-
Notifications
You must be signed in to change notification settings - Fork 36
/
read_only.go
212 lines (185 loc) · 5.01 KB
/
read_only.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
// Copyright 2015 The etcd Authors
// Modified work copyright 2018 The tiglabs Authors.
//
// Licensed under the Apache License, Version 2.0 (the "License");
// you may not use this file except in compliance with the License.
// You may obtain a copy of the License at
//
// http://www.apache.org/licenses/LICENSE-2.0
//
// Unless required by applicable law or agreed to in writing, software
// distributed under the License is distributed on an "AS IS" BASIS,
// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
// See the License for the specific language governing permissions and
// limitations under the License.
package raft
import (
"fmt"
"github.com/tiglabs/raft/logger"
)
// ReadOnlyOption read only option
type ReadOnlyOption int
const (
// ReadOnlySafe guarantees the linearizability of the read only request by
// communicating with the quorum. It is the default and suggested option.
ReadOnlySafe ReadOnlyOption = iota
// ReadOnlyLeaseBased ensures linearizability of the read only request by
// relying on the leader lease. It can be affected by clock drift.
// If the clock drift is unbounded, leader might keep the lease longer than it
// should (clock can move backward/pause without any bound). ReadIndex is not safe
// in that case.
ReadOnlyLeaseBased
)
type readIndexStatus struct {
index uint64
futures []*Future
acks map[uint64]struct{}
}
type readIndexReady struct {
index uint64
futures []*Future
}
type readOnly struct {
id uint64 // raft id
option ReadOnlyOption
// wait leader to commit an entry in current term
committed bool
// ReadIndex requests before leader commit entry in current term
scratch []*Future
// wait quorum ack
pendings map[uint64]*readIndexStatus
pendingQueue []uint64
// quorum acked, wait apply
readys map[uint64]*readIndexReady
readyQueue []uint64
}
func newReadOnly(id uint64, option ReadOnlyOption) *readOnly {
return &readOnly{
id: id,
option: option,
pendings: make(map[uint64]*readIndexStatus),
readys: make(map[uint64]*readIndexReady),
}
}
func (r *readOnly) addPending(index uint64, futures []*Future) {
if status, ok := r.pendings[index]; ok {
status.futures = append(status.futures, futures...)
return
}
// check index valid
if index <= r.lastPending() {
panic(AppPanicError(fmt.Sprintf("[raft->addReadOnly][%v] invalid index[%d]: less than last[%d]", r.id, index, r.lastPending())))
}
r.pendingQueue = append(r.pendingQueue, index)
r.pendings[index] = &readIndexStatus{
index: index,
futures: futures,
acks: make(map[uint64]struct{}),
}
}
func (r *readOnly) addReady(index uint64, futures []*Future) {
if status, ok := r.readys[index]; ok {
status.futures = append(status.futures, futures...)
return
}
r.readyQueue = append(r.readyQueue, index)
r.readys[index] = &readIndexReady{
index: index,
futures: futures,
}
}
func (r *readOnly) add(index uint64, futures []*Future) {
if !r.committed {
r.scratch = append(r.scratch, futures...)
return
}
if r.option == ReadOnlyLeaseBased {
r.addReady(index, futures)
} else {
r.addPending(index, futures)
}
}
func (r *readOnly) commit(index uint64) {
if !r.committed {
r.committed = true
if len(r.scratch) > 0 {
r.add(index, r.scratch)
r.scratch = nil
}
}
}
func (r *readOnly) lastPending() uint64 {
if len(r.pendingQueue) > 0 {
return r.pendingQueue[len(r.pendingQueue)-1]
}
return 0
}
func (r *readOnly) recvAck(index uint64, from uint64, quorum int) {
status, ok := r.pendings[index]
if !ok {
return
}
status.acks[from] = struct{}{}
// add one to include an ack from local node
if len(status.acks)+1 >= quorum {
r.advance(index)
}
}
func (r *readOnly) advance(index uint64) {
var i int
for _, idx := range r.pendingQueue {
if idx > index {
break
}
if rs, ok := r.pendings[idx]; ok {
r.addReady(idx, rs.futures)
delete(r.pendings, idx)
}
i++
}
r.pendingQueue = r.pendingQueue[i:]
}
func (r *readOnly) getReady(applied uint64) (futures []*Future) {
if len(r.readyQueue) == 0 {
return nil
}
var i int
for _, idx := range r.readyQueue {
if idx > applied {
break
}
if rs, ok := r.readys[idx]; ok {
futures = append(futures, rs.futures...)
delete(r.readys, idx)
}
i++
}
r.readyQueue = r.readyQueue[i:]
// TODO: remove this when stable
if logger.IsEnableDebug() {
logger.Debug("raft[%d] get ready index %d, futures len: %d", r.id, applied, len(futures))
}
return
}
func (r *readOnly) containsUpdate(applied uint64) bool {
return len(r.readyQueue) > 0 && applied >= r.readyQueue[0]
}
func (r *readOnly) reset(err error) {
respondReadIndex(r.scratch, err)
for _, status := range r.pendings {
respondReadIndex(status.futures, err)
}
for _, ready := range r.readys {
respondReadIndex(ready.futures, err)
}
r.committed = false
r.scratch = nil
r.pendings = make(map[uint64]*readIndexStatus)
r.pendingQueue = nil
r.readys = make(map[uint64]*readIndexReady)
}
func respondReadIndex(future []*Future, err error) {
for _, f := range future {
f.respond(nil, err)
}
}