Repository navigation
Expand file tree
/
Copy pathfs-v1.go
More file actions
1209 lines (1044 loc) · 37 KB
/
Copy pathfs-v1.go
File metadata and controls
1209 lines (1044 loc) · 37 KB
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
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
744
745
746
747
748
749
750
751
752
753
754
755
756
757
758
759
760
761
762
763
764
765
766
767
768
769
770
771
772
773
774
775
776
777
778
779
780
781
782
783
784
785
786
787
788
789
790
791
792
793
794
795
796
797
798
799
800
801
802
803
804
805
806
807
808
809
810
811
812
813
814
815
816
817
818
819
820
821
822
823
824
825
826
827
828
829
830
831
832
833
834
835
836
837
838
839
840
841
842
843
844
845
846
847
848
849
850
851
852
853
854
855
856
857
858
859
860
861
862
863
864
865
866
867
868
869
870
871
872
873
874
875
876
877
878
879
880
881
882
883
884
885
886
887
888
889
890
891
892
893
894
895
896
897
898
899
900
901
902
903
904
905
906
907
908
909
910
911
912
913
914
915
916
917
918
919
920
921
922
923
924
925
926
927
928
929
930
931
932
933
934
935
936
937
938
939
940
941
942
943
944
945
946
947
948
949
950
951
952
953
954
955
956
957
958
959
960
961
962
963
964
965
966
967
968
969
970
971
972
973
974
975
976
977
978
979
980
981
982
983
984
985
986
987
988
989
990
991
992
993
994
995
996
997
998
999
1000
/*
* MinIO Cloud Storage, (C) 2016, 2017, 2018 MinIO, Inc.
*
* 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 cmd
import (
"bytes"
"context"
"io"
"io/ioutil"
"net/http"
"os"
"path"
"sort"
"strings"
"sync"
"sync/atomic"
"time"
"github.com/minio/minio-go/pkg/s3utils"
"github.com/minio/minio/cmd/logger"
"github.com/minio/minio/pkg/lock"
"github.com/minio/minio/pkg/madmin"
"github.com/minio/minio/pkg/mimedb"
"github.com/minio/minio/pkg/mountinfo"
"github.com/minio/minio/pkg/policy"
)
// Default etag is used for pre-existing objects.
var defaultEtag = "00000000000000000000000000000000-1"
// FSObjects - Implements fs object layer.
type FSObjects struct {
// Disk usage metrics
totalUsed uint64 // ref: https://golang.org/pkg/sync/atomic/#pkg-note-BUG
// Path to be exported over S3 API.
fsPath string
// meta json filename, varies by fs / cache backend.
metaJSONFile string
// Unique value to be used for all
// temporary transactions.
fsUUID string
// This value shouldn't be touched, once initialized.
fsFormatRlk *lock.RLockedFile // Is a read lock on `format.json`.
// FS rw pool.
rwPool *fsIOPool
// ListObjects pool management.
listPool *TreeWalkPool
diskMount bool
appendFileMap map[string]*fsAppendFile
appendFileMapMu sync.Mutex
// To manage the appendRoutine go-routines
nsMutex *nsLockMap
}
// Represents the background append file.
type fsAppendFile struct {
sync.Mutex
parts []PartInfo // List of parts appended.
filePath string // Absolute path of the file in the temp location.
}
// Initializes meta volume on all the fs path.
func initMetaVolumeFS(fsPath, fsUUID string) error {
// This happens for the first time, but keep this here since this
// is the only place where it can be made less expensive
// optimizing all other calls. Create minio meta volume,
// if it doesn't exist yet.
metaBucketPath := pathJoin(fsPath, minioMetaBucket)
if err := os.MkdirAll(metaBucketPath, 0777); err != nil {
return err
}
metaTmpPath := pathJoin(fsPath, minioMetaTmpBucket, fsUUID)
if err := os.MkdirAll(metaTmpPath, 0777); err != nil {
return err
}
metaMultipartPath := pathJoin(fsPath, minioMetaMultipartBucket)
return os.MkdirAll(metaMultipartPath, 0777)
}
// NewFSObjectLayer - initialize new fs object layer.
func NewFSObjectLayer(fsPath string) (ObjectLayer, error) {
ctx := context.Background()
if fsPath == "" {
return nil, errInvalidArgument
}
var err error
if fsPath, err = getValidPath(fsPath); err != nil {
if err == errMinDiskSize {
return nil, err
}
return nil, uiErrUnableToWriteInBackend(err)
}
// Assign a new UUID for FS minio mode. Each server instance
// gets its own UUID for temporary file transaction.
