sync.go 12.8 KB
Newer Older
D
Davies Liu 已提交
1 2
// Copyright (C) 2018-present Juicedata Inc.

T
tonic 已提交
3
package sync
D
Davies Liu 已提交
4 5 6

import (
	"bytes"
D
Davies Liu 已提交
7
	"fmt"
D
Davies Liu 已提交
8 9 10
	"io"
	"io/ioutil"
	"os"
11
	"regexp"
12
	"strings"
D
Davies Liu 已提交
13 14
	"sync"
	"sync/atomic"
D
Davies Liu 已提交
15
	"time"
D
Davies Liu 已提交
16
	"unsafe"
D
Davies Liu 已提交
17

T
tonic 已提交
18
	"github.com/juicedata/juicesync/config"
D
Davies Liu 已提交
19
	"github.com/juicedata/juicesync/object"
T
tonic 已提交
20
	"github.com/juicedata/juicesync/utils"
D
Davies Liu 已提交
21
	"github.com/mattn/go-isatty"
D
Davies Liu 已提交
22 23 24
)

// The max number of key per listing request
D
Davies Liu 已提交
25 26 27 28
const (
	maxResults      = 10240
	defaultPartSize = 5 << 20
	maxBlock        = defaultPartSize * 2
Y
Yujun Zhang 已提交
29
	markDelete      = -1
D
Davies Liu 已提交
30
	markCopyPerms   = -2
D
Davies Liu 已提交
31
)
D
Davies Liu 已提交
32

D
Davies Liu 已提交
33
var (
D
Davies Liu 已提交
34 35 36 37 38
	found       uint64
	missing     uint64
	copied      uint64
	copiedBytes uint64
	failed      uint64
39
	deleted     uint64
D
Davies Liu 已提交
40
	concurrent  chan int
D
Davies Liu 已提交
41
)
D
Davies Liu 已提交
42

T
tonic 已提交
43 44
var logger = utils.GetLogger("juicesync")

45 46
// iterate on all the keys that starts at marker from object storage.
func iterate(store object.ObjectStorage, marker, end string) (<-chan *object.Object, error) {
D
Davies Liu 已提交
47
	start := time.Now()
48
	logger.Debugf("Listing objects from %s marker %q", store, marker)
D
Davies Liu 已提交
49
	objs, err := store.List("", marker, maxResults)
D
Davies Liu 已提交
50 51 52 53
	if err != nil {
		logger.Errorf("Can't list %s: %s", store, err.Error())
		return nil, err
	}
D
Davies Liu 已提交
54
	logger.Debugf("Found %d object from %s in %s", len(objs), store, time.Now().Sub(start))
D
Davies Liu 已提交
55
	out := make(chan *object.Object, maxResults)
D
Davies Liu 已提交
56
	go func() {
D
Davies Liu 已提交
57 58
		lastkey := ""
		first := true
D
Davies Liu 已提交
59
	END:
D
Davies Liu 已提交
60 61 62
		for len(objs) > 0 {
			for _, obj := range objs {
				key := obj.Key
D
Davies Liu 已提交
63
				if !first && key <= lastkey {
64 65
					logger.Fatalf("The keys are out of order: marker %q, last %q current %q", marker, lastkey, key)
				}
D
Davies Liu 已提交
66 67 68
				if end != "" && key >= end {
					break END
				}
D
Davies Liu 已提交
69 70
				lastkey = key
				out <- obj
D
Davies Liu 已提交
71
				first = false
D
Davies Liu 已提交
72 73
			}
			marker = lastkey
D
Davies Liu 已提交
74
			start = time.Now()
75
			logger.Debugf("Continue listing objects from %s marker %q", store, marker)
D
Davies Liu 已提交
76
			objs, err = store.List("", marker, maxResults)
D
Davies Liu 已提交
77 78 79 80 81 82
			for err != nil {
				logger.Warnf("Fail to list: %s, retry again", err.Error())
				// slow down
				time.Sleep(time.Millisecond * 100)
				objs, err = store.List("", marker, maxResults)
			}
D
Davies Liu 已提交
83
			logger.Debugf("Found %d object from %s in %s", len(objs), store, time.Now().Sub(start))
D
Davies Liu 已提交
84 85 86 87 88 89 90 91 92 93 94 95
			if err != nil {
				// Telling that the listing has failed
				out <- nil
				logger.Errorf("Fail to list after %s: %s", marker, err.Error())
				break
			}
		}
		close(out)
	}()
	return out, nil
}

