PT-1978 Implemented duplicated indexes detection (#537)

* PT-1978 Implemented duplicated indexes detection

* WIP

* PT-1978 WIP

* PT-1978 New tool pt-mongodb-check-index

* Removed unused file

* Added version command

* Updated required params

* Added system.profile to the ingored dbs

* Fixed for CR

src/go/pt-mongodb-index-check/README.md

@@ -0,0 +1,48 @@
+# pt-mongodb-index-check
+
+
+
+## Introduction
+
+This tool can perform checks on MongoDB indexes. 
+
+Currently, these checks are available:
+
+### Duplicated indexes
+
+Check for indexes that are the prefix of other indexes. For example if we have these 2 indexes
+
+```
+db.getSiblingDB("testdb").test_col.createIndex({"f1": 1, "f2": -1, "f3": 1, "f4": 1}, {"name": "idx_01"});
+db.getSiblingDB("testdb").test_col.createIndex({"f1": 1, "f2": -1, "f3": 1}, {"name": "idx_02"});
+```
+
+The index `idx_02` is the prefix of `idx_01` because it has the same keys in the same order so, `idx_02` can be dropped.
+
+### Unused indexes.
+
+This check gets the `$indexstats` for all indexes and reports those having `accesses.ops` = 0.
+
+## Usage
+
+Run the program as `pt-mongodb-index-check <command> [flags]`
+
+#### Available commands
+
+| Command          | Description                        |
+| ---------------- | ---------------------------------- |
+| check-duplicated | Run checks for duplicated indexes. |
+| check-unused     | Run check for unused indexes.      |
+| check-all        | Run all checks                     |
+
+#### Available flags
+
+| Flag | Description |
+| ---- | ----------- |
+|--all-databases|Check in all databases excluding system dbs.|
+|--databases=DATABASES,...|Comma separated list of databases to check.|
+|--all-collections|Check in all collections in the selected databases.|
+|--collections=COLLECTIONS,...|Comma separated list of collections to check.|
+|--mongodb.uri=<connection string>|Connection URI|
+|--json|Show output as JSON|
+

src/go/pt-mongodb-index-check/indexes/duplicated.go

@@ -0,0 +1,103 @@
+package indexes
+
+import (
+	"context"
+	"log"
+	"sort"
+	"strings"
+
+	"go.mongodb.org/mongo-driver/bson/primitive"
+	"go.mongodb.org/mongo-driver/mongo"
+)
+
+type collectionIndex struct {
+	Name      string      `bson:"name"`
+	Namespace string      `bson:"ns"`
+	V         int         `bson:"v"`
+	Key       primitive.D `bson:"key"`
+}
+
+func (di collectionIndex) ComparableKey() string {
+	str := ""
+	for _, elem := range di.Key {
+		str += sign(elem) + elem.Key
+	}
+	return str
+}
+
+func sign(elem primitive.E) string {
+	sign := "+"
+	switch elem.Value.(type) {
+	case int32: // internal MongoDB indexes like _id_ or lastUsed have the sign field as int32.
+		if elem.Value.(int32) < 0 {
+			sign = "-"
+		}
+	case float64: // All other indexes have the sign field as float64.
+		if elem.Value.(float64) < 0 {
+			sign = "-"
+		}
+	}
+	return sign
+}
+
+// IndexKey holds the list of fields that are part of an index, along with the field order.
+type IndexKey []primitive.E
+
+// String returns the index fields as a string. The + sign means ascending on this field
+// and a - sign indicates a descending order for that field.
+func (di IndexKey) String() string {
+	str := ""
+	for _, elem := range di {
+		str += sign(elem) + elem.Key + " "
+	}
+
+	return str
+}
+
+// DuplicateIndex represents a duplicated index pair.
+// An index is considered as the duplicate of another one if it is it's prefix.
+// Example: the index +f1-f2 is the prefix of +f1-f2+f3.
+type Duplicate struct {
+	Namespace     string
+	Name          string
+	Key           IndexKey
+	ContainerName string
+	ContainerKey  IndexKey
+}
+
+func FindDuplicated(ctx context.Context, client *mongo.Client, database, collection string) ([]Duplicate, error) {
+	di := []Duplicate{}
+
+	cursor, err := client.Database(database).Collection(collection).Indexes().List(ctx, nil)
+	if err != nil {
+		return nil, err
+	}
+
+	var results []collectionIndex
+	if err = cursor.All(context.TODO(), &results); err != nil {
+		log.Fatal(err)
+	}
+
+	sort.Slice(results, func(i, j int) bool {
+		return results[i].ComparableKey() < results[j].ComparableKey()
+	})
+
+	for i := 0; i < len(results)-1; i++ {
+		for j := i + 1; j < len(results); j++ {
+			if strings.HasPrefix(results[j].ComparableKey(), results[i].ComparableKey()) {
+				idx := Duplicate{
+					Namespace:     database + "." + collection,
+					Name:          results[i].Name,
+					Key:           make([]primitive.E, len(results[i].Key)),
+					ContainerName: results[j].Name,
+					ContainerKey:  make([]primitive.E, len(results[j].Key)),
+				}
+				copy(idx.Key, results[i].Key)
+				copy(idx.ContainerKey, results[j].Key)
+				di = append(di, idx)
+			}
+		}
+	}
+
+	return di, nil
+}

