-
Notifications
You must be signed in to change notification settings - Fork 4
/
Copy pathbackup.go
206 lines (169 loc) · 6.21 KB
/
backup.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
package async_geyser
import (
"context"
"sync"
"time"
"github.com/newrelic/go-agent/v3/newrelic"
"github.com/code-payments/code-server/pkg/code/common"
"github.com/code-payments/code-server/pkg/code/data/account"
"github.com/code-payments/code-server/pkg/metrics"
timelock_token_v1 "github.com/code-payments/code-server/pkg/solana/timelock/v1"
)
// Backup system workers can be found here. This is necessary because we can't rely
// on receiving all updates from Geyser. As a result, we should design backup systems
// to assume Geyser doesn't function/exist at all. Why do we need Geyser if this is
// the case? Real time updates. Backup workers likely won't be able to guarantee
// real time (or near real time) updates at scale.
func (p *service) backupTimelockStateWorker(serviceCtx context.Context, interval time.Duration) error {
log := p.log.WithField("method", "backupTimelockStateWorker")
log.Debug("worker started")
p.metricStatusLock.Lock()
p.backupTimelockStateWorkerStatus = true
p.metricStatusLock.Unlock()
defer func() {
p.metricStatusLock.Lock()
p.backupTimelockStateWorkerStatus = false
p.metricStatusLock.Unlock()
log.Debug("worker stopped")
}()
delay := 0 * time.Second // Initially no delay, so we can run right after a deploy
for {
select {
case <-time.After(delay):
start := time.Now()
func() {
nr := serviceCtx.Value(metrics.NewRelicContextKey).(*newrelic.Application)
m := nr.StartTransaction("async__geyser_consumer_service__backup_timelock_state_worker")
defer m.End()
tracedCtx := newrelic.NewContext(serviceCtx, m)
jobSucceeded := true
// Find and process unlocked timelock accounts unlocking between [+21 - n days, +21 days],
// which enables a retry mechanism across time.
for i := uint8(0); i <= uint8(p.conf.backupTimelockWorkerDaysChecked.Get(tracedCtx)); i++ {
daysUntilUnlock := timelock_token_v1.DefaultNumDaysLocked - i
addresses, slot, err := findUnlockedTimelockV1Accounts(tracedCtx, p.data, daysUntilUnlock)
if err != nil {
m.NoticeError(err)
log.WithError(err).Warn("failure getting unlocked timelock accounts")
jobSucceeded = false
continue
}
log.Infof("found %d timelock accounts unlocking in %d days", len(addresses), daysUntilUnlock)
for _, address := range addresses {
log := log.WithField("account", address)
stateAccount, err := common.NewAccountFromPublicKeyString(address)
if err != nil {
log.WithError(err).Warn("invalid state account address")
continue
}
err = updateTimelockV1AccountCachedState(tracedCtx, p.data, stateAccount, slot)
if err != nil {
m.NoticeError(err)
log.WithError(err).Warn("failure updating cached timelock account state")
jobSucceeded = false
continue
}
}
}
p.metricStatusLock.Lock()
p.unlockedTimelockAccountsSynced = jobSucceeded
p.metricStatusLock.Unlock()
}()
delay = interval - time.Since(start)
case <-serviceCtx.Done():
return serviceCtx.Err()
}
}
}
func (p *service) backupExternalDepositWorker(serviceCtx context.Context, interval time.Duration) error {
log := p.log.WithField("method", "backupExternalDepositWorker")
log.Debug("worker started")
p.metricStatusLock.Lock()
p.backupExternalDepositWorkerStatus = true
p.metricStatusLock.Unlock()
defer func() {
p.metricStatusLock.Lock()
p.backupExternalDepositWorkerStatus = false
p.metricStatusLock.Unlock()
log.Debug("worker stopped")
}()
for {
select {
case <-time.After(interval):
func() {
nr := serviceCtx.Value(metrics.NewRelicContextKey).(*newrelic.Application)
m := nr.StartTransaction("async__geyser_consumer_service__backup_external_deposit_worker")
defer m.End()
tracedCtx := newrelic.NewContext(serviceCtx, m)
accountInfoRecords, err := p.data.GetPrioritizedAccountInfosRequiringDepositSync(tracedCtx, p.conf.backupExternalDepositWorkerCount.Get(tracedCtx))
if err != nil {
if err != account.ErrAccountInfoNotFound {
m.NoticeError(err)
log.WithError(err).Warn("failure getting accounts to sync external deposits")
}
return
}
var wg sync.WaitGroup
for _, accountInfoRecord := range accountInfoRecords {
vault, err := common.NewAccountFromPublicKeyString(accountInfoRecord.TokenAccount)
if err != nil {
log.WithError(err).WithField("account", accountInfoRecord.TokenAccount).Warn("invalid token account")
continue
}
wg.Add(1)
go func(vault *common.Account) {
defer wg.Done()
log := log.WithField("account", vault.PublicKey().ToBase58())
err := fixMissingExternalDeposits(tracedCtx, p.conf, p.data, p.pusher, vault)
if err != nil {
m.NoticeError(err)
log.WithError(err).Warn("failure fixing missing external deposits")
}
}(vault)
}
wg.Wait()
}()
case <-serviceCtx.Done():
return serviceCtx.Err()
}
}
}
func (p *service) backupMessagingWorker(serviceCtx context.Context, interval time.Duration) error {
log := p.log.WithField("method", "backupMessagingWorker")
log.Debug("worker started")
p.metricStatusLock.Lock()
p.backupMessagingWorkerStatus = true
p.metricStatusLock.Unlock()
defer func() {
p.metricStatusLock.Lock()
p.backupMessagingWorkerStatus = false
p.metricStatusLock.Unlock()
log.Debug("worker stopped")
}()
delay := 0 * time.Second // Initially no delay, so we can run right after a deploy
messagingFeeCollector, err := common.NewAccountFromPublicKeyString(p.conf.messagingFeeCollectorPublicKey.Get(serviceCtx))
if err != nil {
return err
}
var checkpoint *string
for {
select {
case <-time.After(delay):
start := time.Now()
func() {
nr := serviceCtx.Value(metrics.NewRelicContextKey).(*newrelic.Application)
m := nr.StartTransaction("async__geyser_consumer_service__backup_messaging_worker")
defer m.End()
tracedCtx := newrelic.NewContext(serviceCtx, m)
checkpoint, err = fixMissingBlockchainMessages(tracedCtx, p.data, p.pusher, messagingFeeCollector, checkpoint)
if err != nil {
m.NoticeError(err)
log.WithError(err).Warn("failure fixing missing messages")
}
}()
delay = interval - time.Since(start)
case <-serviceCtx.Done():
return serviceCtx.Err()
}
}
}