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
|
package weed_server
import (
"context"
"fmt"
"io"
"math"
"os"
"github.com/chrislusf/seaweedfs/weed/glog"
"github.com/chrislusf/seaweedfs/weed/operation"
"github.com/chrislusf/seaweedfs/weed/pb/volume_server_pb"
"github.com/chrislusf/seaweedfs/weed/storage"
"github.com/chrislusf/seaweedfs/weed/storage/erasure_coding"
"github.com/chrislusf/seaweedfs/weed/storage/needle"
)
/*
Steps to apply erasure coding to .dat .idx files
0. ensure the volume is readonly
1. client call VolumeEcShardsGenerate to generate the .ecx and .ec01~.ec14 files
2. client ask master for possible servers to hold the ec files, at least 4 servers
3. client call VolumeEcShardsCopy on above target servers to copy ec files from the source server
4. target servers report the new ec files to the master
5. master stores vid -> [14]*DataNode
6. client checks master. If all 14 slices are ready, delete the original .idx, .idx files
*/
// VolumeEcShardsGenerate generates the .ecx and .ec01 ~ .ec14 files
func (vs *VolumeServer) VolumeEcShardsGenerate(ctx context.Context, req *volume_server_pb.VolumeEcShardsGenerateRequest) (*volume_server_pb.VolumeEcShardsGenerateResponse, error) {
v := vs.store.GetVolume(needle.VolumeId(req.VolumeId))
if v == nil {
return nil, fmt.Errorf("volume %d not found", req.VolumeId)
}
baseFileName := v.FileName()
if v.Collection != req.Collection {
return nil, fmt.Errorf("existing collection:%v unexpected input: %v", v.Collection, req.Collection)
}
// write .ecx file
if err := erasure_coding.WriteSortedEcxFile(baseFileName); err != nil {
return nil, fmt.Errorf("WriteSortedEcxFile %s: %v", baseFileName, err)
}
// write .ec01 ~ .ec14 files
if err := erasure_coding.WriteEcFiles(baseFileName); err != nil {
return nil, fmt.Errorf("WriteEcFiles %s: %v", baseFileName, err)
}
return &volume_server_pb.VolumeEcShardsGenerateResponse{}, nil
}
// VolumeEcShardsCopy copy the .ecx and some ec data slices
func (vs *VolumeServer) VolumeEcShardsCopy(ctx context.Context, req *volume_server_pb.VolumeEcShardsCopyRequest) (*volume_server_pb.VolumeEcShardsCopyResponse, error) {
location := vs.store.FindFreeLocation()
if location == nil {
return nil, fmt.Errorf("no space left")
}
baseFileName := storage.VolumeFileName(req.Collection, location.Directory, int(req.VolumeId))
err := operation.WithVolumeServerClient(req.SourceDataNode, vs.grpcDialOption, func(client volume_server_pb.VolumeServerClient) error {
// copy ecx file
if err := vs.doCopyFile(ctx, client, req.VolumeId, math.MaxUint32, math.MaxInt64, baseFileName, ".ecx"); err != nil {
return err
}
// copy ec data slices
for _, shardId := range req.ShardIds {
if err := vs.doCopyFile(ctx, client, req.VolumeId, math.MaxUint32, math.MaxInt64, baseFileName, erasure_coding.ToExt(int(shardId))); err != nil {
return err
}
}
return nil
})
if err != nil {
return nil, fmt.Errorf("VolumeEcShardsCopy volume %d: %v", req.VolumeId, err)
}
return &volume_server_pb.VolumeEcShardsCopyResponse{}, nil
}
// VolumeEcShardsDelete local delete the .ecx and some ec data slices if not needed, assuming current server has the source volume
func (vs *VolumeServer) VolumeEcShardsDelete(ctx context.Context, req *volume_server_pb.VolumeEcShardsDeleteRequest) (*volume_server_pb.VolumeEcShardsDeleteResponse, error) {
foundExistingVolume, err := vs.doDeleteUnmountedShards(ctx, req)
if err != nil {
return nil, err
}
if !foundExistingVolume {
err = vs.doDeleteMountedShards(ctx, req)
}
return &volume_server_pb.VolumeEcShardsDeleteResponse{}, err
}