T
tonic 已提交
96
func copyObject(src, dst object.ObjectStorage, obj *object.Object) error {
D
Davies Liu 已提交
97 98 99 100
	concurrent <- 1
	defer func() {
		<-concurrent
	}()
D
Davies Liu 已提交
101
	key := obj.Key
102 103 104 105 106 107 108 109 110 111 112
	if strings.HasPrefix(src.String(), "file://") || strings.HasPrefix(dst.String(), "file://") {
		in, e := src.Get(key, 0, -1)
		if e != nil {
			if src.Exists(key) != nil {
				return nil
			}
			return e
		}
		defer in.Close()
		return dst.Put(key, in)
	}
D
Davies Liu 已提交
113 114 115 116 117
	firstBlock := -1
	if obj.Size > maxBlock {
		firstBlock = maxBlock
	}
	in, e := src.Get(key, 0, int64(firstBlock))
D
Davies Liu 已提交
118 119 120 121 122 123 124 125 126 127 128
	if e != nil {
		if src.Exists(key) != nil {
			return nil
		}
		return e
	}
	data, err := ioutil.ReadAll(in)
	in.Close()
	if err != nil {
		return err
	}
D
Davies Liu 已提交
129
	if firstBlock == -1 {
D
Davies Liu 已提交
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
		return dst.Put(key, bytes.NewReader(data))
	}

	// download the object into disk first
	f, err := ioutil.TempFile("", "rep")
	if err != nil {
		return err
	}
	os.Remove(f.Name()) // will be deleted after Close()
	defer f.Close()
	if _, err := f.Write(data); err != nil {
		return err
	}
	if in, e = src.Get(key, int64(len(data)), -1); e != nil {
		return e
	}
	_, e = io.Copy(f, in)
	in.Close()
	if e != nil {
		return e
	}
	// upload
	if _, e = f.Seek(0, 0); e != nil {
		return e
	}
	return dst.Put(key, f)
}

D
Davies Liu 已提交
158 159 160 161 162 163 164 165 166 167 168 169 170 171
func try(n int, f func() error) (err error) {
	for i := 0; i < n; i++ {
		err = f()
		if err == nil {
			return
		}
		time.Sleep(time.Second * time.Duration(i*i))
	}
	return
}

func copyInParallel(src, dst object.ObjectStorage, obj *object.Object) error {
	if obj.Size < maxBlock {
		return try(3, func() error {
T
tonic 已提交
172
			return copyObject(src, dst, obj)
D
Davies Liu 已提交
173 174 175 176 177
		})
	}
	upload, err := dst.CreateMultipartUpload(obj.Key)
	if err != nil {
		return try(3, func() error {
T
tonic 已提交
178
			return copyObject(src, dst, obj)
D
Davies Liu 已提交
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
		})
	}
	partSize := int64(upload.MinPartSize)
	if partSize == 0 {
		partSize = defaultPartSize
	}
	if obj.Size > partSize*int64(upload.MaxCount) {
		partSize = obj.Size / int64(upload.MaxCount)
		partSize = ((partSize-1)>>20 + 1) << 20 // align to MB
	}
	n := int((obj.Size-1)/partSize) + 1
	key := obj.Key
	logger.Debugf("Copying object %s as %d parts (size: %d): %s", key, n, partSize, upload.UploadID)
	parts := make([]*object.Part, n)
	errs := make(chan error, n)
	for i := 0; i < n; i++ {
		go func(num int) {
			sz := partSize
			if num == n-1 {
				sz = obj.Size - int64(num)*partSize
			}
			var err error
			concurrent <- 1
			defer func() {
				<-concurrent
			}()
			data := make([]byte, sz)
			err = try(3, func() error {
				r, err := src.Get(key, int64(num)*partSize, int64(sz))
				if err != nil {
					return nil
				}
				_, err = io.ReadFull(r, data)
				return err
			})
			if err == nil {
				err = try(3, func() error {
					// PartNumber starts from 1
					parts[num], err = dst.UploadPart(key, upload.UploadID, num+1, data)
					return err
				})
			}
			if err != nil {
				errs <- fmt.Errorf("part %d: %s", num, err.Error())
				logger.Warningf("Failed to copy %s part %d: %s", key, num, err.Error())
			} else {
				errs <- nil
				logger.Debugf("Copied %s part %d", key, num)
			}
		}(i)
	}
	for i := 0; i < n; i++ {
		if err = <-errs; err != nil {
			break
		}
	}
	if err == nil {
		err = try(3, func() error {
			return dst.CompleteUpload(key, upload.UploadID, parts)
		})
	}
	if err != nil {
		dst.AbortUpload(key, upload.UploadID)
		return fmt.Errorf("multipart: %s", err.Error())
	}
	return nil
}

