aboutsummaryrefslogtreecommitdiff
path: root/weed/wdclient/masterclient.go
blob: da46a440b691f6d902effd8ff54d36161287d95c (plain)
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
package wdclient

import (
	"context"
	"fmt"
	"math/rand"
	"sync"
	"time"

	"github.com/seaweedfs/seaweedfs/weed/stats"

	"github.com/seaweedfs/seaweedfs/weed/util"
	"google.golang.org/grpc"

	"github.com/seaweedfs/seaweedfs/weed/glog"
	"github.com/seaweedfs/seaweedfs/weed/pb"
	"github.com/seaweedfs/seaweedfs/weed/pb/master_pb"
)

type MasterClient struct {
	FilerGroup        string
	clientType        string
	clientHost        pb.ServerAddress
	rack              string
	currentMaster     pb.ServerAddress
	currentMasterLock sync.RWMutex
	masters           pb.ServerDiscovery
	grpcDialOption    grpc.DialOption

	*vidMap
	vidMapCacheSize  int
	OnPeerUpdate     func(update *master_pb.ClusterNodeUpdate, startFrom time.Time)
	OnPeerUpdateLock sync.RWMutex
}

func NewMasterClient(grpcDialOption grpc.DialOption, filerGroup string, clientType string, clientHost pb.ServerAddress, clientDataCenter string, rack string, masters pb.ServerDiscovery) *MasterClient {
	return &MasterClient{
		FilerGroup:      filerGroup,
		clientType:      clientType,
		clientHost:      clientHost,
		rack:            rack,
		masters:         masters,
		grpcDialOption:  grpcDialOption,
		vidMap:          newVidMap(clientDataCenter),
		vidMapCacheSize: 5,
	}
}

func (mc *MasterClient) SetOnPeerUpdateFn(onPeerUpdate func(update *master_pb.ClusterNodeUpdate, startFrom time.Time)) {
	mc.OnPeerUpdateLock.Lock()
	mc.OnPeerUpdate = onPeerUpdate
	mc.OnPeerUpdateLock.Unlock()
}

func (mc *MasterClient) GetLookupFileIdFunction() LookupFileIdFunctionType {
	return mc.LookupFileIdWithFallback
}

func (mc *MasterClient) LookupFileIdWithFallback(fileId string) (fullUrls []string, err error) {
	fullUrls, err = mc.vidMap.LookupFileId(fileId)
	if err == nil && len(fullUrls) > 0 {
		return
	}
	err = pb.WithMasterClient(false, mc.GetMaster(context.Background()), mc.grpcDialOption, false, func(client master_pb.SeaweedClient) error {
		resp, err := client.LookupVolume(context.Background(), &master_pb.LookupVolumeRequest{
			VolumeOrFileIds: []string{fileId},
		})
		if err != nil {
			return fmt.Errorf("LookupVolume %s failed: %v", fileId, err)
		}
		for vid, vidLocation := range resp.VolumeIdLocations {
			for _, vidLoc := range vidLocation.Locations {
				loc := Location{
					Url:        vidLoc.Url,
					PublicUrl:  vidLoc.PublicUrl,
					GrpcPort:   int(vidLoc.GrpcPort),
					DataCenter: vidLoc.DataCenter,
				}
				mc.vidMap.addLocation(uint32(vid), loc)
				httpUrl := "http://" + loc.Url + "/" + fileId
				// Prefer same data center
				if mc.DataCenter != "" && mc.DataCenter == loc.DataCenter {
					fullUrls = append([]string{httpUrl}, fullUrls...)
				} else {
					fullUrls = append(fullUrls, httpUrl)
				}
			}
		}
		return nil
	})
	return
}

func (mc *MasterClient) getCurrentMaster() pb.ServerAddress {
	mc.currentMasterLock.RLock()
	defer mc.currentMasterLock.RUnlock()
	return mc.currentMaster
}

func (mc *MasterClient) setCurrentMaster(master pb.ServerAddress) {
	mc.currentMasterLock.Lock()
	mc.currentMaster = master
	mc.currentMasterLock.Unlock()
}

func (mc *MasterClient) GetMaster(ctx context.Context) pb.ServerAddress {
	mc.WaitUntilConnected(ctx)
	return mc.getCurrentMaster()
}

func (mc *MasterClient) GetMasters(ctx context.Context) []pb.ServerAddress {
	mc.WaitUntilConnected(ctx)
	return mc.masters.GetInstances()
}

func (mc *MasterClient) WaitUntilConnected(ctx context.Context) {
	for {
		select {
		case <-ctx.Done():
			glog.V(0).Infof("Connection wait stopped: %v", ctx.Err())
			return
		default:
			if mc.getCurrentMaster() != "" {
				return
			}
			time.Sleep(time.Duration(rand.Int31n(200)) * time.Millisecond)
			print(".")
		}
	}
}

func (mc *MasterClient) KeepConnectedToMaster(ctx context.Context) {
	glog.V(1).Infof("%s.%s masterClient bootstraps with masters %v", mc.FilerGroup, mc.clientType, mc.masters)
	for {
		select {
		case <-ctx.Done():
			glog.V(0).Infof("Connection to masters stopped: %v", ctx.Err())
			return
		default:
			mc.tryAllMasters(ctx)
			time.Sleep(time.Second)
		}
	}
}

