forked from ipfs-cluster/ipfs-cluster
-
Notifications
You must be signed in to change notification settings - Fork 0
/
cluster.go
2068 lines (1795 loc) · 54.4 KB
/
cluster.go
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
package ipfscluster
import (
"context"
"errors"
"fmt"
"mime/multipart"
"sync"
"time"
"github.com/ipfs/ipfs-cluster/adder"
"github.com/ipfs/ipfs-cluster/adder/sharding"
"github.com/ipfs/ipfs-cluster/adder/single"
"github.com/ipfs/ipfs-cluster/api"
"github.com/ipfs/ipfs-cluster/pstoremgr"
"github.com/ipfs/ipfs-cluster/rpcutil"
"github.com/ipfs/ipfs-cluster/state"
"github.com/ipfs/ipfs-cluster/version"
cid "github.com/ipfs/go-cid"
ds "github.com/ipfs/go-datastore"
host "github.com/libp2p/go-libp2p-core/host"
peer "github.com/libp2p/go-libp2p-core/peer"
peerstore "github.com/libp2p/go-libp2p-core/peerstore"
rpc "github.com/libp2p/go-libp2p-gorpc"
dual "github.com/libp2p/go-libp2p-kad-dht/dual"
"github.com/libp2p/go-libp2p/p2p/discovery"
ma "github.com/multiformats/go-multiaddr"
ocgorpc "github.com/lanzafame/go-libp2p-ocgorpc"
trace "go.opencensus.io/trace"
)
// ReadyTimeout specifies the time before giving up
// during startup (waiting for consensus to be ready)
// It may need adjustment according to timeouts in the
// consensus layer.
var ReadyTimeout = 30 * time.Second
const (
pingMetricName = "ping"
bootstrapCount = 3
reBootstrapInterval = 30 * time.Second
mdnsServiceTag = "_ipfs-cluster-discovery._udp"
maxAlerts = 1000
)
var (
errFollowerMode = errors.New("this peer is configured to be in follower mode. Write operations are disabled")
)
// Cluster is the main IPFS cluster component. It provides
// the go-API for it and orchestrates the components that make up the system.
type Cluster struct {
ctx context.Context
cancel func()
id peer.ID
config *Config
host host.Host
dht *dual.DHT
discovery discovery.Service
datastore ds.Datastore
rpcServer *rpc.Server
rpcClient *rpc.Client
peerManager *pstoremgr.Manager
consensus Consensus
apis []API
ipfs IPFSConnector
tracker PinTracker
monitor PeerMonitor
allocator PinAllocator
informers []Informer
tracer Tracer
alerts []api.Alert
alertsMux sync.Mutex
doneCh chan struct{}
readyCh chan struct{}
readyB bool
wg sync.WaitGroup
// peerAdd
paMux sync.Mutex
// shutdown function and related variables
shutdownLock sync.Mutex
shutdownB bool
removed bool
}
// NewCluster builds a new IPFS Cluster peer. It initializes a LibP2P host,
// creates and RPC Server and client and sets up all components.
//
// The new cluster peer may still be performing initialization tasks when
// this call returns (consensus may still be bootstrapping). Use Cluster.Ready()
// if you need to wait until the peer is fully up.