fsUUID := mustGetUUID()
// Initialize meta volume, if volume already exists ignores it.
if err = initMetaVolumeFS(fsPath, fsUUID); err != nil {
return nil, err
}
// Initialize `format.json`, this function also returns.
rlk, err := initFormatFS(ctx, fsPath)
if err != nil {
return nil, err
}
// Initialize fs objects.
fs := &FSObjects{
fsPath: fsPath,
metaJS
8000
ONFile: fsMetaJSONFile,
fsUUID: fsUUID,
rwPool: &fsIOPool{
readersMap: make(map[string]*lock.RLockedFile),
},
nsMutex: newNSLock(false),
listPool: NewTreeWalkPool(globalLookupTimeout),
appendFileMap: make(map[string]*fsAppendFile),
diskMount: mountinfo.IsLikelyMountPoint(fsPath),
}
// Once the filesystem has initialized hold the read lock for
// the life time of the server. This is done to ensure that under
// shared backend mode for FS, remote servers do not migrate
// or cause changes on backend format.
fs.fsFormatRlk = rlk
if !fs.diskMount {
go fs.diskUsage(GlobalServiceDoneCh)
}
go fs.cleanupStaleMultipartUploads(ctx, GlobalMultipartCleanupInterval, GlobalMultipartExpiry, GlobalServiceDoneCh)
// Return successfully initialized object layer.
return fs, nil
}
// Shutdown - should be called when process shuts down.
func (fs *FSObjects) Shutdown(ctx context.Context) error {
fs.fsFormatRlk.Close()
// Cleanup and delete tmp uuid.
return fsRemoveAll(ctx, pathJoin(fs.fsPath, minioMetaTmpBucket, fs.fsUUID))
}
// diskUsage returns du information for the posix path, in a continuous routine.
func (fs *FSObjects) diskUsage(doneCh chan struct{}) {
usageFn := func(ctx context.Context, entry string) error {
if globalHTTPServer != nil {
// Wait at max 1 minute for an inprogress request
// before proceeding to count the usage.
waitCount := 60
// Any requests in progress, delay the usage.
for globalHTTPServer.GetRequestCount() > 0 && waitCount > 0 {
waitCount--
time.Sleep(1 * time.Second)
}
}
select {
case <-doneCh:
return errWalkAbort
default:
fi, err := os.Stat(entry)
if err != nil {
err = osErrToFSFileErr(err)
return err
}
atomic.AddUint64(&fs.totalUsed, uint64(fi.Size()))
}
return nil
}
// Return this routine upon errWalkAbort, continue for any other error on purpose
// so that we can start the routine freshly in another 12 hours.
if err := getDiskUsage(context.Background(), fs.fsPath, usageFn); err == errWalkAbort {
return
}
for {
select {
case <-doneCh:
return
case <-time.After(globalUsageCheckInterval):
var usage uint64
usageFn = func(ctx context.Context, entry string) error {
if globalHTTPServer != nil {
// Wait at max 1 minute for an inprogress request
// before proceeding to count the usage.
waitCount := 60
// Any requests in progress, delay the usage.
for globalHTTPServer.GetRequestCount() > 0 && waitCount > 0 {
waitCount--
time.Sleep(1 * time.Second)
}
}
fi, err := os.Stat(entry)
if err != nil {
err = osErrToFSFileErr(err)
return err
}
usage = usage + uint64(fi.Size())
return nil
}
if err := getDiskUsage(context.Background(), fs.fsPath, usageFn); err != nil {
continue
}
atomic.StoreUint64(&fs.totalUsed, usage)
}
}
}
// StorageInfo - returns underlying storage statistics.
func (fs *FSObjects) StorageInfo(ctx context.Context) StorageInfo {
di, err := getDiskInfo(fs.fsPath)
if err != nil {
return StorageInfo{}
}
used := di.Total - di.Free
if !fs.diskMount {
used = atomic.LoadUint64(&fs.totalUsed)
}
storageInfo := StorageInfo{
Used: used,
Total: di.Total,
Available: di.Free,
}
storageInfo.Backend.Type = BackendFS
return storageInfo
}
/// Bucket operations
// getBucketDir - will convert incoming bucket names to
// corresponding valid bucket names on the backend in a platform
// compatible way for all operating systems.