src/go/pt-mongodb-index-check/indexes/duplicated_test.go

@@ -0,0 +1,116 @@
+package indexes
+
+import (
+	"context"
+	"fmt"
+	"testing"
+	"time"
+
+	"github.com/AlekSi/pointer"
+	tu "github.com/percona/percona-toolkit/src/go/internal/testutils"
+	"github.com/stretchr/testify/assert"
+	"go.mongodb.org/mongo-driver/bson/primitive"
+	"go.mongodb.org/mongo-driver/mongo"
+	"go.mongodb.org/mongo-driver/mongo/options"
+	"gopkg.in/mgo.v2/bson"
+)
+
+func TestDuplicateIndexes(t *testing.T) {
+	ctx, cancel := context.WithTimeout(context.Background(), 5*time.Second)
+	defer cancel()
+
+	client, err := tu.TestClient(ctx, tu.MongoDBShard1PrimaryPort)
+	if err != nil {
+		t.Fatalf("cannot get a new MongoDB client: %s", err)
+	}
+
+	dbname := "test_db"
+	collname := "test_col"
+
+	database := client.Database(dbname)
+	database.Drop(ctx)       //nolint:errcheck
+	defer database.Drop(ctx) //nolint:errcheck
+
+	_, err = database.Collection(collname).InsertOne(ctx, bson.M{"f1": 1, "f2": "2", "f3": "a", "f4": "c"})
+	assert.NoError(t, err)
+
+	testCases := []primitive.D{
+		{{"f1", 1}, {"f2", -1}, {"f3", 1}, {"f4", 1}},
+		{{"f1", 1}, {"f2", -1}, {"f3", 1}, {"f4", 1}}, // this will throw a duplicate index error
+		{{"f1", 1}, {"f2", -1}, {"f3", 1}},
+		{{"f1", 1}, {"f2", -1}},
+		{{"f3", -1}},
+	}
+
+	errCount := 0
+	for i, tc := range testCases {
+		mod := mongo.IndexModel{
+			Keys: tc,
+			Options: &options.IndexOptions{
+				Name: pointer.ToString(fmt.Sprintf("idx_%02d", i)),
+			},
+		}
+		_, err := database.Collection(collname).Indexes().CreateOne(ctx, mod)
+		if err != nil {
+			errCount++
+		}
+	}
+	/*
+	 At this point we have 5 indexes: _id: (MongoDB's default), idx_00, idx_02, idx_03, idx_04.
+	 idx_01 wasn't created since it duplicates idx_00 and errCount=1.
+	*/
+
+	assert.Equal(t, 1, errCount)
+
+	want := []Duplicate{
+		{
+			Name:      "idx_03",
+			Namespace: "test_db.test_col",
+			Key: IndexKey{
+				{Key: "f1", Value: int32(1)},
+				{Key: "f2", Value: int32(-1)},
+			},
+			ContainerName: "idx_02",
+			ContainerKey: IndexKey{
+				{Key: "f1", Value: int32(1)},
+				{Key: "f2", Value: int32(-1)},
+				{Key: "f3", Value: int32(1)},
+			},
+		},
+		{
+			Name:      "idx_03",
+			Namespace: "test_db.test_col",
+			Key: IndexKey{
+				{Key: "f1", Value: int32(1)},
+				{Key: "f2", Value: int32(-1)},
+			},
+			ContainerName: "idx_00",
+			ContainerKey: IndexKey{
+				{Key: "f1", Value: int32(1)},
+				{Key: "f2", Value: int32(-1)},
+				{Key: "f3", Value: int32(1)},
+				{Key: "f4", Value: int32(1)},
+			},
+		},
+		{
+			Name:      "idx_02",
+			Namespace: "test_db.test_col",
+			Key: IndexKey{
+				{Key: "f1", Value: int32(1)},
+				{Key: "f2", Value: int32(-1)},
+				{Key: "f3", Value: int32(1)},
+			},
+			ContainerName: "idx_00",
+			ContainerKey: IndexKey{
+				{Key: "f1", Value: int32(1)},
+				{Key: "f2", Value: int32(-1)},
+				{Key: "f3", Value: int32(1)},
+				{Key: "f4", Value: int32(1)},
+			},
+		},
+	}
+
+	di, err := FindDuplicated(ctx, client, dbname, collname)
+	assert.NoError(t, err)
+	assert.Equal(t, want, di)
+}