func (mc *MasterClient) FindLeaderFromOtherPeers(myMasterAddress pb.ServerAddress) (leader string) {
	for _, master := range mc.masters.GetInstances() {
		if master == myMasterAddress {
			continue
		}
		if grpcErr := pb.WithMasterClient(false, master, mc.grpcDialOption, false, func(client master_pb.SeaweedClient) error {
			ctx, cancel := context.WithTimeout(context.Background(), 120*time.Millisecond)
			defer cancel()
			resp, err := client.GetMasterConfiguration(ctx, &master_pb.GetMasterConfigurationRequest{})
			if err != nil {
				return err
			}
			leader = resp.Leader
			return nil
		}); grpcErr != nil {
			glog.V(0).Infof("connect to %s: %v", master, grpcErr)
		}
		if leader != "" {
			glog.V(0).Infof("existing leader is %s", leader)
			return
		}
	}
	glog.V(0).Infof("No existing leader found!")
	return
}

func (mc *MasterClient) tryAllMasters(ctx context.Context) {
	var nextHintedLeader pb.ServerAddress
	mc.masters.RefreshBySrvIfAvailable()
	for _, master := range mc.masters.GetInstances() {
		nextHintedLeader = mc.tryConnectToMaster(ctx, master)
		for nextHintedLeader != "" {
			select {
			case <-ctx.Done():
				glog.V(0).Infof("Connection attempt to all masters stopped: %v", ctx.Err())
				return
			default:
				nextHintedLeader = mc.tryConnectToMaster(ctx, nextHintedLeader)
			}
		}
		mc.setCurrentMaster("")
	}
}

func (mc *MasterClient) tryConnectToMaster(ctx context.Context, master pb.ServerAddress) (nextHintedLeader pb.ServerAddress) {
	glog.V(1).Infof("%s.%s masterClient Connecting to master %v", mc.FilerGroup, mc.clientType, master)
	stats.MasterClientConnectCounter.WithLabelValues("total").Inc()
	gprcErr := pb.WithMasterClient(true, master, mc.grpcDialOption, false, func(client master_pb.SeaweedClient) error {
		ctx, cancel := context.WithCancel(ctx)
		defer cancel()

		stream, err := client.KeepConnected(ctx)
		if err != nil {
			glog.V(1).Infof("%s.%s masterClient failed to keep connected to %s: %v", mc.FilerGroup, mc.clientType, master, err)
			stats.MasterClientConnectCounter.WithLabelValues(stats.FailedToKeepConnected).Inc()
			return err
		}

		if err = stream.Send(&master_pb.KeepConnectedRequest{
			FilerGroup:    mc.FilerGroup,
			DataCenter:    mc.DataCenter,
			Rack:          mc.rack,
			ClientType:    mc.clientType,
			ClientAddress: string(mc.clientHost),
			Version:       util.Version(),
		}); err != nil {
			glog.V(0).Infof("%s.%s masterClient failed to send to %s: %v", mc.FilerGroup, mc.clientType, master, err)
			stats.MasterClientConnectCounter.WithLabelValues(stats.FailedToSend).Inc()
			return err
		}
		glog.V(1).Infof("%s.%s masterClient Connected to %v", mc.FilerGroup, mc.clientType, master)

		resp, err := stream.Recv()
		if err != nil {
			glog.V(0).Infof("%s.%s masterClient failed to receive from %s: %v", mc.FilerGroup, mc.clientType, master, err)
			stats.MasterClientConnectCounter.WithLabelValues(stats.FailedToReceive).Inc()
			return err
		}

		// check if it is the leader to determine whether to reset the vidMap
		if resp.VolumeLocation != nil {
			if resp.VolumeLocation.Leader != "" && string(master) != resp.VolumeLocation.Leader {
				glog.V(0).Infof("master %v redirected to leader %v", master, resp.VolumeLocation.Leader)
				nextHintedLeader = pb.ServerAddress(resp.VolumeLocation.Leader)
				stats.MasterClientConnectCounter.WithLabelValues(stats.RedirectedToLeader).Inc()
				return nil
			}
			mc.resetVidMap()
			mc.updateVidMap(resp)
		} else {
			mc.resetVidMap()
		}
		mc.setCurrentMaster(master)

		for {
			resp, err := stream.Recv()
			if err != nil {
				glog.V(0).Infof("%s.%s masterClient failed to receive from %s: %v", mc.FilerGroup, mc.clientType, master, err)
				stats.MasterClientConnectCounter.WithLabelValues(stats.FailedToReceive).Inc()
				return err
			}

			if resp.VolumeLocation != nil {
				// maybe the leader is changed
				if resp.VolumeLocation.Leader != "" && string(mc.GetMaster(ctx)) != resp.VolumeLocation.Leader {
					glog.V(0).Infof("currentMaster %v redirected to leader %v", mc.GetMaster(ctx), resp.VolumeLocation.Leader)
					nextHintedLeader = pb.ServerAddress(resp.VolumeLocation.Leader)
					stats.MasterClientConnectCounter.WithLabelValues(stats.RedirectedToLeader).Inc()
					return nil
				}
				mc.updateVidMap(resp)
			}