func (fs *FSObjects) getBucketDir(ctx context.Context, bucket string) (string, error) {
if bucket == "" || bucket == "." || bucket == ".." {
return "", errVolumeNotFound
}
bucketDir := pathJoin(fs.fsPath, bucket)
return bucketDir, nil
}
func (fs *FSObjects) statBucketDir(ctx context.Context, bucket string) (os.FileInfo, error) {
bucketDir, err := fs.getBucketDir(ctx, bucket)
if err != nil {
return nil, err
}
st, err := fsStatVolume(ctx, bucketDir)
if err != nil {
return nil, err
}
return st, nil
}
// MakeBucketWithLocation - create a new bucket, returns if it
// already exists.
func (fs *FSObjects) MakeBucketWithLocation(ctx context.Context, bucket, location string) error {
bucketLock := fs.nsMutex.NewNSLock(bucket, "")
if err := bucketLock.GetLock(globalObjectTimeout); err != nil {
return err
}
defer bucketLock.Unlock()
// Verify if bucket is valid.
if s3utils.CheckValidBucketNameStrict(bucket) != nil {
return BucketNameInvalid{Bucket: bucket}
}
bucketDir, err := fs.getBucketDir(ctx, bucket)
if err != nil {
return toObjectErr(err, bucket)
}
if err = fsMkdir(ctx, bucketDir); err != nil {
return toObjectErr(err, bucket)
}
return nil
}
// GetBucketInfo - fetch bucket metadata info.
func (fs *FSObjects) GetBucketInfo(ctx context.Context, bucket string) (bi BucketInfo, e error) {
bucketLock := fs.nsMutex.NewNSLock(bucket, "")
if e := bucketLock.GetRLock(globalObjectTimeout); e != nil {
return bi, e
}
defer bucketLock.RUnlock()
st, err := fs.statBucketDir(ctx, bucket)
if err != nil {
return bi, toObjectErr(err, bucket)
}
// As os.Stat() doesn't carry other than ModTime(), use ModTime() as CreatedTime.
createdTime := st.ModTime()
return BucketInfo{
Name: bucket,
Created: createdTime,
}, nil
}
// ListBuckets - list all s3 compatible buckets (directories) at fsPath.
func (fs *FSObjects) ListBuckets(ctx context.Context) ([]BucketInfo, error) {
if err := checkPathLength(fs.fsPath); err != nil {
logger.LogIf(ctx, err)
return nil, err
}
var bucketInfos []BucketInfo
entries, err := readDir((fs.fsPath))
if err != nil {
logger.LogIf(ctx, errDiskNotFound)
return nil, toObjectErr(errDiskNotFound)
}
for _, entry := range entries {
// Ignore all reserved bucket names and invalid bucket names.
if isReservedOrInvalidBucket(entry, false) {
continue
}
var fi os.FileInfo
fi, err = fsStatVolume(ctx, pathJoin(fs.fsPath, entry))
// There seems like no practical reason to check for errors
// at this point, if there are indeed errors we can simply
// just ignore such buckets and list only those which
// return proper Stat information instead.
if err != nil {
// Ignore any errors returned here.
continue
}
bucketInfos = append(bucketInfos, BucketInfo{
Name: fi.Name(),
// As os.Stat() doesnt carry CreatedTime, use ModTime() as CreatedTime.
Created: fi.ModTime(),
})
}
// Sort bucket infos by bucket name.
sort.Sort(byBucketName(bucketInfos))
// Succes.
return bucketInfos, nil
}
// DeleteBucket - delete a bucket and all the metadata associated
// with the bucket including pending multipart, object metadata.
func (fs *FSObjects) DeleteBucket(ctx context.Context, bucket string) error {
bucketLock := fs.nsMutex.NewNSLock(bucket, "")
if err := bucketLock.GetLock(globalObjectTimeout); err != nil {
logger.LogIf(ctx, err)
return err
}
defer bucketLock.Unlock()
bucketDir, err := fs.getBucketDir(ctx, bucket)
if err != nil {
return toObjectErr(err, bucket)
}
// Attempt to delete regular bucket.
if err = fsRemoveDir(ctx, bucketDir); err != nil {
return toObjectErr(err, bucket)
}
// Cleanup all the bucket metadata.
minioMetadataBucketDir := pathJoin(fs.fsPath, minioMetaBucket, bucketMetaPrefix, bucket)
if err = fsRemoveAll(ctx, minioMetadataBucketDir); err != nil {
return toObjectErr(err, bucket)
}
// Delete all bucket metadata.
deleteBucketMetadata(ctx, bucket, fs)
return nil
}
/// Object Operations
// CopyObject - copy object source object to destination object.