src/go/pt-mongodb-index-check/indexes/unused.go

@@ -0,0 +1,62 @@
+package indexes
+
+import (
+	"context"
+
+	"github.com/pkg/errors"
+	"go.mongodb.org/mongo-driver/bson/primitive"
+	"go.mongodb.org/mongo-driver/mongo"
+	"gopkg.in/mgo.v2/bson"
+)
+
+var systemDBs = []string{"admin", "config", "local", "system.profile"} //nolint:gochecknoglobals
+// IndexStat hold an index usage statistics.
+type IndexStat struct {
+	Accesses struct {
+		Ops   int64              `bson:"ops"`
+		Since primitive.DateTime `bson:"since"`
+	} `bson:"accesses"`
+	Spec struct {
+		Name      string      `bson:"name"`
+		Namespace string      `bson:"ns"`
+		V         int32       `bson:"v"`
+		Key       primitive.D `bson:"key"`
+	} `bson:"spec"`
+	Name string      `bson:"name"`
+	Key  primitive.D `bson:"key"`
+	Host string      `bson:"host"`
+}
+
+func in(search string, items []string) bool {
+	for _, item := range items {
+		if search == item {
+			return true
+		}
+	}
+	return false
+}
+
+// FindUnusedIndexes returns a list of unused indexes for the given database and collection.
+func FindUnused(ctx context.Context, client *mongo.Client, database, collection string) ([]IndexStat, error) {
+	aggregation := mongo.Pipeline{
+		{{Key: "$indexStats", Value: primitive.M{}}},
+		{{Key: "$match", Value: primitive.M{"accesses.ops": 0}}},
+		{{Key: "$match", Value: primitive.M{"name": bson.M{"$ne": "_id_"}}}},
+	}
+
+	if in(database, systemDBs) {
+		return nil, nil
+	}
+
+	cursor, err := client.Database(database).Collection(collection).Aggregate(ctx, aggregation)
+	if err != nil {
+		return nil, errors.Wrap(err, "cannot run $indexStats for unused indexes")
+	}
+
+	var stats []IndexStat
+	if err = cursor.All(ctx, &stats); err != nil {
+		return nil, errors.Wrap(err, "cannot get $indexStats for unused indexes")
+	}
+
+	return stats, nil
+}

src/go/pt-mongodb-index-check/indexes/unused_test.go

@@ -0,0 +1,81 @@
+package indexes
+
+import (
+	"context"
+	"fmt"
+	"math/rand"
+	"sort"
+	"testing"
+	"time"
+
+	"github.com/AlekSi/pointer"
+	tu "github.com/percona/percona-toolkit/src/go/internal/testutils"
+	"github.com/stretchr/testify/assert"
+	"go.mongodb.org/mongo-driver/bson/primitive"
+	"go.mongodb.org/mongo-driver/mongo"
+	"go.mongodb.org/mongo-driver/mongo/options"
+	"gopkg.in/mgo.v2/bson"
+)
+
+func TestUnusedIndexes(t *testing.T) {
+	ctx, cancel := context.WithTimeout(context.Background(), 5*time.Second)
+	defer cancel()
+
+	client, err := tu.TestClient(ctx, tu.MongoDBShard1PrimaryPort)
+	if err != nil {
+		t.Fatalf("cannot get a new MongoDB client: %s", err)
+	}
+
+	dbname := "test_db"
+	collname := "test_col"
+
+	database := client.Database(dbname)
+	database.Drop(ctx)       //nolint:errcheck
+	defer database.Drop(ctx) //nolint:errcheck
+
+	testCases := []primitive.D{
+		{{"f1", 1}, {"f2", -1}, {"f3", 1}, {"f4", 1}},
+		{{"f3", -1}},
+		{{"f4", -1}},
+	}
+
+	errCount := 0
+	for i, tc := range testCases {
+		mod := mongo.IndexModel{
+			Keys: tc,
+			Options: &options.IndexOptions{
+				Name: pointer.ToString(fmt.Sprintf("idx_%02d", i)),
+			},
+		}
+		_, err := database.Collection(collname).Indexes().CreateOne(ctx, mod)
+		if err != nil {
+			errCount++
+		}
+	}
+
+	for i := 0; i < 100; i++ {
+		_, err = database.Collection(collname).InsertOne(ctx,
+			bson.M{"f1": rand.Int63n(1000), "f2": rand.Int63n(1000), "f3": rand.Int63n(1000), "f4": rand.Int63n(1000)})
+		assert.NoError(t, err)
+	}
+
+	// Make use of idx_02: {"f4": -1} to exclude it from the results so we ensure only unused indexes
+	// are being listed.
+	_, err = database.Collection(collname).Find(ctx, primitive.M{"f4": primitive.M{"$gt": 500}})
+	assert.NoError(t, err)
+
+	want := []string{"idx_00", "idx_01"}
+
+	ui, err := FindUnused(ctx, client, dbname, collname)
+	assert.NoError(t, err)
+
+	got := make([]string, 0, len(ui))
+	for _, idx := range ui {
+		// compare only names because the index struct has a timestamp in it and it is variable.
+		got = append(got, idx.Name)
+	}
+
+	sort.Strings(got)
+
+	assert.Equal(t, want, got)
+}