func NewCluster(
ctx context.Context,
host host.Host,
dht *dual.DHT,
cfg *Config,
datastore ds.Datastore,
consensus Consensus,
apis []API,
ipfs IPFSConnector,
tracker PinTracker,
monitor PeerMonitor,
allocator PinAllocator,
informers []Informer,
tracer Tracer,
) (*Cluster, error) {
err := cfg.Validate()
if err != nil {
return nil, err
}
if host == nil {
return nil, errors.New("cluster host is nil")
}
if len(informers) == 0 {
return nil, errors.New("no informers are passed")
}
ctx, cancel := context.WithCancel(ctx)
listenAddrs := ""
for _, addr := range host.Addrs() {
listenAddrs += fmt.Sprintf(" %s/p2p/%s\n", addr, host.ID().Pretty())
}
logger.Infof("IPFS Cluster v%s listening on:\n%s\n", version.Version, listenAddrs)
peerManager := pstoremgr.New(ctx, host, cfg.GetPeerstorePath())
var mdns discovery.Service
if cfg.MDNSInterval > 0 {
mdns, err := discovery.NewMdnsService(ctx, host, cfg.MDNSInterval, mdnsServiceTag)
if err != nil {
logger.Warnf("mDNS could not be started: %s", err)
} else {
mdns.RegisterNotifee(peerManager)
}
}
c := &Cluster{
ctx: ctx,
cancel: cancel,
id: host.ID(),
config: cfg,
host: host,
dht: dht,
discovery: mdns,
datastore: datastore,
consensus: consensus,
apis: apis,
ipfs: ipfs,
tracker: tracker,
monitor: monitor,
allocator: allocator,
informers: informers,
tracer: tracer,
alerts: []api.Alert{},
peerManager: peerManager,
shutdownB: false,
removed: false,
doneCh: make(chan struct{}),
readyCh: make(chan struct{}),
readyB: false,
}
// Import known cluster peers from peerstore file and config. Set
// a non permanent TTL.
c.peerManager.ImportPeersFromPeerstore(false, peerstore.AddressTTL)
c.peerManager.ImportPeers(c.config.PeerAddresses, false, peerstore.AddressTTL)
// Attempt to connect to some peers (up to bootstrapCount)
connectedPeers := c.peerManager.Bootstrap(bootstrapCount)
// We cannot warn when count is low as this as this is normal if going
// to Join() later.
logger.Debugf("bootstrap count %d", len(connectedPeers))
// Log a ping metric for every connected peer. This will make them
// visible as peers without having to wait for them to send one.
for _, p := range connectedPeers {
if err := c.logPingMetric(ctx, p); err != nil {
logger.Warn(err)
}
}
// After setupRPC components can do their tasks with a fully operative
// routed libp2p host with some connections and a working DHT (hopefully).
err = c.setupRPC()
if err != nil {
c.Shutdown(ctx)
return nil, err
}
c.setupRPCClients()
// Note: It is very important to first call Add() once in a non-racy
// place
c.wg.Add(1)
go func() {
defer c.wg.Done()
c.ready(ReadyTimeout)
c.run()
}()
return c, nil
}
func (c *Cluster) setupRPC() error {
rpcServer, err := newRPCServer(c)
if err != nil {
return err
}
c.rpcServer = rpcServer
var rpcClient *rpc.Client
if c.config.Tracing {
csh := &ocgorpc.ClientHandler{}
rpcClient = rpc.NewClientWithServer(
c.host,
version.RPCProtocol,
rpcServer,
rpc.WithClientStatsHandler(csh),
)
} else {
rpcClient = rpc.NewClientWithServer(c.host, version.RPCProtocol, rpcServer)
}
c.rpcClient = rpcClient
return nil
}
func (c *Cluster) setupRPCClients() {
c.tracker.SetClient(c.rpcClient)
c.ipfs.SetClient(c.rpcClient)
for _, api := range c.apis {
api.SetClient(c.rpcClient)
}
c.consensus.SetClient(c.rpcClient)
c.monitor.SetClient(c.rpcClient)
c.allocator.SetClient(c.rpcClient)
for _, informer := range c.informers {
informer.SetClient(c.rpcClient)
}
}
// watchPinset triggers recurrent operations that loop on the pinset.