// if source object and destination object are same we only
// update metadata.
func (fs *FSObjects) CopyObject(ctx context.Context, srcBucket, srcObject, dstBucket, dstObject string, srcInfo ObjectInfo, srcOpts, dstOpts ObjectOptions) (oi ObjectInfo, e error) {
cpSrcDstSame := isStringEqual(pathJoin(srcBucket, srcObject), pathJoin(dstBucket, dstObject))
if !cpSrcDstSame {
objectDWLock := fs.nsMutex.NewNSLock(dstBucket, dstObject)
if err := objectDWLock.GetLock(globalObjectTimeout); err != nil {
return oi, err
}
defer objectDWLock.Unlock()
}
if _, err := fs.statBucketDir(ctx, srcBucket); err != nil {
return oi, toObjectErr(err, srcBucket)
}
if cpSrcDstSame && srcInfo.metadataOnly {
fsMetaPath := pathJoin(fs.fsPath, minioMetaBucket, bucketMetaPrefix, srcBucket, srcObject, fs.metaJSONFile)
wlk, err := fs.rwPool.Write(fsMetaPath)
if err != nil {
logger.LogIf(ctx, err)
return oi, toObjectErr(err, srcBucket, srcObject)
}
// This close will allow for locks to be synchronized on `fs.json`.
defer wlk.Close()
// Save objects' metadata in `fs.json`.
fsMeta := newFSMetaV1()
if _, err = fsMeta.ReadFrom(ctx, wlk); err != nil && err != io.EOF {
return oi, toObjectErr(err, srcBucket, srcObject)
}
fsMeta.Meta = srcInfo.UserDefined
fsMeta.Meta["etag"] = srcInfo.ETag
if _, err = fsMeta.WriteTo(wlk); err != nil {
return oi, toObjectErr(err, srcBucket, srcObject)
}
// Stat the file to get file size.
fi, err := fsStatFile(ctx, pathJoin(fs.fsPath, srcBucket, srcObject))
if err != nil {
return oi, toObjectErr(err, srcBucket, srcObject)
}
// Return the new object info.
return fsMeta.ToObjectInfo(srcBucket, srcObject, fi), nil
}
if err := checkPutObjectArgs(ctx, dstBucket, dstObject, fs, srcInfo.PutObjReader.Size()); err != nil {
return ObjectInfo{}, err
}
objInfo, err := fs.putObject(ctx, dstBucket, dstObject, srcInfo.PutObjReader, ObjectOptions{ServerSideEncryption: dstOpts.ServerSideEncryption, UserDefined: srcInfo.UserDefined})
if err != nil {
return oi, toObjectErr(err, dstBucket, dstObject)
}
return objInfo, nil
}
// GetObjectNInfo - returns object info and a reader for object
// content.
func (fs *FSObjects) GetObjectNInfo(ctx context.Context, bucket, object string, rs *HTTPRangeSpec, h http.Header, lockType LockType, opts ObjectOptions) (gr *GetObjectReader, err error) {
if err = checkGetObjArgs(ctx, bucket, object); err != nil {
return nil, err
}
if _, err = fs.statBucketDir(ctx, bucket); err != nil {
return nil, toObjectErr(err, bucket)
}
var nsUnlocker = func() {}
if lockType != noLock {
// Lock the object before reading.
lock := fs.nsMutex.NewNSLock(bucket, object)
switch lockType {
case writeLock:
if err = lock.GetLock(globalObjectTimeout); err != nil {
logger.LogIf(ctx, err)
return nil, err
}
nsUnlocker = lock.Unlock
case readLock:
if err = lock.GetRLock(globalObjectTimeout); err != nil {
logger.LogIf(ctx, err)
return nil, err
}
nsUnlocker = lock.RUnlock
}
}
// Otherwise we get the object info
var objInfo ObjectInfo
if objInfo, err = fs.getObjectInfo(ctx, bucket, object); err != nil {
nsUnlocker()
return nil, toObjectErr(err, bucket, object)
}
// For a directory, we need to send an reader that returns no bytes.
if hasSuffix(object, slashSeparator) {
// The lock taken above is released when
// objReader.Close() is called by the caller.
return NewGetObjectReaderFromReader(bytes.NewBuffer(nil), objInfo, opts.CheckCopyPrecondFn, nsUnlocker)
}
// Take a rwPool lock for NFS gateway type deployment
rwPoolUnlocker := func() {}
if bucket != minioMetaBucket && lockType != noLock {
fsMetaPath := pathJoin(fs.fsPath, minioMetaBucket, bucketMetaPrefix, bucket, object, fs.metaJSONFile)
_, err = fs.rwPool.Open(fsMetaPath)
if err != nil && err != errFileNotFound {
logger.LogIf(ctx, err)
nsUnlocker()
return nil, toObjectErr(err, bucket, object)
}
// Need to clean up lock after getObject is
// completed.
