]
}
// SegmentPath returns the path of the named segment.
func (t *Table) SegmentPath(name string) string {
return filepath.Join(t.Path, name)
}
// AcquireSegment returns a segment by name. Returns nil if segment does not exist.
// Must call ReleaseSegment when finished with the segment.
func (t *Table) AcquireSegment(name string) (Segment, error) {
t.mu.RLock()
defer t.mu.RUnlock()
if s := t.ldbSegments[name]; s != nil {
return s, nil
}
return t.segments.Acquire(name)
}
// ReleaseSegment releases a given segment.
func (t *Table) ReleaseSegment(s Segment) {
switch s.(type) {
case *LDBSegment:
return
default:
t.segments.Release()
}
}
// SegmentNames a sorted list of all segments names.
func (t *Table) SegmentNames() []string {
t.mu.RLock()
defer t.mu.RUnlock()
a := make([]string, 0, len(t.ldbSegments)+t.segments.Len())
for _, s := range t.ldbSegments {
a = append(a, s.Name())
}
for _, s := range t.segments.Slice() {
a = append(a, s.Name())
}
sort.Strings(a)
return a
}
// SegmentsSlice returns a sorted slice of all segments.
func (t *Table) SegmentSlice() []Segment {
t.mu.RLock()
defer t.mu.RUnlock()
return t.segmentSlice()
}
func (t *Table) segmentSlice() []Segment {
a := make([]Segment, 0, len(t.ldbSegments)+t.segments.Len())
for _, s := range t.ldbSegments {
a = append(a, s)
}
for _, s := range t.segments.Slice() {
a = append(a, s)
}
SortSegments(a)
return a
}
func (t *Table) ldbSegmentSlice() []*LDBSegment {
a := make([]*LDBSegment, 0, len(t.ldbSegments))
for _, s := range t.ldbSegments {
a = append(a, s)
}
sort.Slice(a, func(i, j int) bool { return a[i].Name() < a[j].Name() })
return a
}
// CreateSegmentIfNotExists returns a mutable segment by name.
// Creates a new segment if it does not exist.
func (t *Table) CreateSegmentIfNotExists(name string) (*LDBSegment, error) {
t.mu.RLock()
if s := t.ldbSegments[name]; s != nil {
t.mu.RUnlock()
return s, nil
}
t.mu.RUnlock()
t.mu.Lock()
defer t.mu.Unlock()
// Recheck under write lock.
if s := t.ldbSegments[name]; s != nil {
return s, nil
}
// Uncompact segment if it has already become compacted.
if t.segments.Contains(name) {
return t.uncompact(context.TODO(), name)
}
// Ensure segment name can become active.
if name < t.active {
log.Error("cannot new non-active create segment", "name", name, "active", t.active)
return nil, ErrImmutableSegment
}
// Create new mutable segment.
ldbSegment := NewLDBSegment(name, t.SegmentPath(name))
if err := ldbSegment.Open(); err != nil {
return nil, err
}
t.ldbSegments[name] = ldbSegment
// Set as active segment.
t.active = name
// Compact under lock.
// TODO(benbjohnson): Run compaction in background if too slow.
if err := t.compact(context.TODO()); err != nil {
return nil, err
}
return ldbSegment, nil
}
// Has returns true if key exists in the table.
func (t *Table) Has(key []byte) (bool, error) {
name := t.Partitioner.Partition(key)
s, err := t.AcquireSegment(name)
if err != nil {
return false, err
} else if s == nil {
return false, common.ErrNotFound
}
defer t.ReleaseSegment(s)
return s.Has(key)
}
// Get returns the value associated with key.
func (t *Table) Get(key []byte) ([]byte, error) {
name := t.Partitioner.Partition(key)
s, err := t.AcquireSegment(name)
if err != nil {
return nil, err
} else if s == nil {
return nil, nil
}
defer t.ReleaseSegment(s)
return s.Get(key)
}
// Put associates a value with key.
func (t *Table) Put(key, value []byte) error {
// Ignore if value is the same.
if v, err := t.Get(key); err != nil && err != common.ErrNotFound {
return err
} else if bytes.Equal(v, value) {
return nil
}
s, err := t.CreateSegmentIfNotExists(t.Partitioner.Partition(key))
if err != nil {
return err
}
return s.Put(key, value)
}
// Delete removes key from the database.