T
tonic 已提交
247
func doSync(src, dst object.ObjectStorage, srckeys, dstkeys <-chan *object.Object, config *config.Config) {
R
root 已提交
248
	todo := make(chan *object.Object, 10240)
D
Davies Liu 已提交
249
	wg := sync.WaitGroup{}
T
tonic 已提交
250 251
	concurrent = make(chan int, config.Threads)
	for i := 0; i < config.Threads; i++ {
D
Davies Liu 已提交
252 253 254 255
		wg.Add(1)
		go func() {
			defer wg.Done()
			for {
D
Davies Liu 已提交
256
				obj, ok := <-todo
D
Davies Liu 已提交
257 258 259
				if !ok {
					break
				}
D
Davies Liu 已提交
260
				start := time.Now()
D
Davies Liu 已提交
261
				var err error
Y
Yujun Zhang 已提交
262
				if obj.Size == markDelete {
T
tonic 已提交
263 264
					if config.DeleteSrc {
						if config.Dry {
Y
Yujun Zhang 已提交
265 266 267 268 269 270 271 272 273 274 275
							logger.Debugf("Will delete %s from %s", obj.Key, src)
							continue
						}
						if err = try(3, func() error {
							return src.Delete(obj.Key)
						}); err == nil {
							logger.Debugf("Deleted %s from %s", obj.Key, src)
							atomic.AddUint64(&deleted, 1)
						} else {
							logger.Errorf("Failed to delete %s from %s: %s", obj.Key, src, err.Error())
							atomic.AddUint64(&failed, 1)
D
Davies Liu 已提交
276
						}
277
					}
T
tonic 已提交
278 279
					if config.DeleteDst {
						if config.Dry {
Y
Yujun Zhang 已提交
280 281 282 283 284 285 286 287 288 289 290 291
							logger.Debugf("Will delete %s from %s", obj.Key, dst)
							continue
						}
						if err = try(3, func() error {
							return dst.Delete(obj.Key)
						}); err == nil {
							logger.Debugf("Deleted %s from %s", obj.Key, dst)
							atomic.AddUint64(&deleted, 1)
						} else {
							logger.Errorf("Failed to delete %s from %s: %s", obj.Key, dst, err.Error())
							atomic.AddUint64(&failed, 1)
						}
292 293 294
					}
					continue
				}
D
Davies Liu 已提交
295

T
tonic 已提交
296
				if config.Dry {
Y
Yujun Zhang 已提交
297 298
					logger.Debugf("Will copy %s (%d bytes)", obj.Key, obj.Size)
					continue
D
Davies Liu 已提交
299
				}
D
Davies Liu 已提交
300 301 302 303 304 305 306 307 308 309 310 311
				if config.Perms && obj.Size == markCopyPerms {
					fi := (*object.File)(unsafe.Pointer(obj))
					if err := dst.(object.FileSystem).Chmod(obj.Key, fi.Mode); err != nil {
						logger.Warnf("Chmod %s to %d: %s", obj.Key, err)
					}
					if err := dst.(object.FileSystem).Chown(obj.Key, fi.Owner, fi.Group); err != nil {
						logger.Warnf("Chown %s to (%s,%s): %s", obj.Key, fi.Owner, fi.Group, err)
					}
					atomic.AddUint64(&copied, 1)
					logger.Debugf("Copied permissions for %s in %s", obj.Key, time.Now().Sub(start))
					continue
				}
Y
Yujun Zhang 已提交
312
				err = copyInParallel(src, dst, obj)
D
Davies Liu 已提交
313
				if err != nil {
D
Davies Liu 已提交
314
					atomic.AddUint64(&failed, 1)
D
Davies Liu 已提交
315
					logger.Errorf("Failed to copy %s: %s", obj.Key, err.Error())
D
Davies Liu 已提交
316
				} else {
317 318 319 320 321 322
					if mc, ok := dst.(object.MtimeChanger); ok {
						err := mc.Chtimes(obj.Key, obj.Mtime)
						if err != nil {
							logger.Warnf("Update mtime of %s: %s", obj.Key, err)
						}
					}
D
Davies Liu 已提交
323 324 325 326 327 328 329 330 331
					if config.Perms {
						fi := (*object.File)(unsafe.Pointer(obj))
						if err := dst.(object.FileSystem).Chmod(obj.Key, fi.Mode); err != nil {
							logger.Warnf("Chmod %s to %o: %s", obj.Key, fi.Mode, err)
						}
						if err := dst.(object.FileSystem).Chown(obj.Key, fi.Owner, fi.Group); err != nil {
							logger.Warnf("Chown %s to (%s,%s): %s", obj.Key, fi.Owner, fi.Group, err)
						}
					}
D
Davies Liu 已提交
332
					atomic.AddUint64(&copied, 1)
D
Davies Liu 已提交
333
					atomic.AddUint64(&copiedBytes, uint64(obj.Size))
D
Davies Liu 已提交
334
					logger.Debugf("Copied %s (%d bytes) in %s", obj.Key, obj.Size, time.Now().Sub(start))
D
Davies Liu 已提交
335 336 337 338 339
				}
			}
		}()
	}