rwPoolUnlocker = func() { fs.rwPool.Close(fsMetaPath) }
}
objReaderFn, off, length, rErr := NewGetObjectReader(rs, objInfo, opts.CheckCopyPrecondFn, nsUnlocker, rwPoolUnlocker)
if rErr != nil {
return nil, rErr
}
// Read the object, doesn't exist returns an s3 compatible error.
fsObjPath := pathJoin(fs.fsPath, bucket, object)
readCloser, size, err := fsOpenFile(ctx, fsObjPath, off)
if err != nil {
rwPoolUnlocker()
nsUnlocker()
return nil, toObjectErr(err, bucket, object)
}
reader := io.LimitReader(readCloser, length)
closeFn := func() {
readCloser.Close()
}
// Check if range is valid
if off > size || off+length > size {
err = InvalidRange{off, length, size}
logger.LogIf(ctx, err)
closeFn()
rwPoolUnlocker()
nsUnlocker()
return nil, err
}
return objReaderFn(reader, h, opts.CheckCopyPrecondFn, closeFn)
}
// GetObject - reads an object from the disk.
// Supports additional parameters like offset and length
// which are synonymous with HTTP Range requests.
//
// startOffset indicates the starting read location of the object.
// length indicates the total length of the object.
func (fs *FSObjects) GetObject(ctx context.Context, bucket, object string, offset int64, length int64, writer io.Writer, etag string, opts ObjectOptions) (err error) {
if err = checkGetObjArgs(ctx, bucket, object); err != nil {
return err
}
// Lock the object before reading.
objectLock := fs.nsMutex.NewNSLock(bucket, object)
if err := objectLock.GetRLock(globalObjectTimeout); err != nil {
logger.LogIf(ctx, err)
return err
}
defer objectLock.RUnlock()
return fs.getObject(ctx, bucket, object, offset, length, writer, etag, true)
}
// getObject - wrapper for GetObject
func (fs *FSObjects) getObject(ctx context.Context, bucket, object string, offset int64, length int64, writer io.Writer, etag string, lock bool) (err error) {
if _, err = fs.statBucketDir(ctx, bucket); err != nil {
return toObjectErr(err, bucket)
}
// Offset cannot be negative.
if offset < 0 {
logger.LogIf(ctx, errUnexpected)
return toObjectErr(errUnexpected, bucket, object)
}
// Writer cannot be nil.
if writer == nil {
logger.LogIf(ctx, errUnexpected)
return toObjectErr(errUnexpected, bucket, object)
}
// If its a directory request, we return an empty body.
if hasSuffix(object, slashSeparator) {
_, err = writer.Write([]byte(""))
logger.LogIf(ctx, err)
return toObjectErr(err, bucket, object)
}
if bucket != minioMetaBucket {
fsMetaPath := pathJoin(fs.fsPath, minioMetaBucket, bucketMetaPrefix, bucket, object, fs.metaJSONFile)
if lock {
_, err = fs.rwPool.Open(fsMetaPath)
if err != nil && err != errFileNotFound {
logger.LogIf(ctx, err)
return toObjectErr(err, bucket, object)
}
defer fs.rwPool.Close(fsMetaPath)
}
}
if etag != "" && etag != defaultEtag {
objEtag, perr := fs.getObjectETag(ctx, bucket, object, lock)
if perr != nil {
return toObjectErr(perr, bucket, object)
}
if objEtag != etag {
logger.LogIf(ctx, InvalidETag{})
return toObjectErr(InvalidETag{}, bucket, object)
}
}
// Read the object, doesn't exist returns an s3 compatible error.
fsObjPath := pathJoin(fs.fsPath, bucket, object)
reader, size, err := fsOpenFile(ctx, fsObjPath, offset)
if err != nil {
return toObjectErr(err, bucket, object)
}
defer reader.Close()
bufSize := int64(readSizeV1)
if length > 0 && bufSize > length {
bufSize = length
}
// For negative length we read everything.
if length < 0 {
length = size - offset
}
// Reply back invalid range if the input offset and length fall out of range.
if offset > size || offset+length > size {
err = InvalidRange{offset, length, size}
logger.LogIf(ctx, err)
return err
}
// Allocate a staging buffer.