src/go/pt-mongodb-index-check/main.go

@@ -0,0 +1,189 @@
+package main
+
+import (
+	"bytes"
+	"context"
+	"encoding/json"
+	"fmt"
+	"strings"
+	"text/template"
+	"time"
+
+	"github.com/alecthomas/kong"
+	"github.com/pkg/errors"
+	"go.mongodb.org/mongo-driver/bson/primitive"
+	"go.mongodb.org/mongo-driver/mongo"
+	"go.mongodb.org/mongo-driver/mongo/options"
+
+	"github.com/percona/percona-toolkit/src/go/pt-mongodb-index-check/indexes"
+	"github.com/percona/percona-toolkit/src/go/pt-mongodb-index-check/templates"
+
+	log "github.com/sirupsen/logrus"
+)
+
+type cmdlineArgs struct {
+	CheckUnused     struct{} `cmd:"" name:"check-unused" help:"Check for unused indexes."`
+	CheckDuplicated struct{} `cmd:"" name:"check-duplicates" help:"Check for duplicated indexes."`
+	CheckAll        struct{} `cmd:"" name:"check-all" help:"Check for unused and duplicated indexes."`
+	ShowHelp        struct{} `cmd:"" default:"1"`
+	Version         struct{} `cmd:"" name:"version"`
+
+	AllDatabases bool     `name:"all-databases" xor:"db" help:"Check in all databases excluding system dbs"`
+	Databases    []string `name:"databases" xor:"db" help:"Comma separated list of databases to check"`
+
+	AllCollections bool     `name:"all-collections" xor:"colls" help:"Check in all collections in the selected databases."`
+	Collections    []string `name:"collections" xor:"colls" help:"Comma separated list of collections to check"`
+	URI            string   `name:"mongodb.uri" required:"" placeholder:"mongodb://host:port/admindb?options" help:"Connection URI"`
+	JSON           bool     `name:"json" help:"Show output as JSON"`
+}
+
+type response struct {
+	Unused     []indexes.IndexStat
+	Duplicated []indexes.Duplicate
+}
+
+const (
+	TOOLNAME = "pt-mongodb-index-check"
+)
+
+var (
+	Build     string = "2020-04-23" //nolint
+	GoVersion string = "1.14.1"     //nolint
+	Version   string = "3.4.0"      //nolint
+	Commit    string                //nolint
+)
+
+func main() {
+	var args cmdlineArgs
+	kongctx := kong.Parse(&args, kong.UsageOnError())
+
+	if kongctx.Command() == "version" {
+		fmt.Println(TOOLNAME)
+		fmt.Printf("Version %s\n", Version)
+		fmt.Printf("Build: %s using %s\n", Build, GoVersion)
+		fmt.Printf("Commit: %s\n", Commit)
+		return
+	}
+
+	ctx, cancel := context.WithTimeout(context.Background(), 20*time.Second)
+	defer cancel()
+
+	if !strings.HasPrefix(args.URI, "mongodb") && !strings.HasPrefix(args.URI, "mongodb+srv") {
+		args.URI = "mongodb://" + args.URI
+	}
+
+	client, err := mongo.Connect(ctx, options.Client().ApplyURI(args.URI))
+	if err != nil {
+		log.Fatalf("Cannot connect to the database: %q", err)
+	}
+
+	if args.AllDatabases {
+		args.Databases, err = client.ListDatabaseNames(context.TODO(), primitive.D{})
+		if err != nil {
+			log.Fatalf("cannot list all databases: %s", err)
+		}
+	}
+	if args.AllCollections {
+		args.Collections = nil
+	}
+
+	resp := response{}
+
+	switch kongctx.Command() {
+	case "check-unused":
+		resp.Unused = findUnused(ctx, client, args.Databases, args.Collections)
+	case "check-duplicates":
+		resp.Duplicated = findDuplicated(ctx, client, args.Databases, args.Collections)
+	case "check-all":
+		resp.Unused = findUnused(ctx, client, args.Databases, args.Collections)