func (c *Cluster) watchPinset() {
ctx, span := trace.StartSpan(c.ctx, "cluster/watchPinset")
defer span.End()
stateSyncTicker := time.NewTicker(c.config.StateSyncInterval)
recoverTicker := time.NewTicker(c.config.PinRecoverInterval)
for {
select {
case <-stateSyncTicker.C:
logger.Debug("auto-triggering StateSync()")
c.StateSync(ctx)
case <-recoverTicker.C:
logger.Debug("auto-triggering RecoverAllLocal()")
c.RecoverAllLocal(ctx)
case <-c.ctx.Done():
stateSyncTicker.Stop()
recoverTicker.Stop()
return
}
}
}
func (c *Cluster) sendInformerMetric(ctx context.Context, informer Informer) (*api.Metric, error) {
ctx, span := trace.StartSpan(ctx, "cluster/sendInformerMetric")
defer span.End()
metric := informer.GetMetric(ctx)
metric.Peer = c.id
return metric, c.monitor.PublishMetric(ctx, metric)
}
func (c *Cluster) sendInformersMetrics(ctx context.Context) ([]*api.Metric, error) {
ctx, span := trace.StartSpan(ctx, "cluster/sendInformersMetrics")
defer span.End()
var metrics []*api.Metric
for _, informer := range c.informers {
m, err := c.sendInformerMetric(ctx, informer)
if err != nil {
return nil, err
}
metrics = append(metrics, m)
}
return metrics, nil
}
// pushInformerMetrics loops and publishes informers metrics using the
// cluster monitor. Metrics are pushed normally at a TTL/2 rate. If an error
// occurs, they are pushed at a TTL/4 rate.
func (c *Cluster) pushInformerMetrics(ctx context.Context, informer Informer) {
ctx, span := trace.StartSpan(ctx, "cluster/pushInformerMetrics")
defer span.End()
timer := time.NewTimer(0) // fire immediately first
// retries counts how many retries we have made
retries := 0
// retryWarnMod controls how often do we log
// "error broadcasting metric".
// It will do it in the first error, and then on every
// 10th.
retryWarnMod := 10
for {
select {
case <-ctx.Done():
return
case <-timer.C:
// wait
}
metric, err := c.sendInformerMetric(ctx, informer)
if err != nil {
if (retries % retryWarnMod) == 0 {
logger.Errorf("error broadcasting metric: %s", err)
retries++
}
// retry sooner
timer.Reset(metric.GetTTL() / 4)
continue
}
retries = 0
// send metric again in TTL/2
timer.Reset(metric.GetTTL() / 2)
}
}
func (c *Cluster) sendPingMetric(ctx context.Context) (*api.Metric, error) {
ctx, span := trace.StartSpan(ctx, "cluster/sendPingMetric")
defer span.End()
metric := &api.Metric{
Name: pingMetricName,
Peer: c.id,
Valid: true,
}
metric.SetTTL(c.config.MonitorPingInterval * 2)
return metric, c.monitor.PublishMetric(ctx, metric)
}
// logPingMetric logs a ping metric as if it had been sent from PID. It is
// used to make peers appear available as soon as we connect to them (without
// having to wait for them to broadcast a metric).
//
// We avoid specifically sending a metric to a peer when we "connect" to it
// because: a) this requires an extra. OPEN RPC endpoint (LogMetric) that can
// be called by everyone b) We have no way of verifying that the peer ID in a
// metric pushed is actually the issuer of the metric (something the regular
// "pubsub" way of pushing metrics allows (by verifying the signature on the
// message). Thus, this reduces chances of abuse until we have something
// better.
func (c *Cluster) logPingMetric(ctx context.Context, pid peer.ID) error {
m := &api.Metric{
Name: pingMetricName,
Peer: pid,
Valid: true,
}
m.SetTTL(c.config.MonitorPingInterval * 2)
return c.monitor.LogMetric(ctx, m)
}
func (c *Cluster) pushPingMetrics(ctx context.Context) {
ctx, span := trace.StartSpan(ctx, "cluster/pushPingMetrics")
defer span.End()
ticker := time.NewTicker(c.config.MonitorPingInterval)
for {
c.sendPingMetric(ctx)
select {
case <-ctx.Done():
return
case <-ticker.C:
}
}
}
// Alerts returns the last alerts recorded by this cluster peer with the most
// recent first.
func (c *Cluster) Alerts() []api.Alert {
alerts := make([]api.Alert, len(c.alerts))
c.alertsMux.Lock()
{
total := len(alerts)
for i, a := range c.alerts {
alerts[total-1-i] = a
}
}
c.alertsMux.Unlock()
return alerts
}
// read the alerts channel from the monitor and triggers repins
func (c *Cluster) alertsHandler() {
for {
select {
case <-c.ctx.Done():
return
case alrt := <-c.monitor.Alerts():
// Follower peers do not care about alerts.