buf := make([]byte, int(bufSize))
_, err = io.CopyBuffer(writer, io.LimitReader(reader, length), buf)
// The writer will be closed incase of range queries, which will emit ErrClosedPipe.
if err == io.ErrClosedPipe {
err = nil
}
return toObjectErr(err, bucket, object)
}
// Create a new fs.json file, if the existing one is corrupt. Should happen very rarely.
func (fs *FSObjects) createFsJSON(object, fsMetaPath string) error {
fsMeta := newFSMetaV1()
fsMeta.Meta = make(map[string]string)
fsMeta.Meta["etag"] = GenETag()
contentType := mimedb.TypeByExtension(path.Ext(object))
fsMeta.Meta["content-type"] = contentType
wlk, werr := fs.rwPool.Create(fsMetaPath)
if werr == nil {
_, err := fsMeta.WriteTo(wlk)
wlk.Close()
return err
}
return werr
}
// Used to return default etag values when a pre-existing object's meta data is queried.
func (fs *FSObjects) defaultFsJSON(object string) fsMetaV1 {
fsMeta := newFSMetaV1()
fsMeta.Meta = make(map[string]string)
fsMeta.Meta["etag"] = defaultEtag
contentType := mimedb.TypeByExtension(path.Ext(object))
fsMeta.Meta["content-type"] = contentType
return fsMeta
}
// getObjectInfo - wrapper for reading object metadata and constructs ObjectInfo.
func (fs *FSObjects) getObjectInfo(ctx context.Context, bucket, object string) (oi ObjectInfo, e error) {
fsMeta := fsMetaV1{}
if hasSuffix(object, slashSeparator) {
// Since we support PUT of a "directory" object, we allow HEAD.
if !fsIsDir(ctx, pathJoin(fs.fsPath, bucket, object)) {
return oi, errFileNotFound
}
fi, err := fsStatDir(ctx, pathJoin(fs.fsPath, bucket, object))
if err != nil {
return oi, err
}
return fsMeta.ToObjectInfo(bucket, object, fi), nil
}
fsMetaPath := pathJoin(fs.fsPath, minioMetaBucket, bucketMetaPrefix, bucket, object, fs.metaJSONFile)
// Read `fs.json` to perhaps contend with
// parallel Put() operations.
rlk, err := fs.rwPool.Open(fsMetaPath)
if err == nil {
// Read from fs metadata only if it exists.
_, rerr := fsMeta.ReadFrom(ctx, rlk.LockedFile)
fs.rwPool.Close(fsMetaPath)
if rerr != nil {
if rerr != io.EOF {
return oi, rerr
}
// Set Default ETag, if fs.json is empty
fsMeta = fs.defaultFsJSON(object)
}
}
// Return a default etag and content-type based on the object's extension.
if err == errFileNotFound {
fsMeta = fs.defaultFsJSON(object)
}
// Ignore if `fs.json` is not available, this is true for pre-existing data.
if err != nil && err != errFileNotFound {
logger.LogIf(ctx, err)
return oi, err
}
// Stat the file to get file size.
fi, err := fsStatFile(ctx, pathJoin(fs.fsPath, bucket, object))
if err != nil {
return oi, err
}
return fsMeta.ToObjectInfo(bucket, object, fi), nil
}
// getObjectInfoWithLock - reads object metadata and replies back ObjectInfo.
func (fs *FSObjects) getObjectInfoWithLock(ctx context.Context, bucket, object string) (oi ObjectInfo, e error) {
// Lock the object before reading.
objectLock := fs.nsMutex.NewNSLock(bucket, object)
if err := objectLock.GetRLock(globalObjectTimeout); err != nil {
return oi, err
}
defer objectLock.RUnlock()
if err := checkGetObjArgs(ctx, bucket, object); err != nil {
return oi, err
}
if _, err := fs.statBucketDir(ctx, bucket); err != nil {
return oi, err
}
if strings.HasSuffix(object, slashSeparator) && !fs.isObjectDir(bucket, object) {
return oi, errFileNotFound
}
return fs.getObjectInfo(ctx, bucket, object)
}
// GetObjectInfo - reads object metadata and replies back ObjectInfo.