func (t *Table) Delete(key []byte) error {
s, err := t.AcquireSegment(t.Partitioner.Partition(key))
if err != nil {
return err
} else if s == nil {
return nil
}
defer t.ReleaseSegment(s)
switch s := s.(type) {
case MutableSegment:
return s.Delete(key)
default:
return ErrImmutableSegment
}
}
func (t *Table) NewBatch() common.Batch {
return &tableBatch{table: t, batches: make(map[string]*ldbSegmentBatch)}
}
// Compact converts LDB segments into immutable file segments.
func (t *Table) Compact(ctx context.Context) error {
t.mu.Lock()
defer t.mu.Unlock()
return t.compact(ctx)
}
func (t *Table) compact(ctx context.Context) error {
// Retrieve LDB segments. Exit if too few mutable segments.
ldbSegmentSlice := t.ldbSegmentSlice()
if len(ldbSegmentSlice) < t.MinMutableSegmentCount {
return nil
}
for _, ldbSegment := range ldbSegmentSlice[:len(ldbSegmentSlice)-t.MinMutableSegmentCount] {
startTime := time.Now()
if fi, err := os.Stat(ldbSegment.Path()); err != nil {
return err
} else if time.Since(fi.ModTime()) < t.MinCompactionAge {
log.Debug("LDB segment too young, skipping compaction", "table", t.Name, "name", ldbSegment.Name(), "elapsed", time.Since(startTime))
continue
}
newSegment, err := t.SegmentCompactor.CompactSegment(ctx, t.Name, ldbSegment)
if err != nil {
return err
}
t.segments.Add(newSegment)
delete(t.ldbSegments, ldbSegment.Name())
log.Info("Compacted segment", "table", t.Name, "name", ldbSegment.Name(), "elapsed", time.Since(startTime))
}
return nil
}
// uncompact converts an immutable segment to a mutable segment.
func (t *Table) uncompact(ctx context.Context, name string) (*LDBSegment, error) {
startTime := time.Now()
s, err := t.segments.Acquire(name)
if err != nil {
return nil, err
}
defer t.segments.Release()
ldbSegment, err := t.SegmentCompactor.UncompactSegment(ctx, t.Name, s)
if err != nil {
return nil, err
}
t.ldbSegments[name] = ldbSegment
t.segments.Remove(name)
log.Info("Uncompacted segment", "table", t.Name, "name", name, "elapsed", time.Since(startTime))
return ldbSegment, nil
}
type tableBatch struct {
table *Table
batches map[string]*ldbSegmentBatch
size int
}
func (b *tableBatch) Put(key, value []byte) error {
// Ignore if value is the same.
if v, err := b.table.Get(key); err != nil && err != common.ErrNotFound {
return err
} else if bytes.Equal(v, value) {
return nil
}
name := b.table.Partitioner.Partition(key)
ldbSegment, err := b.table.CreateSegmentIfNotExists(name)
if err != nil {
log.Error("tableBatch.Put: error", "table", b.table.Name, "segment", name, "key", fmt.Sprintf("%x", key))
return err
}
sb := b.batches[name]
if sb == nil {
sb = ldbSegment.newBatch()
b.batches[name] = sb
}
if err := sb.Put(key, value); err != nil {
return err
}
b.size += len(value)
return nil
}
func (b *tableBatch) Delete(key []byte) error {
// Ignore if value doesn't exist.
if _, err := b.table.Get(key); err == common.ErrNotFound {
return nil
} else if err != nil {
return err
}
name := b.table.Partitioner.Partition(key)
ldbSegment, err := b.table.CreateSegmentIfNotExists(name)
if err != nil {
log.Error("tableBatch.Delete: error", "table", b.table.Name, "segment", name, "key", fmt.Sprintf("%x", key))
return err
}
sb := b.batches[name]
if sb == nil {
sb = ldbSegment.newBatch()
b.batches[name] = sb
}
if err := sb.Delete(key); err != nil {
return err
}
return nil
}
func (b *tableBatch) Write() error {
for _, sb := range b.batches {
if err := sb.Write(); err != nil {
return err
}
}
return nil
}
func (b *tableBatch) ValueSize() int {
return b.size
}
func (b *tableBatch) Reset() {
for _, sb := range b.batches {
sb.Reset()
}
b.size = 0
}