D
Davies Liu 已提交
340
	var dstobj *object.Object
D
Davies Liu 已提交
341
	hasMore := true
D
Davies Liu 已提交
342
OUT:
D
Davies Liu 已提交
343 344
	for obj := range srckeys {
		if obj == nil {
D
Davies Liu 已提交
345
			logger.Errorf("Listing failed, stop syncing, waiting for pending ones")
346
			hasMore = false
D
Davies Liu 已提交
347 348
			break
		}
D
Davies Liu 已提交
349
		atomic.AddUint64(&found, 1)
D
Davies Liu 已提交
350
		for hasMore && (dstobj == nil || obj.Key > dstobj.Key) {
D
Davies Liu 已提交
351
			var ok bool
T
tonic 已提交
352
			if config.DeleteDst && dstobj != nil && dstobj.Key < obj.Key {
Y
Yujun Zhang 已提交
353
				dstobj.Size = markDelete
354 355
				todo <- dstobj
			}
D
Davies Liu 已提交
356
			dstobj, ok = <-dstkeys
D
Davies Liu 已提交
357
			if !ok {
D
Davies Liu 已提交
358 359
				hasMore = false
			} else if dstobj == nil {
D
Davies Liu 已提交
360
				// Listing failed, stop
D
Davies Liu 已提交
361
				logger.Errorf("Listing failed, stop syncing, waiting for pending ones")
362
				hasMore = false
D
Davies Liu 已提交
363 364 365
				break OUT
			}
		}
D
Davies Liu 已提交
366
		// FIXME: there is a race when source is modified during coping
367 368
		if !hasMore || obj.Key < dstobj.Key ||
			obj.Key == dstobj.Key && (obj.Size != dstobj.Size || config.Update && obj.Mtime.After(dstobj.Mtime)) {
D
Davies Liu 已提交
369
			todo <- obj
D
Davies Liu 已提交
370
			atomic.AddUint64(&missing, 1)
T
tonic 已提交
371
		} else if config.DeleteSrc && dstobj != nil && obj.Key == dstobj.Key && obj.Size == dstobj.Size {
Y
Yujun Zhang 已提交
372
			obj.Size = markDelete
373
			todo <- obj
D
Davies Liu 已提交
374 375 376 377 378 379 380
		} else if config.Perms {
			f1 := (*object.File)(unsafe.Pointer(obj))
			f2 := (*object.File)(unsafe.Pointer(dstobj))
			if f2.Mode != f1.Mode || f2.Owner != f1.Owner || f2.Group != f1.Group {
				obj.Size = markCopyPerms
				todo <- obj
			}
D
Davies Liu 已提交
381
		}
382 383 384 385
		if dstobj != nil && dstobj.Key == obj.Key {
			dstobj = nil
		}
	}
T
tonic 已提交
386
	if config.DeleteDst && hasMore {
387
		if dstobj != nil {
Y
Yujun Zhang 已提交
388
			dstobj.Size = markDelete
389 390 391 392
			todo <- dstobj
		}
		for obj := range dstkeys {
			if obj != nil {
Y
Yujun Zhang 已提交
393
				obj.Size = markDelete
394 395 396
				todo <- obj
			}
		}
D
Davies Liu 已提交
397 398 399 400 401
	}
	close(todo)
	wg.Wait()
}