// VolumeEcShardsDelete local delete the .ecx and some ec data slices if not needed, assuming current server has the source volume
func (vs *VolumeServer) doDeleteUnmountedShards(ctx context.Context, req *volume_server_pb.VolumeEcShardsDeleteRequest) (foundVolume bool, err error) {
v := vs.store.GetVolume(needle.VolumeId(req.VolumeId))
if v == nil {
return false, nil
}
baseFileName := v.FileName()
for _, shardId := range req.ShardIds {
if err := os.Remove(baseFileName + erasure_coding.ToExt(int(shardId))); err != nil {
return true, err
}
}
if req.ShouldDeleteEcx {
if err := os.Remove(baseFileName + ".ecx"); err != nil {
return true, err
}
}
return true, nil
}
// VolumeEcShardsDelete local delete the .ecx and some ec data slices if not needed, assuming current server has the source volume
func (vs *VolumeServer) doDeleteMountedShards(ctx context.Context, req *volume_server_pb.VolumeEcShardsDeleteRequest) (error) {
ecv, found := vs.store.FindEcVolume(needle.VolumeId(req.VolumeId))
if !found {
return fmt.Errorf("volume %d not found", req.VolumeId)
}
for _, shardId := range req.ShardIds {
if shard, found := ecv.DeleteEcVolumeShard(erasure_coding.ShardId(shardId)); found {
shard.Destroy()
}
}
if len(ecv.Shards) == 0 {
vs.store.DestroyEcVolume(needle.VolumeId(req.VolumeId))
}
return nil
}
func (vs *VolumeServer) VolumeEcShardsMount(ctx context.Context, req *volume_server_pb.VolumeEcShardsMountRequest) (*volume_server_pb.VolumeEcShardsMountResponse, error) {
for _, shardId := range req.ShardIds {
err := vs.store.MountEcShards(req.Collection, needle.VolumeId(req.VolumeId), erasure_coding.ShardId(shardId))
if err != nil {
glog.Errorf("ec shard mount %v: %v", req, err)
} else {
glog.V(2).Infof("ec shard mount %v", req)
}
if err != nil {
return nil, fmt.Errorf("mount %d.%d: %v", req.VolumeId, shardId, err)
}
}
return &volume_server_pb.VolumeEcShardsMountResponse{}, nil
}
func (vs *VolumeServer) VolumeEcShardsUnmount(ctx context.Context, req *volume_server_pb.VolumeEcShardsUnmountRequest) (*volume_server_pb.VolumeEcShardsUnmountResponse, error) {
for _, shardId := range req.ShardIds {
err := vs.store.UnmountEcShards(needle.VolumeId(req.VolumeId), erasure_coding.ShardId(shardId))
if err != nil {
glog.Errorf("ec shard unmount %v: %v", req, err)
} else {
glog.V(2).Infof("ec shard unmount %v", req)
}
if err != nil {
return nil, fmt.Errorf("unmount %d.%d: %v", req.VolumeId, shardId, err)
}
}
return &volume_server_pb.VolumeEcShardsUnmountResponse{}, nil
}
func (vs *VolumeServer) VolumeEcShardRead(req *volume_server_pb.VolumeEcShardReadRequest, stream volume_server_pb.VolumeServer_VolumeEcShardReadServer) error {
ecVolume, found := vs.store.FindEcVolume(needle.VolumeId(req.VolumeId))
if !found {
return fmt.Errorf("not found ec volume id %d", req.VolumeId)
}
ecShard, found := ecVolume.FindEcVolumeShard(erasure_coding.ShardId(req.ShardId))
if !found {
return fmt.Errorf("not found ec shard %d.%d", req.VolumeId, req.ShardId)
}
bufSize := req.Size
if bufSize > BufferSizeLimit {
bufSize = BufferSizeLimit
}
buffer := make([]byte, bufSize)
startOffset, bytesToRead := req.Offset, req.Size
for bytesToRead > 0 {
bytesread, err := ecShard.ReadAt(buffer, startOffset)
// println(fileName, "read", bytesread, "bytes, with target", bytesToRead)
if bytesread > 0 {
if int64(bytesread) > bytesToRead {
bytesread = int(bytesToRead)
}
err = stream.Send(&volume_server_pb.VolumeEcShardReadResponse{
Data: buffer[:bytesread],
})
if err != nil {
// println("sending", bytesread, "bytes err", err.Error())
return err
}
bytesToRead -= int64(bytesread)
}
if err != nil {
if err != io.EOF {
return err
}
return nil
}
}
return nil
}
|