func (fs *FSObjects) GetObjectInfo(ctx context.Context, bucket, object string, opts ObjectOptions) (oi ObjectInfo, e error) {
oi, err := fs.getObjectInfoWithLock(ctx, bucket, object)
if err == errCorruptedFormat || err == io.EOF {
objectLock := fs.nsMutex.NewNSLock(bucket, object)
if err = objectLock.GetLock(globalObjectTimeout); err != nil {
return oi, toObjectErr(err, bucket, object)
}
fsMetaPath := pathJoin(fs.fsPath, minioMetaBucket, bucketMetaPrefix, bucket, object, fs.metaJSONFile)
err = fs.createFsJSON(object, fsMetaPath)
objectLock.Unlock()
if err != nil {
return oi, toObjectErr(err, bucket, object)
}
oi, err = fs.getObjectInfoWithLock(ctx, bucket, object)
}
return oi, toObjectErr(err, bucket, object)
}
// This function does the following check, suppose
// object is "a/b/c/d", stat makes sure that objects ""a/b/c""
// "a/b" and "a" do not exist.
func (fs *FSObjects) parentDirIsObject(ctx context.Context, bucket, parent string) bool {
var isParentDirObject func(string) bool
isParentDirObject = func(p string) bool {
if p == "." || p == "/" {
return false
}
if fsIsFile(ctx, pathJoin(fs.fsPath, bucket, p)) {
// If there is already a file at prefix "p", return true.
return true
}
// Check if there is a file as one of the parent paths.
return isParentDirObject(path.Dir(p))
}
return isParentDirObject(parent)
}
// PutObject - creates an object upon reading from the input stream
// until EOF, writes data directly to configured filesystem path.
// Additionally writes `fs.json` which carries the necessary metadata
// for future object operations.
func (fs *FSObjects) PutObject(ctx context.Context, bucket string, object string, r *PutObjReader, opts ObjectOptions) (objInfo ObjectInfo, retErr error) {
if err := checkPutObjectArgs(ctx, bucket, object, fs, r.Size()); err != nil {
return ObjectInfo{}, err
}
// Lock the object.
objectLock := fs.nsMutex.NewNSLock(bucket, object)
if err := objectLock.GetLock(globalObjectTimeout); err != nil {
logger.LogIf(ctx, err)
return objInfo, err
}
defer objectLock.Unlock()
return fs.putObject(ctx, bucket, object, r, opts)
}
// putObject - wrapper for PutObject
func (fs *FSObjects) putObject(ctx context.Context, bucket string, object string, r *PutObjReader, opts ObjectOptions) (objInfo ObjectInfo, retErr error) {
data := r.Reader
// No metadata is set, allocate a new one.
meta := make(map[string]string)
for k, v := range opts.UserDefined {
meta[k] = v
}
var err error
// Validate if bucket name is valid and exists.
if _, err = fs.statBucketDir(ctx, bucket); err != nil {
return ObjectInfo{}, toObjectErr(err, bucket)
}
fsMeta := newFSMetaV1()
fsMeta.Meta = meta
// This is a special case with size as '0' and object ends
// with a slash separator, we treat it like a valid operation
// and return success.
if isObjectDir(object, data.Size()) {
// Check if an object is present as one of the parent dir.
if fs.parentDirIsObject(ctx, bucket, path.Dir(object)) {
return ObjectInfo{}, toObjectErr(errFileParentIsFile, bucket, object)
}
if err = mkdirAll(pathJoin(fs.fsPath, bucket, object), 0777); err != nil {
logger.LogIf(ctx, err)
return ObjectInfo{}, toObjectErr(err, bucket, object)
}
var fi os.FileInfo
if fi, err = fsStatDir(ctx, pathJoin(fs.fsPath, bucket, object)); err != nil {
return ObjectInfo{}, toObjectErr(err, bucket, object)
}
return fsMeta.ToObjectInfo(bucket, object, fi), nil
}
// Check if an object is present as one of the parent dir.
if fs.parentDirIsObject(ctx, bucket, path.Dir(object)) {
return ObjectInfo{}, toObjectErr(errFileParentIsFile, bucket, object)
}
// Validate input data size and it can never be less than zero.
if data.Size() < -1 {
logger.LogIf(ctx, errInvalidArgument)
return ObjectInfo{}, errInvalidArgument
}
var wlk *lock.LockedFile
if bucket != minioMetaBucket {
bucketMetaDir := pathJoin(fs.fsPath, minioMetaBucket, bucketMetaPrefix)
fsMetaPath := pathJoin(bucketMetaDir, bucket, object, fs.metaJSONFile)
wlk, err = fs.rwPool.Create(fsMetaPath)
if err != nil {
logger.LogIf(ctx, err)
return ObjectInfo{}, toObjectErr(err, bucket, object)
}
// This close will allow for locks to be synchronized on `fs.json`.
