/usr/share/gocode/src/github.com/influxdata/influxdb/cmd/influx_tsm/tracker.go is in golang-github-influxdb-influxdb-dev 1.1.1+dfsg1-4.
This file is owned by root:root, with mode 0o644.
The actual contents of the file can be viewed below.
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 | package main
import (
"fmt"
"log"
"runtime"
"sync"
"sync/atomic"
"time"
"github.com/influxdata/influxdb/cmd/influx_tsm/stats"
"github.com/influxdata/influxdb/cmd/influx_tsm/tsdb"
)
// tracker will orchestrate and track the conversions of non-TSM shards to TSM
type tracker struct {
Stats stats.Stats
shards tsdb.ShardInfos
opts options
pg ParallelGroup
wg sync.WaitGroup
}
// newTracker will setup and return a clean tracker instance
func newTracker(shards tsdb.ShardInfos, opts options) *tracker {
t := &tracker{
shards: shards,
opts: opts,
pg: NewParallelGroup(runtime.GOMAXPROCS(0)),
}
return t
}
func (t *tracker) Run() error {
conversionStart := time.Now()
// Backup each directory.
if !opts.SkipBackup {
databases := t.shards.Databases()
fmt.Printf("Backing up %d databases...\n", len(databases))
t.wg.Add(len(databases))
for i := range databases {
db := databases[i]
go t.pg.Do(func() {
defer t.wg.Done()
start := time.Now()
log.Printf("Backup of database '%v' started", db)
err := backupDatabase(db)
if err != nil {
log.Fatalf("Backup of database %v failed: %v\n", db, err)
}
log.Printf("Database %v backed up (%v)\n", db, time.Now().Sub(start))
})
}
t.wg.Wait()
} else {
fmt.Println("Database backup disabled.")
}
t.wg.Add(len(t.shards))
for i := range t.shards {
si := t.shards[i]
go t.pg.Do(func() {
defer func() {
atomic.AddUint64(&t.Stats.CompletedShards, 1)
t.wg.Done()
}()
start := time.Now()
log.Printf("Starting conversion of shard: %v", si.FullPath(opts.DataPath))
if err := convertShard(si, t); err != nil {
log.Fatalf("Failed to convert %v: %v\n", si.FullPath(opts.DataPath), err)
}
log.Printf("Conversion of %v successful (%v)\n", si.FullPath(opts.DataPath), time.Since(start))
})
}
done := make(chan struct{})
go func() {
t.wg.Wait()
close(done)
}()
WAIT_LOOP:
for {
select {
case <-done:
break WAIT_LOOP
case <-time.After(opts.UpdateInterval):
t.StatusUpdate()
}
}
t.Stats.TotalTime = time.Since(conversionStart)
return nil
}
func (t *tracker) StatusUpdate() {
shardCount := atomic.LoadUint64(&t.Stats.CompletedShards)
pointCount := atomic.LoadUint64(&t.Stats.PointsRead)
pointWritten := atomic.LoadUint64(&t.Stats.PointsWritten)
log.Printf("Still Working: Completed Shards: %d/%d Points read/written: %d/%d", shardCount, len(t.shards), pointCount, pointWritten)
}
func (t *tracker) PrintStats() {
preSize := t.shards.Size()
postSize := int64(t.Stats.TsmBytesWritten)
fmt.Printf("\nSummary statistics\n========================================\n")
fmt.Printf("Databases converted: %d\n", len(t.shards.Databases()))
fmt.Printf("Shards converted: %d\n", len(t.shards))
fmt.Printf("TSM files created: %d\n", t.Stats.TsmFilesCreated)
fmt.Printf("Points read: %d\n", t.Stats.PointsRead)
fmt.Printf("Points written: %d\n", t.Stats.PointsWritten)
fmt.Printf("NaN filtered: %d\n", t.Stats.NanFiltered)
fmt.Printf("Inf filtered: %d\n", t.Stats.InfFiltered)
fmt.Printf("Points without fields filtered: %d\n", t.Stats.FieldsFiltered)
fmt.Printf("Disk usage pre-conversion (bytes): %d\n", preSize)
fmt.Printf("Disk usage post-conversion (bytes): %d\n", postSize)
fmt.Printf("Reduction factor: %d%%\n", 100*(preSize-postSize)/preSize)
fmt.Printf("Bytes per TSM point: %.2f\n", float64(postSize)/float64(t.Stats.PointsWritten))
fmt.Printf("Total conversion time: %v\n", t.Stats.TotalTime)
fmt.Println()
}
|