// They can do nothing about them.
if c.config.FollowerMode {
continue
}
logger.Warnf("metric alert for %s: Peer: %s.", alrt.Name, alrt.Peer)
c.alertsMux.Lock()
{
if len(c.alerts) > maxAlerts {
c.alerts = c.alerts[:0]
}
c.alerts = append(c.alerts, *alrt)
}
c.alertsMux.Unlock()
if alrt.Name != pingMetricName {
continue // only handle ping alerts
}
if c.config.DisableRepinning {
logger.Debugf("repinning is disabled. Will not re-allocate pins on alerts")
return
}
cState, err := c.consensus.State(c.ctx)
if err != nil {
logger.Warn(err)
return
}
list, err := cState.List(c.ctx)
if err != nil {
logger.Warn(err)
return
}
distance, err := c.distances(c.ctx, alrt.Peer)
if err != nil {
logger.Warn(err)
return
}
for _, pin := range list {
if containsPeer(pin.Allocations, alrt.Peer) && distance.isClosest(pin.Cid) {
c.repinFromPeer(c.ctx, alrt.Peer, pin)
}
}
}
}
}
// detects any changes in the peerset and saves the configuration. When it
// detects that we have been removed from the peerset, it shuts down this peer.
func (c *Cluster) watchPeers() {
ticker := time.NewTicker(c.config.PeerWatchInterval)
defer ticker.Stop()
for {
select {
case <-c.ctx.Done():
return
case <-ticker.C:
// logger.Debugf("%s watching peers", c.id)
hasMe := false
peers, err := c.consensus.Peers(c.ctx)
if err != nil {
logger.Error(err)
continue
}
for _, p := range peers {
if p == c.id {
hasMe = true
break
}
}
if !hasMe {
c.shutdownLock.Lock()
defer c.shutdownLock.Unlock()
logger.Info("peer no longer in peerset. Initiating shutdown")
c.removed = true
go c.Shutdown(c.ctx)
return
}
}
}
}
// reBootstrap regularly attempts to bootstrap (re-connect to peers from the
// peerstore). This should ensure that we auto-recover from situations in
// which the network was completely gone and we lost all peers.
func (c *Cluster) reBootstrap() {
ticker := time.NewTicker(reBootstrapInterval)
defer ticker.Stop()
for {
select {
case <-c.ctx.Done():
return
case <-ticker.C:
connected := c.peerManager.Bootstrap(bootstrapCount)
for _, p := range connected {
logger.Infof("reconnected to %s", p)
}
}
}
}
// find all Cids pinned to a given peer and triggers re-pins on them.
func (c *Cluster) vacatePeer(ctx context.Context, p peer.ID) {
ctx, span := trace.StartSpan(ctx, "cluster/vacatePeer")
defer span.End()
if c.config.DisableRepinning {
logger.Warnf("repinning is disabled. Will not re-allocate cids from %s", p.Pretty())
return
}
cState, err := c.consensus.State(ctx)
if err != nil {
logger.Warn(err)
return
}
list, err := cState.List(ctx)
if err != nil {
logger.Warn(err)
return
}
for _, pin := range list {
if containsPeer(pin.Allocations, p) {
c.repinFromPeer(ctx, p, pin)
}
}
}
// repinFromPeer triggers a repin on a given pin object blacklisting one of the
// allocations.