D
Davies Liu 已提交
402
func showProgress() {
D
Davies Liu 已提交
403
	var lastCopied, lastBytes uint64
D
Davies Liu 已提交
404
	var lastTime = time.Now()
D
Davies Liu 已提交
405
	for {
D
Davies Liu 已提交
406 407 408 409
		if found == 0 {
			time.Sleep(time.Millisecond * 10)
			continue
		}
D
Davies Liu 已提交
410
		same := atomic.LoadUint64(&found) - atomic.LoadUint64(&missing)
411
		var width uint64 = 45
D
Davies Liu 已提交
412 413 414 415 416 417 418 419 420 421 422 423 424 425 426
		a := width * same / found
		b := width * copied / found
		var bar [80]byte
		var i uint64
		for i = 0; i < width; i++ {
			if i < a {
				bar[i] = '='
			} else if i < a+b {
				bar[i] = '+'
			} else {
				bar[i] = ' '
			}
		}
		now := time.Now()
		fps := float64(copied-lastCopied) / now.Sub(lastTime).Seconds()
D
Davies Liu 已提交
427
		bw := float64(copiedBytes-lastBytes) / now.Sub(lastTime).Seconds() / 1024 / 1024
D
Davies Liu 已提交
428
		lastCopied = copied
R
root 已提交
429
		lastBytes = copiedBytes
D
Davies Liu 已提交
430
		lastTime = now
431
		fmt.Printf("[%s] % 8d % 2d%% % 4.0f/s % 4.1f MB/s \r", string(bar[:]), found, (found-missing+copied)*100/found, fps, bw)
D
Davies Liu 已提交
432
		time.Sleep(time.Millisecond * 300)
D
Davies Liu 已提交
433 434 435
	}
}

436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465 466 467 468 469 470 471 472 473 474 475 476 477 478 479 480
func compileExp(patterns []string) []*regexp.Regexp {
	var rs []*regexp.Regexp
	for _, p := range patterns {
		r, err := regexp.CompilePOSIX(p)
		if err != nil {
			logger.Fatalf("invalid regular expression `%s`: %s", p, err)
		}
		rs = append(rs, r)
	}
	return rs
}

func findAny(s string, ps []*regexp.Regexp) bool {
	for _, p := range ps {
		if p.FindString(s) != "" {
			return true
		}
	}
	return false
}

func filter(keys <-chan *object.Object, include, exclude []string) <-chan *object.Object {
	inc := compileExp(include)
	exc := compileExp(exclude)
	r := make(chan *object.Object)
	go func() {
		for o := range keys {
			if o == nil {
				break
			}
			if findAny(o.Key, exc) {
				logger.Debugf("exclude %s", o.Key)
				continue
			}
			if len(inc) > 0 && !findAny(o.Key, inc) {
				logger.Debugf("%s is not included", o.Key)
				continue
			}
			r <- o
		}
		close(r)
	}()
	return r
}

D
Davies Liu 已提交
481
// Sync syncs all the keys between to object storage
T
tonic 已提交
482 483
func Sync(src, dst object.ObjectStorage, config *config.Config) error {
	start, end := config.Start, config.End
Y
Yujun Zhang 已提交
484 485 486 487 488 489 490
	logger.Infof("Syncing from %s to %s", src, dst)
	if start != "" {
		logger.Infof("first key: %q", start)
	}
	if end != "" {
		logger.Infof("last key: %q", end)
	}
491
	logger.Debugf("maxResults: %d, defaultPartSize: %d, maxBlock: %d", maxResults, defaultPartSize, maxBlock)
T
tonic 已提交
492

493
	srcCh, err := iterate(src, start, end)
D
Davies Liu 已提交
494 495 496
	if err != nil {
		return err
	}
T
tonic 已提交
497

498
	dstCh, err := iterate(dst, start, end)
D
Davies Liu 已提交
499 500 501
	if err != nil {
		return err
	}
502 503 504 505
	if config.Exclude != nil {
		srcCh = filter(srcCh, config.Include, config.Exclude)
		dstCh = filter(dstCh, config.Include, config.Exclude)
	}
D
Davies Liu 已提交
506

D
Davies Liu 已提交
507
	tty := isatty.IsTerminal(os.Stdout.Fd())
T
tonic 已提交
508
	if tty && !config.Verbose && !config.Quiet {
D
Davies Liu 已提交
509 510
		go showProgress()
	}
T
tonic 已提交
511
	doSync(src, dst, srcCh, dstCh, config)
512
	logger.Infof("Found: %d, copied: %d, deleted: %d, failed: %d", found, copied, deleted, failed)
513 514 515
	if failed > 0 {
		return fmt.Errorf("Failed to copy %d objects", failed)
	}
D
Davies Liu 已提交
516 517
	return nil
}