roaring-migrate bug;performance improvements

This commit is contained in:
Todd Gruben 2022-02-11 11:19:44 -06:00
parent a21dd96b2d
commit 8d5cdbdd77

View file

@ -4,6 +4,7 @@ package main
import (
"encoding/json"
"fmt"
"io"
"io/ioutil"
"os"
"path/filepath"
@ -17,11 +18,20 @@ import (
"github.com/molecula/featurebase/v3/rbf/cfg"
"github.com/molecula/featurebase/v3/roaring"
txkey "github.com/molecula/featurebase/v3/short_txkey"
"github.com/molecula/featurebase/v3/vprint"
"github.com/spf13/cobra"
)
var visited map[string]int64
const (
Version = "1.0"
)
func main() {
visited = make(map[string]int64)
var dataDir, backupPath string
var verbose bool
cmdMigrate := &cobra.Command{
Use: "roaring-migrate",
Short: "convert roaring pilosa backup to rbf",
@ -29,7 +39,7 @@ func main() {
Run: func(cmd *cobra.Command, args []string) {
nodes := strings.Split(dataDir, ",")
for _, nodePath := range nodes {
err := Migrate(nodePath, backupPath)
err := Migrate(nodePath, backupPath, verbose)
if err != nil {
fmt.Println("Error", err)
return
@ -40,6 +50,7 @@ func main() {
}
cmdMigrate.Flags().StringVarP(&dataDir, "data-dir", "d", "", "source directories for each node seperated by commas")
cmdMigrate.Flags().StringVarP(&backupPath, "backup-dir", "b", "", "location of backup directory")
cmdMigrate.Flags().BoolVar(&verbose, "verbose", false, "addition progress information")
err := cmdMigrate.MarkFlagRequired("data-dir")
if err != nil {
fmt.Println("Error setting flag data-dir")
@ -52,6 +63,9 @@ func main() {
os.Exit(1)
return
}
if verbose {
vprint.VV("Version: %v", Version)
}
err = cmdMigrate.Execute()
if err != nil {
@ -94,6 +108,14 @@ type local struct {
Fields []*pilosa.FieldInfo `json:"fields,omitempty"`
}
func fileExists(filename string) (bool, int64) {
info, err := os.Stat(filename)
if os.IsNotExist(err) {
return false, 0
}
return !info.IsDir(), info.Size()
}
func BuildSchema(dataDir string) ([]byte, error) {
//need to find all the ".meta" files and load as field options
@ -197,20 +219,34 @@ func (d *rbfFile) getDB(path, index string, shard uint64) (*rbf.DB, error) {
return d.working, nil
}
func (d *rbfFile) Close() error {
defer func() error {
//cleanup the tempdirectory
err := os.RemoveAll(d.temp)
if err != nil {
return err
}
return nil
}()
if d.last != "" {
d.working.Close()
//if d.last exists only keep the biggest
err := os.MkdirAll(filepath.Dir(d.last), 0777)
if err != nil {
return err
exists, sz := fileExists(d.last)
src := filepath.Join(d.temp, "data")
if !exists {
err := os.MkdirAll(filepath.Dir(d.last), 0777)
if err != nil {
return err
}
} else {
_, sz2 := fileExists(src)
if sz > sz2 {
return nil
}
}
// move the datafile backup shard
err = os.Rename(filepath.Join(d.temp, "data"), d.last)
if err != nil {
return err
}
//cleanup the tempdirectory
err = os.RemoveAll(d.temp)
err := os.Rename(src, d.last)
if err != nil {
return err
}
@ -218,19 +254,26 @@ func (d *rbfFile) Close() error {
return nil
}
func copyFile(src, dest string) error {
input, err := ioutil.ReadFile(src)
from, err := os.Open(src)
if err != nil {
return err
}
defer from.Close()
err = ioutil.WriteFile(dest, input, 0644)
to, err := os.OpenFile(dest, os.O_RDWR|os.O_CREATE, 0644)
if err != nil {
return err
}
defer to.Close()
_, err = io.Copy(to, from)
if err != nil {
return err
}
return nil
}
func Migrate(dataDir, backupPath string) error {
func Migrate(dataDir, backupPath string, verbose bool) error {
dataDir = strings.TrimSuffix(dataDir, "/")
err := os.MkdirAll(backupPath, 0777)
@ -279,7 +322,21 @@ func Migrate(dataDir, backupPath string) error {
bm := roaring.NewSliceBitmap()
for _, filename := range raw {
index, field, view, shard := Extract(filename)
sz, before := visited[filename]
fi, _ := os.Stat(dataDir + filename)
if field != "_exists" {
if !before {
visited[filename] = fi.Size()
} else {
if fi.Size() <= sz {
continue //skipp it
}
visited[filename] = fi.Size()
}
}
if verbose {
vprint.VV("processing: %v", dataDir+filename)
}
content, err := ioutil.ReadFile(dataDir + filename)
if err != nil {
return err
@ -293,35 +350,19 @@ func Migrate(dataDir, backupPath string) error {
if err != nil {
return err
}
tx, err := db.Begin(true)
if err != nil {
return err
}
key := string(txkey.Prefix(index, field, view, shard))
itr, ok := bm.Containers.Iterator(0)
if ok {
for itr.Next() {
k, v := itr.Value()
tx.PutContainer(key, k, v)
}
}
tx, err := db.Begin(true)
tx.AddRoaring(key, bm)
err = tx.Commit()
if err != nil {
return err
}
}
cache.Close()
keys := FetchIndexKeys(dataDir)
for _, filename := range keys {
fmt.Println("index keys", filename)
content, err := ioutil.ReadFile(filepath.Join(dataDir, filename))
if err != nil {
return err
}
srcFile := filepath.Join(dataDir, filename)
parts := strings.Split(filename, "/")
destFile := filepath.Join(backupPath, "indexes", parts[1], "translate", parts[3])
err = writeIfBigger(destFile, content)
err = writeIfBigger(destFile, srcFile)
if err != nil {
return err
}
@ -331,13 +372,10 @@ func Migrate(dataDir, backupPath string) error {
keys = FetchRowkeys(dataDir)
for _, filename := range keys {
fmt.Println("field", filename)
content, err := ioutil.ReadFile(dataDir + filename)
if err != nil {
return err
}
srcFile := dataDir + filename
parts := strings.Split(filename, "/")
destFile := filepath.Join(backupPath, "indexes", parts[1], "fields", parts[2], "translate")
err = writeIfBigger(destFile, content)
err = writeIfBigger(destFile, srcFile)
if err != nil {
return err
}
@ -345,16 +383,21 @@ func Migrate(dataDir, backupPath string) error {
return nil
}
func writeIfBigger(dst string, content []byte) error {
func writeIfBigger(dst string, srcFile string) error {
if stats, err := os.Stat(dst); os.IsNotExist(err) {
err = os.MkdirAll(filepath.Dir(dst), 0777)
if err != nil {
return err
}
return ioutil.WriteFile(dst, content, 0644)
return copyFile(srcFile, dst)
} else {
if stats.Size() < int64(len(content)) {
return ioutil.WriteFile(dst, content, 0644)
stats2, err := os.Stat(srcFile)
if err != nil {
return err
}
if stats.Size() < stats2.Size() {
vprint.VV("Bigger %v %v", stats.Size(), stats2.Size())
return copyFile(srcFile, dst)
}
}
return nil //simply skip it