func (c *Cluster) repinFromPeer(ctx context.Context, p peer.ID, pin *api.Pin) {
ctx, span := trace.StartSpan(ctx, "cluster/repinFromPeer")
defer span.End()
pin.Allocations = nil // force re-allocations
_, ok, err := c.pin(ctx, pin, []peer.ID{p})
if ok && err == nil {
logger.Infof("repinned %s out of %s", pin.Cid, p.Pretty())
}
}
// run launches some go-routines which live throughout the cluster's life
func (c *Cluster) run() {
c.wg.Add(1)
go func() {
defer c.wg.Done()
c.watchPinset()
}()
c.wg.Add(1)
go func() {
defer c.wg.Done()
c.pushPingMetrics(c.ctx)
}()
c.wg.Add(len(c.informers))
for _, informer := range c.informers {
go func(inf Informer) {
defer c.wg.Done()
c.pushInformerMetrics(c.ctx, inf)
}(informer)
}
c.wg.Add(1)
go func() {
defer c.wg.Done()
c.watchPeers()
}()
c.wg.Add(1)
go func() {
defer c.wg.Done()
c.alertsHandler()
}()
c.wg.Add(1)
go func() {
defer c.wg.Done()
c.reBootstrap()
}()
}
func (c *Cluster) ready(timeout time.Duration) {
ctx, span := trace.StartSpan(c.ctx, "cluster/ready")
defer span.End()
// We bootstrapped first because with dirty state consensus
// may have a peerset and not find a leader so we cannot wait
// for it.
timer := time.NewTimer(timeout)
select {
case <-timer.C:
logger.Error("***** ipfs-cluster consensus start timed out (tips below) *****")
logger.Error(`
**************************************************
This peer was not able to become part of the cluster.
This might be due to one or several causes:
- Check the logs above this message for errors
- Check that there is connectivity to the "peers" multiaddresses
- Check that all cluster peers are using the same "secret"
- Check that this peer is reachable on its "listen_multiaddress" by all peers
- Check that the current cluster is healthy (has a leader). Otherwise make
sure to start enough peers so that a leader election can happen.
- Check that the peer(s) you are trying to connect to is running the
same version of IPFS-cluster.
**************************************************
`)
c.Shutdown(ctx)
return
case <-c.consensus.Ready(ctx):
// Consensus ready means the state is up to date. Every item
// in the state that is not pinned will appear as PinError so
// we can proceed to recover all of those in the tracker.
c.RecoverAllLocal(ctx)
case <-c.ctx.Done():
return
}
// Cluster is ready.
peers, err := c.consensus.Peers(ctx)
if err != nil {
logger.Error(err)
c.Shutdown(ctx)
return
}
logger.Info("Cluster Peers (without including ourselves):")
if len(peers) == 1 {
logger.Info(" - No other peers")
}
for _, p := range peers {
if p != c.id {
logger.Infof(" - %s", p.Pretty())
}
}
close(c.readyCh)
c.shutdownLock.Lock()
c.readyB = true
c.shutdownLock.Unlock()
logger.Info("** IPFS Cluster is READY **")
}
// Ready returns a channel which signals when this peer is
// fully initialized (including consensus).
func (c *Cluster) Ready() <-chan struct{} {
return c.readyCh
}
// Shutdown performs all the necessary operations to shutdown
// the IPFS Cluster peer:
// * Save peerstore with the current peers
// * Remove itself from consensus when LeaveOnShutdown is set
// * It Shutdowns all the components
// * Collects all goroutines
//
// Shutdown does not close the libp2p host, the DHT, the datastore or
// generally anything that Cluster did not create.
func (c *Cluster) Shutdown(ctx context.Context) error {
_, span := trace.StartSpan(ctx, "cluster/Shutdown")
defer span.End()
ctx = trace.NewContext(c.ctx, span)
c.shutdownLock.Lock()
defer c.shutdownLock.Unlock()
if c.shutdownB {
logger.Debug("Cluster is already shutdown")
return nil
}
logger.Info("shutting down Cluster")
// Cancel discovery service (this shutdowns announcing). Handling
// entries is cancelled along with the context below.
