forked from influxdata/influxdb
-
Notifications
You must be signed in to change notification settings - Fork 0
/
service.go
125 lines (104 loc) · 3.39 KB
/
service.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
// Package retention provides the retention policy enforcement service.
package retention // import "github.com/influxdata/influxdb/services/retention"
import (
"fmt"
"sync"
"time"
"github.com/influxdata/influxdb/services/meta"
"go.uber.org/zap"
)
// Service represents the retention policy enforcement service.
type Service struct {
MetaClient interface {
Databases() []meta.DatabaseInfo
DeleteShardGroup(database, policy string, id uint64) error
PruneShardGroups() error
}
TSDBStore interface {
ShardIDs() []uint64
DeleteShard(shardID uint64) error
}
config Config
wg sync.WaitGroup
done chan struct{}
logger *zap.Logger
}
// NewService returns a configured retention policy enforcement service.
func NewService(c Config) *Service {
return &Service{
config: c,
logger: zap.NewNop(),
}
}
// Open starts retention policy enforcement.
func (s *Service) Open() error {
if !s.config.Enabled || s.done != nil {
return nil
}
s.logger.Info("Starting retention policy enforcement service", zap.String("check-interval", s.config.CheckInterval.String()))
s.done = make(chan struct{})
s.wg.Add(1)
go func() { defer s.wg.Done(); s.run() }()
return nil
}
// Close stops retention policy enforcement.
func (s *Service) Close() error {
if !s.config.Enabled || s.done == nil {
return nil
}
s.logger.Info("Retention policy enforcement service closing.")
close(s.done)
s.wg.Wait()
s.done = nil
return nil
}
// WithLogger sets the logger on the service.
func (s *Service) WithLogger(log *zap.Logger) {
s.logger = log.With(zap.String("service", "retention"))
}
func (s *Service) run() {
ticker := time.NewTicker(time.Duration(s.config.CheckInterval))
defer ticker.Stop()
for {
select {
case <-s.done:
return
case <-ticker.C:
s.logger.Info("Retention policy shard deletion check commencing.")
type deletionInfo struct {
db string
rp string
}
deletedShardIDs := make(map[uint64]deletionInfo)
dbs := s.MetaClient.Databases()
for _, d := range dbs {
for _, r := range d.RetentionPolicies {
for _, g := range r.ExpiredShardGroups(time.Now().UTC()) {
if err := s.MetaClient.DeleteShardGroup(d.Name, r.Name, g.ID); err != nil {
s.logger.Info(fmt.Sprintf("Failed to delete shard group %d from database %s, retention policy %s: %v. Retry in %v.", g.ID, d.Name, r.Name, err, s.config.CheckInterval))
continue
}
s.logger.Info(fmt.Sprintf("Deleted shard group %d from database %s, retention policy %s.", g.ID, d.Name, r.Name))
// Store all the shard IDs that may possibly need to be removed locally.
for _, sh := range g.Shards {
deletedShardIDs[sh.ID] = deletionInfo{db: d.Name, rp: r.Name}
}
}
}
}
// Remove shards if we store them locally
for _, id := range s.TSDBStore.ShardIDs() {
if info, ok := deletedShardIDs[id]; ok {
if err := s.TSDBStore.DeleteShard(id); err != nil {
s.logger.Error(fmt.Sprintf("Failed to delete shard ID %d from database %s, retention policy %s: %v. Will retry in %v", id, info.db, info.rp, err, s.config.CheckInterval))
continue
}
s.logger.Info(fmt.Sprintf("Shard ID %d from database %s, retention policy %s, deleted.", id, info.db, info.rp))
}
}
if err := s.MetaClient.PruneShardGroups(); err != nil {
s.logger.Info(fmt.Sprintf("Problem pruning shard groups: %s. Will retry in %v", err, s.config.CheckInterval))
}
}
}
}