File size: 5,613 Bytes
95d599c | 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 | // _ _
// __ _____ __ ___ ___ __ _| |_ ___
// \ \ /\ / / _ \/ _` \ \ / / |/ _` | __/ _ \
// \ V V / __/ (_| |\ V /| | (_| | || __/
// \_/\_/ \___|\__,_| \_/ |_|\__,_|\__\___|
//
// Copyright © 2016 - 2025 Weaviate B.V. All rights reserved.
//
// CONTACT: hello@weaviate.io
//
package bootstrap
import (
"context"
"fmt"
"math/rand"
"time"
"github.com/getsentry/sentry-go"
"github.com/sirupsen/logrus"
cmd "github.com/weaviate/weaviate/cluster/proto/api"
"github.com/weaviate/weaviate/cluster/resolver"
entSentry "github.com/weaviate/weaviate/entities/sentry"
)
// PeerJoiner is the interface we expect to be able to talk to the other peers to either Join or Notify them
type PeerJoiner interface {
Join(_ context.Context, leaderAddr string, _ *cmd.JoinPeerRequest) (*cmd.JoinPeerResponse, error)
Notify(_ context.Context, leaderAddr string, _ *cmd.NotifyPeerRequest) (*cmd.NotifyPeerResponse, error)
}
// Bootstrapper is used to bootstrap this node by attempting to join it to a RAFT cluster.
type Bootstrapper struct {
peerJoiner PeerJoiner
addrResolver resolver.ClusterStateReader
isStoreReady func() bool
localRaftAddr string
localNodeID string
voter bool
retryPeriod time.Duration
jitter time.Duration
}
// NewBootstrapper constructs a new bootsrapper
func NewBootstrapper(peerJoiner PeerJoiner, raftID string, raftAddr string, voter bool, r resolver.ClusterStateReader, isStoreReady func() bool) *Bootstrapper {
return &Bootstrapper{
peerJoiner: peerJoiner,
addrResolver: r,
retryPeriod: time.Second,
jitter: time.Second,
localNodeID: raftID,
localRaftAddr: raftAddr,
isStoreReady: isStoreReady,
voter: voter,
}
}
// Do iterates over a list of servers in an attempt to join this node to a cluster.
func (b *Bootstrapper) Do(ctx context.Context, serverPortMap map[string]int, lg *logrus.Logger, stop chan struct{}) error {
if entSentry.Enabled() {
transaction := sentry.StartTransaction(ctx, "raft.bootstrap",
sentry.WithOpName("init"),
sentry.WithDescription("Attempt to bootstrap a raft cluster"),
)
ctx = transaction.Context()
defer transaction.Finish()
}
ticker := time.NewTicker(jitter(b.retryPeriod, b.jitter))
defer ticker.Stop()
for {
select {
case <-stop:
return nil
case <-ctx.Done():
return ctx.Err()
case <-ticker.C:
if b.isStoreReady() {
lg.WithField("action", "bootstrap").Info("node reporting ready, exiting bootstrap process")
return nil
}
remoteNodes := ResolveRemoteNodes(b.addrResolver, serverPortMap)
// We were not able to resolve any nodes to an address
if len(remoteNodes) == 0 {
lg.WithField("action", "bootstrap").WithField("join_list", serverPortMap).Warn("unable to resolve any node address to join")
continue
}
// Always try to join an existing cluster first
joiner := NewJoiner(b.peerJoiner, b.localNodeID, b.localRaftAddr, b.voter)
if leader, err := joiner.Do(ctx, lg, remoteNodes); err != nil {
lg.WithFields(logrus.Fields{
"action": "bootstrap",
"servers": remoteNodes,
"voter": b.voter,
}).WithError(err).Warning("failed to join cluster")
} else {
lg.WithFields(logrus.Fields{
"action": "bootstrap",
"leader": leader,
}).Info("successfully joined cluster")
return nil
}
// We are a voter, we resolve other peers but we're unable to join them. We're in the situation where we are
// bootstrapping a new cluster and now we want to notify the other nodes.
// Each node on notify will build a list of notified node. Once bootstrap expect is reached the nodes will
// bootstrap together.
if b.voter {
// notify other servers about readiness of this node to be joined
if err := b.notify(ctx, remoteNodes); err != nil {
lg.WithFields(logrus.Fields{
"action": "bootstrap",
"servers": remoteNodes,
}).WithError(err).Error("failed to notify peers")
continue
}
lg.WithFields(logrus.Fields{
"action": "bootstrap",
"servers": remoteNodes,
}).Info("notified peers this node is ready to join as voter")
}
}
}
}
// notify attempts to notify all nodes in remoteNodes that this server is ready to bootstrap
func (b *Bootstrapper) notify(ctx context.Context, remoteNodes map[string]string) (err error) {
if entSentry.Enabled() {
span := sentry.StartSpan(ctx, "raft.bootstrap.notify",
sentry.WithOpName("notify"),
sentry.WithDescription("Attempt to notify existing node(s) to join a cluster"),
)
ctx = span.Context()
span.SetData("servers", remoteNodes)
defer span.Finish()
}
for _, addr := range remoteNodes {
req := &cmd.NotifyPeerRequest{Id: b.localNodeID, Address: b.localRaftAddr}
_, err = b.peerJoiner.Notify(ctx, addr, req)
if err != nil {
return err
}
}
return
}
// ResolveRemoteNodes returns a list of remoteNodes addresses resolved using addrResolver. The nodes id used are
// taken from serverPortMap keys and ports from the values
func ResolveRemoteNodes(addrResolver resolver.ClusterStateReader, serverPortMap map[string]int) map[string]string {
candidates := make(map[string]string, len(serverPortMap))
for name, raftPort := range serverPortMap {
if addr := addrResolver.NodeAddress(name); addr != "" {
candidates[name] = fmt.Sprintf("%s:%d", addr, raftPort)
}
}
return candidates
}
// jitter introduce some jitter to a given duration d + [0, 1) * jit -> [d, d+jit]
func jitter(d time.Duration, jit time.Duration) time.Duration {
return d + time.Duration(float64(jit)*rand.Float64())
}
|