if c.discovery != nil {
c.discovery.Close()
}
// Try to store peerset file for all known peers whatsoever
// if we got ready (otherwise, don't overwrite anything)
if c.readyB {
// Ignoring error since it's a best-effort
c.peerManager.SavePeerstoreForPeers(c.host.Peerstore().Peers())
}
// Only attempt to leave if:
// - consensus is initialized
// - cluster was ready (no bootstrapping error)
// - We are not removed already (means watchPeers() called us)
if c.consensus != nil && c.config.LeaveOnShutdown && c.readyB && !c.removed {
c.removed = true
_, err := c.consensus.Peers(ctx)
if err == nil {
// best effort
logger.Warn("attempting to leave the cluster. This may take some seconds")
err := c.consensus.RmPeer(ctx, c.id)
if err != nil {
logger.Error("leaving cluster: " + err.Error())
}
}
}
if con := c.consensus; con != nil {
if err := con.Shutdown(ctx); err != nil {
logger.Errorf("error stopping consensus: %s", err)
return err
}
}
// We left the cluster or were removed. Remove any consensus-specific
// state.
if c.removed && c.readyB {
err := c.consensus.Clean(ctx)
if err != nil {
logger.Error("cleaning consensus: ", err)
}
}
if err := c.monitor.Shutdown(ctx); err != nil {
logger.Errorf("error stopping monitor: %s", err)
return err
}
for _, api := range c.apis {
if err := api.Shutdown(ctx); err != nil {
logger.Errorf("error stopping API: %s", err)
return err
}
}
if err := c.ipfs.Shutdown(ctx); err != nil {
logger.Errorf("error stopping IPFS Connector: %s", err)
return err
}
if err := c.tracker.Shutdown(ctx); err != nil {
logger.Errorf("error stopping PinTracker: %s", err)
return err
}
for _, inf := range c.informers {
if err := inf.Shutdown(ctx); err != nil {
logger.Errorf("error stopping informer: %s", err)
return err
}
}
if err := c.tracer.Shutdown(ctx); err != nil {
logger.Errorf("error stopping Tracer: %s", err)
return err
}
c.cancel()
c.wg.Wait()
c.shutdownB = true
close(c.doneCh)
return nil
}
// Done provides a way to learn if the Peer has been shutdown
// (for example, because it has been removed from the Cluster)
func (c *Cluster) Done() <-chan struct{} {
return c.doneCh
}
// ID returns information about the Cluster peer
func (c *Cluster) ID(ctx context.Context) *api.ID {
_, span := trace.StartSpan(ctx, "cluster/ID")
defer span.End()
ctx = trace.NewContext(c.ctx, span)
// ignore error since it is included in response object
ipfsID, err := c.ipfs.ID(ctx)
if err != nil {
ipfsID = &api.IPFSID{
Error: err.Error(),
}
}
var addrs []api.Multiaddr
mAddrs, err := peer.AddrInfoToP2pAddrs(&peer.AddrInfo{ID: c.id, Addrs: c.host.Addrs()})
if err == nil {
for _, mAddr := range mAddrs {
addrs = append(addrs, api.NewMultiaddrWithValue(mAddr))
}
}
peers := []peer.ID{}
// This method might get called very early by a remote peer
// and might catch us when consensus is not set
if c.consensus != nil {
peers, _ = c.consensus.Peers(ctx)
}
clusterPeerInfos := c.peerManager.PeerInfos(peers)
addresses := []api.Multiaddr{}
for _, pinfo := range clusterPeerInfos {
addrs, err := peer.AddrInfoToP2pAddrs(&pinfo)
if err != nil {
continue
}
for _, a := range addrs {
addresses = append(addresses, api.NewMultiaddrWithValue(a))
}
}
id := &api.ID{
ID: c.id,
//PublicKey: c.host.Peerstore().PubKey(c.id),
Addresses: addrs,
ClusterPeers: peers,
ClusterPeersAddresses: addresses,
Version: version.Version.String(),
RPCProtocolVersion: version.RPCProtocol,
IPFS: ipfsID,
Peername: c.config.Peername,
}
if err != nil {
id.Error = err.Error()
}
return id
}
// PeerAdd adds a new peer to this Cluster.
//
// For it to work well, the new peer should be discoverable
// (part of our peerstore or connected to one of the existing peers)
// and reachable. Since PeerAdd allows to add peers which are
// not running, or reachable, it is recommended to call Join() from the
// new peer instead.
