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
|
package weed_server
import (
"bytes"
"crypto/md5"
"fmt"
"golang.org/x/exp/slices"
"hash"
"io"
"net/http"
"strconv"
"strings"
"sync"
"sync/atomic"
"time"
"github.com/seaweedfs/seaweedfs/weed/filer"
"github.com/seaweedfs/seaweedfs/weed/glog"
"github.com/seaweedfs/seaweedfs/weed/operation"
"github.com/seaweedfs/seaweedfs/weed/pb/filer_pb"
"github.com/seaweedfs/seaweedfs/weed/security"
"github.com/seaweedfs/seaweedfs/weed/stats"
"github.com/seaweedfs/seaweedfs/weed/util"
)
var bufPool = sync.Pool{
New: func() interface{} {
return new(bytes.Buffer)
},
}
func (fs *FilerServer) uploadReaderToChunks(w http.ResponseWriter, r *http.Request, reader io.Reader, chunkSize int32, fileName, contentType string, contentLength int64, so *operation.StorageOption) (fileChunks []*filer_pb.FileChunk, md5Hash hash.Hash, chunkOffset int64, uploadErr error, smallContent []byte) {
query := r.URL.Query()
isAppend := isAppend(r)
if query.Has("offset") {
offset := query.Get("offset")
offsetInt, err := strconv.ParseInt(offset, 10, 64)
if err != nil || offsetInt < 0 {
err = fmt.Errorf("invalid 'offset': '%s'", offset)
return nil, nil, 0, err, nil
}
if isAppend && offsetInt > 0 {
err = fmt.Errorf("cannot set offset when op=append")
return nil, nil, 0, err, nil
}
chunkOffset = offsetInt
}
md5Hash = md5.New()
var partReader = io.NopCloser(io.TeeReader(reader, md5Hash))
var wg sync.WaitGroup
var bytesBufferCounter int64
bytesBufferLimitCond := sync.NewCond(new(sync.Mutex))
var fileChunksLock sync.Mutex
for {
// need to throttle used byte buffer
bytesBufferLimitCond.L.Lock()
for atomic.LoadInt64(&bytesBufferCounter) >= 4 {
glog.V(4).Infof("waiting for byte buffer %d", bytesBufferCounter)
bytesBufferLimitCond.Wait()
}
atomic.AddInt64(&bytesBufferCounter, 1)
bytesBufferLimitCond.L.Unlock()
bytesBuffer := bufPool.Get().(*bytes.Buffer)
glog.V(4).Infof("received byte buffer %d", bytesBufferCounter)
limitedReader := io.LimitReader(partReader, int64(chunkSize))
bytesBuffer.Reset()
dataSize, err := bytesBuffer.ReadFrom(limitedReader)
// data, err := io.ReadAll(limitedReader)
if err != nil || dataSize == 0 {
bufPool.Put(bytesBuffer)
atomic.AddInt64(&bytesBufferCounter, -1)
bytesBufferLimitCond.Signal()
uploadErr = err
break
}
if chunkOffset == 0 && !isAppend {
if dataSize < fs.option.SaveToFilerLimit || strings.HasPrefix(r.URL.Path, filer.DirectoryEtcRoot) {
chunkOffset += dataSize
smallContent = make([]byte, dataSize)
bytesBuffer.Read(smallContent)
bufPool.Put(bytesBuffer)
atomic.AddInt64(&bytesBufferCounter, -1)
bytesBufferLimitCond.Signal()
break
}
}
wg.Add(1)
go func(offset int64) {
defer func() {
bufPool.Put(bytesBuffer)
atomic.AddInt64(&bytesBufferCounter, -1)
bytesBufferLimitCond.Signal()
wg.Done()
}()
chunk, toChunkErr := fs.dataToChunk(fileName, contentType, bytesBuffer.Bytes(), offset, so)
if toChunkErr != nil {
uploadErr = toChunkErr
}
if chunk != nil {
fileChunksLock.Lock()
fileChunks = append(fileChunks, chunk)
fileChunksLock.Unlock()
glog.V(4).Infof("uploaded %s chunk %d to %s [%d,%d)", fileName, len(fileChunks), chunk.FileId, offset, offset+int64(chunk.Size))
}
}(chunkOffset)
// reset variables for the next chunk
chunkOffset = chunkOffset + dataSize
// if last chunk was not at full chunk size, but already exhausted the reader
if dataSize < int64(chunkSize) {
break
}
}
wg.Wait()
if uploadErr != nil {
fs.filer.DeleteChunks(fileChunks)
return nil, md5Hash, 0, uploadErr, nil
}
slices.SortFunc(fileChunks, func(a, b *filer_pb.FileChunk) bool {
return a.Offset < b.Offset
})
return fileChunks, md5Hash, chunkOffset, nil, smallContent
}
func (fs *FilerServer) doUpload(urlLocation string, limitedReader io.Reader, fileName string, contentType string, pairMap map[string]string, auth security.EncodedJwt) (*operation.UploadResult, error, []byte) {
stats.FilerRequestCounter.WithLabelValues("chunkUpload").Inc()
start := time.Now()
defer func() {
stats.FilerRequestHistogram.WithLabelValues("chunkUpload").Observe(time.Since(start).Seconds())
}()
uploadOption := &operation.UploadOption{
UploadUrl: urlLocation,
Filename: fileName,
Cipher: fs.option.Cipher,
IsInputCompressed: false,
MimeType: contentType,
PairMap: pairMap,
Jwt: auth,
}
uploadResult, err, data := operation.Upload(limitedReader, uploadOption)
if uploadResult != nil && uploadResult.RetryCount > 0 {
stats.FilerRequestCounter.WithLabelValues("chunkUploadRetry").Add(float64(uploadResult.RetryCount))
}
return uploadResult, err, data
}
func (fs *FilerServer) dataToChunk(fileName, contentType string, data []byte, chunkOffset int64, so *operation.StorageOption) (*filer_pb.FileChunk, error) {
dataReader := util.NewBytesReader(data)
// retry to assign a different file id
var fileId, urlLocation string
var auth security.EncodedJwt
var uploadErr error
var uploadResult *operation.UploadResult
for i := 0; i < 3; i++ {
// assign one file id for one chunk
fileId, urlLocation, auth, uploadErr = fs.assignNewFileInfo(so)
if uploadErr != nil {
glog.V(4).Infof("retry later due to assign error: %v", uploadErr)
time.Sleep(time.Duration(i+1) * 251 * time.Millisecond)
continue
}
// upload the chunk to the volume server
uploadResult, uploadErr, _ = fs.doUpload(urlLocation, dataReader, fileName, contentType, nil, auth)
if uploadErr != nil {
glog.V(4).Infof("retry later due to upload error: %v", uploadErr)
time.Sleep(time.Duration(i+1) * 251 * time.Millisecond)
continue
}
break
}
if uploadErr != nil {
glog.Errorf("upload error: %v", uploadErr)
return nil, uploadErr
}
// if last chunk exhausted the reader exactly at the border
if uploadResult.Size == 0 {
return nil, nil
}
return uploadResult.ToPbFileChunk(fileId, chunkOffset), nil
}
|