			if resp.ClusterNodeUpdate != nil {
				update := resp.ClusterNodeUpdate
				mc.OnPeerUpdateLock.RLock()
				if mc.OnPeerUpdate != nil {
					if update.FilerGroup == mc.FilerGroup {
						if update.IsAdd {
							glog.V(0).Infof("+ %s@%s noticed %s.%s %s\n", mc.clientType, mc.clientHost, update.FilerGroup, update.NodeType, update.Address)
						} else {
							glog.V(0).Infof("- %s@%s noticed %s.%s %s\n", mc.clientType, mc.clientHost, update.FilerGroup, update.NodeType, update.Address)
						}
						stats.MasterClientConnectCounter.WithLabelValues(stats.OnPeerUpdate).Inc()
						mc.OnPeerUpdate(update, time.Now())
					}
				}
				mc.OnPeerUpdateLock.RUnlock()
			}
			if err := ctx.Err(); err != nil {
				glog.V(0).Infof("Connection attempt to master stopped: %v", err)
				return err
			}
		}
	})
	if gprcErr != nil {
		stats.MasterClientConnectCounter.WithLabelValues(stats.Failed).Inc()
		glog.V(1).Infof("%s.%s masterClient failed to connect with master %v: %v", mc.FilerGroup, mc.clientType, master, gprcErr)
	}
	return
}

func (mc *MasterClient) updateVidMap(resp *master_pb.KeepConnectedResponse) {
	if resp.VolumeLocation.IsEmptyUrl() {
		glog.V(0).Infof("updateVidMap ignore short heartbeat: %+v", resp)
		return
	}
	// process new volume location
	loc := Location{
		Url:        resp.VolumeLocation.Url,
		PublicUrl:  resp.VolumeLocation.PublicUrl,
		DataCenter: resp.VolumeLocation.DataCenter,
		GrpcPort:   int(resp.VolumeLocation.GrpcPort),
	}
	for _, newVid := range resp.VolumeLocation.NewVids {
		glog.V(2).Infof("%s.%s: %s masterClient adds volume %d", mc.FilerGroup, mc.clientType, loc.Url, newVid)
		mc.addLocation(newVid, loc)
	}
	for _, deletedVid := range resp.VolumeLocation.DeletedVids {
		glog.V(2).Infof("%s.%s: %s masterClient removes volume %d", mc.FilerGroup, mc.clientType, loc.Url, deletedVid)
		mc.deleteLocation(deletedVid, loc)
	}
	for _, newEcVid := range resp.VolumeLocation.NewEcVids {
		glog.V(2).Infof("%s.%s: %s masterClient adds ec volume %d", mc.FilerGroup, mc.clientType, loc.Url, newEcVid)
		mc.addEcLocation(newEcVid, loc)
	}
	for _, deletedEcVid := range resp.VolumeLocation.DeletedEcVids {
		glog.V(2).Infof("%s.%s: %s masterClient removes ec volume %d", mc.FilerGroup, mc.clientType, loc.Url, deletedEcVid)
		mc.deleteEcLocation(deletedEcVid, loc)
	}
	glog.V(1).Infof("updateVidMap(%s) %s.%s: %s volume add: %d, del: %d, add ec: %d del ec: %d",
		resp.VolumeLocation.DataCenter, mc.FilerGroup, mc.clientType, loc.Url,
		len(resp.VolumeLocation.NewVids), len(resp.VolumeLocation.DeletedVids),
		len(resp.VolumeLocation.NewEcVids), len(resp.VolumeLocation.DeletedEcVids))
}

func (mc *MasterClient) WithClient(streamingMode bool, fn func(client master_pb.SeaweedClient) error) error {
	getMasterF := func() pb.ServerAddress { return mc.GetMaster(context.Background()) }
	return mc.WithClientCustomGetMaster(getMasterF, streamingMode, fn)
}

func (mc *MasterClient) WithClientCustomGetMaster(getMasterF func() pb.ServerAddress, streamingMode bool, fn func(client master_pb.SeaweedClient) error) error {
	return util.Retry("master grpc", func() error {
		return pb.WithMasterClient(streamingMode, getMasterF(), mc.grpcDialOption, false, func(client master_pb.SeaweedClient) error {
			return fn(client)
		})
	})
}

func (mc *MasterClient) resetVidMap() {
	tail := &vidMap{
		vid2Locations:   mc.vid2Locations,
		ecVid2Locations: mc.ecVid2Locations,
		DataCenter:      mc.DataCenter,
		cache:           mc.cache,
	}

	nvm := newVidMap(mc.DataCenter)
	nvm.cache = tail
	mc.vidMap = nvm

	//trim
	for i := 0; i < mc.vidMapCacheSize && tail.cache != nil; i++ {
		if i == mc.vidMapCacheSize-1 {
			tail.cache = nil
		} else {
			tail = tail.cache
		}
	}
}