//
// The new peer ID will be passed to the consensus
// component to be added to the peerset.
func (c *Cluster) PeerAdd(ctx context.Context, pid peer.ID) (*api.ID, error) {
_, span := trace.StartSpan(ctx, "cluster/PeerAdd")
defer span.End()
ctx = trace.NewContext(c.ctx, span)
// starting 10 nodes on the same box for testing
// causes deadlock and a global lock here
// seems to help.
c.paMux.Lock()
defer c.paMux.Unlock()
logger.Debugf("peerAdd called with %s", pid.Pretty())
// Let the consensus layer be aware of this peer
err := c.consensus.AddPeer(ctx, pid)
if err != nil {
logger.Error(err)
id := &api.ID{ID: pid, Error: err.Error()}
return id, err
}
logger.Info("Peer added ", pid.Pretty())
addedID, err := c.getIDForPeer(ctx, pid)
if err != nil {
return addedID, err
}
if !containsPeer(addedID.ClusterPeers, c.id) {
addedID.ClusterPeers = append(addedID.ClusterPeers, c.id)
}
return addedID, nil
}
// PeerRemove removes a peer from this Cluster.
//
// The peer will be removed from the consensus peerset.
// This may first trigger repinnings for all content if not disabled.
func (c *Cluster) PeerRemove(ctx context.Context, pid peer.ID) error {
_, span := trace.StartSpan(ctx, "cluster/PeerRemove")
defer span.End()
ctx = trace.NewContext(c.ctx, span)
// We need to repin before removing the peer, otherwise, it won't
// be able to submit the pins.
logger.Infof("re-allocating all CIDs directly associated to %s", pid)
c.vacatePeer(ctx, pid)
err := c.consensus.RmPeer(ctx, pid)
if err != nil {
logger.Error(err)
return err
}
logger.Info("Peer removed ", pid.Pretty())
return nil
}
// Join adds this peer to an existing cluster by bootstrapping to a
// given multiaddress. It works by calling PeerAdd on the destination
// cluster and making sure that the new peer is ready to discover and contact
// the rest.
func (c *Cluster) Join(ctx context.Context, addr ma.Multiaddr) error {
_, span := trace.StartSpan(ctx, "cluster/Join")
defer span.End()
ctx = trace.NewContext(c.ctx, span)
logger.Debugf("Join(%s)", addr)
// Add peer to peerstore so we can talk to it
pid, err := c.peerManager.ImportPeer(addr, false, peerstore.PermanentAddrTTL)
if err != nil {
return err
}
if pid == c.id {
return nil
}
// Note that PeerAdd() on the remote peer will
// figure out what our real address is (obviously not
// ListenAddr).
var myID api.ID
err = c.rpcClient.CallContext(
ctx,
pid,
"Cluster",
"PeerAdd",
c.id,
&myID,
)
if err != nil {
logger.Error(err)
return err
}
// Log a fake but valid metric from the peer we are
// contacting. This will signal a CRDT component that
// we know that peer since we have metrics for it without
// having to wait for the next metric round.
if err := c.logPingMetric(ctx, pid); err != nil {
logger.Warn(err)
}
// Broadcast our metrics to the world
_, err = c.sendInformersMetrics(ctx)
if err != nil {
logger.Warn(err)
}
_, err = c.sendPingMetric(ctx)
if err != nil {
logger.Warn(err)
}
// We need to trigger a DHT bootstrap asap for this peer to not be
// lost if the peer it bootstrapped to goes down. We do this manually
// by triggering 1 round of bootstrap in the background.
// Note that our regular bootstrap process is still running in the
// background since we created the cluster.
c.wg.Add(1)
go func() {
defer c.wg.Done()
select {
case err := <-c.dht.LAN.RefreshRoutingTable():
if err != nil {
// this error is quite chatty
// on single peer clusters
logger.Debug(err)
}
case <-c.ctx.Done():
return
}
select {
case err := <-c.dht.WAN.RefreshRoutingTable():
if err != nil {
// this error is quite chatty
// on single peer clusters
logger.Debug(err)
}
case <-c.ctx.Done():
return
}