defer wlk.Close()
defer func() {
// Remove meta file when PutObject encounters any error
if retErr != nil {
tmpDir := pathJoin(fs.fsPath, minioMetaTmpBucket, fs.fsUUID)
fsRemoveMeta(ctx, bucketMetaDir, fsMetaPath, tmpDir)
}
}()
}
// Uploaded object will first be written to the temporary location which will eventually
// be renamed to the actual location. It is first written to the temporary location
// so that cleaning it up will be easy if the server goes down.
tempObj := mustGetUUID()
// Allocate a buffer to Read() from request body
bufSize := int64(readSizeV1)
if size := data.Size(); size > 0 && bufSize > size {
bufSize = size
}
buf := make([]byte, int(bufSize))
fsTmpObjPath := pathJoin(fs.fsPath, minioMetaTmpBucket, fs.fsUUID, tempObj)
bytesWritten, err := fsCreateFile(ctx, fsTmpObjPath, data, buf, data.Size())
if err != nil {
fsRemoveFile(ctx, fsTmpObjPath)
return ObjectInfo{}, toObjectErr(err, bucket, object)
}
fsMeta.Meta["etag"] = r.MD5CurrentHexString()
// Should return IncompleteBody{} error when reader has fewer
// bytes than specified in request header.
if bytesWritten < data.Size() {
fsRemoveFile(ctx, fsTmpObjPath)
return ObjectInfo{}, IncompleteBody{}
}
// Delete the temporary object in the case of a
// failure. If PutObject succeeds, then there would be
// nothing to delete.
defer fsRemoveFile(ctx, fsTmpObjPath)
// Entire object was written to the temp location, now it's safe to rename it to the actual location.
fsNSObjPath := pathJoin(fs.fsPath, bucket, object)
// Deny if WORM is enabled
if globalWORMEnabled {
if _, err = fsStatFile(ctx, fsNSObjPath); err == nil {
return ObjectInfo{}, ObjectAlreadyExists{Bucket: bucket, Object: object}
}
}
if err = fsRenameFile(ctx, fsTmpObjPath, fsNSObjPath); err != nil {
return ObjectInfo{}, toObjectErr(err, bucket, object)
}
if bucket != minioMetaBucket {
// Write FS metadata after a successful namespace operation.
if _, err = fsMeta.WriteTo(wlk); err != nil {
return ObjectInfo{}, toObjectErr(err, bucket, object)
}
}
// Stat the file to fetch timestamp, size.
fi, err := fsStatFile(ctx, pathJoin(fs.fsPath, bucket, object))
if err != nil {
return ObjectInfo{}, toObjectErr(err, bucket, object)
}
// Success.
return fsMeta.ToObjectInfo(bucket, object, fi), nil
}
// DeleteObjects - deletes an object from a bucket, this operation is destructive
// and there are no rollbacks supported.
func (fs *FSObjects) DeleteObjects(ctx context.Context, bucket string, objects []string) ([]error, error) {
errs := make([]error, len(objects))
for idx, object := range objects {
errs[idx] = fs.DeleteObject(ctx, bucket, object)
}
return errs, nil
}
// DeleteObject - deletes an object from a bucket, this operation is destructive
// and there are no rollbacks supported.
func (fs *FSObjects) DeleteObject(ctx context.Context, bucket, object string) error {
// Acquire a write lock before deleting the object.
objectLock := fs.nsMutex.NewNSLock(bucket, object)
if err := objectLock.GetLock(globalOperationTimeout); err != nil {
return err
}
defer objectLock.Unlock()
if err := checkDelObjArgs(ctx, bucket, object); err != nil {
return err
}
if _, err := fs.statBucketDir(ctx, bucket); err != nil {
return toObjectErr(err, bucket)
}
minioMetaBucketDir := pathJoin(fs.fsPath, minioMetaBucket)
fsMetaPath := pathJoin(minioMetaBucketDir, bucketMetaPrefix, bucket, object, fs.metaJSONFile)
if bucket != minioMetaBucket {
rwlk, lerr := fs.rwPool.Write(fsMetaPath)
if lerr == nil {
// This close will allow for fs locks to be synchronized on `fs.json`.
defer rwlk.Close()
}
if lerr != nil && lerr != errFileNotFound {
logger.LogIf(ctx, lerr)
return toObjectErr(lerr, bucket, object)
}
}
// Delete the object.
if err := fsDeleteFile(ctx, pathJoin(fs.fsPath, bucket), pathJoin(fs.fsPath, bucket, object)); err != nil {
return toObjectErr(err, bucket, object)