+		resp.Duplicated = findDuplicated(ctx, client, args.Databases, args.Collections)
+	default:
+		kong.DefaultHelpPrinter(kong.HelpOptions{}, kongctx)
+	}
+
+	fmt.Println(output(resp, args.JSON))
+}
+
+func output(resp response, asJson bool) string {
+	if asJson {
+		jsonStr, err := json.MarshalIndent(resp, "", "\t")
+		if err != nil {
+			log.Fatal("cannot encode the response as json")
+		}
+		return string(jsonStr)
+	}
+
+	buf := new(bytes.Buffer)
+
+	t := template.Must(template.New("duplicated").Parse(templates.Duplicated))
+	if err := t.Execute(buf, resp.Duplicated); err != nil {
+		log.Fatal(errors.Wrap(err, "cannot parse clusterwide section of the output template"))
+	}
+
+	t = template.Must(template.New("unused").Parse(templates.Unused))
+	if err := t.Execute(buf, resp.Unused); err != nil {
+		log.Fatal(errors.Wrap(err, "cannot parse clusterwide section of the output template"))
+	}
+
+	return buf.String()
+}
+
+func findUnused(ctx context.Context, client *mongo.Client, databases []string, collections []string) []indexes.IndexStat {
+	unused := []indexes.IndexStat{}
+	var err error
+
+	colls := make([]string, len(collections))
+	copy(colls, collections)
+
+	for _, database := range databases {
+		if len(collections) == 0 {
+			colls, err = client.Database(database).ListCollectionNames(ctx, primitive.D{})
+			if err != nil {
+				log.Errorf("cannot get the list of collections for the database %s", database)
+				continue
+			}
+		}
+
+		for _, collection := range colls {
+			idx, err := indexes.FindUnused(ctx, client, database, collection)
+			if err != nil {
+				log.Errorf("error while checking unused indexes in %s.%s: %s", database, collection, err)
+				continue
+			}
+
+			unused = append(unused, idx...)
+		}
+	}
+
+	return unused
+}
+
+func findDuplicated(ctx context.Context, client *mongo.Client, databases []string, collections []string) []indexes.Duplicate {
+	duplicated := []indexes.Duplicate{}
+	var err error
+
+	colls := make([]string, len(collections))
+	copy(colls, collections)
+
+	for _, database := range databases {
+		if len(collections) == 0 {
+			colls, err = client.Database(database).ListCollectionNames(ctx, primitive.D{})
+			if err != nil {
+				log.Errorf("cannot get the list of collections for the database %s", database)
+				continue
+			}
+		}
+
+		for _, collection := range colls {
+			dups, err := indexes.FindDuplicated(ctx, client, database, collection)
+			if err != nil {
+				log.Errorf("error while checking duplicated indexes in %s.%s: %s", database, collection, err)
+				continue
+			}
+
+			duplicated = append(duplicated, dups...)
+		}
+	}
+
+	return duplicated
+}

src/go/pt-mongodb-index-check/templates/duplicated.go

@@ -0,0 +1,10 @@
+package templates
+
+// {{if $i}},{{end}} adds a comma after the first element.
+// When $i == 0 (first element) {{ if $i }} returns false (0)
+
+var Duplicated = `
+Duplicated indexes 
+{{ range . }}
+{{ .Namespace }}, index '{{ .Name }}', with fields { {{- range $i, $val := .Key }}{{if $i}}, {{end}}{{ $val.Key }}:{{ $val.Value }}{{ end -}} } is the prefix of '{{ .ContainerName }}' with fields { {{- range $i, $val := .ContainerKey }}{{if $i}}, {{end}}{{ $val.Key }}:{{ $val.Value }}{{ end -}} }{{ end}}
+`

src/go/pt-mongodb-index-check/templates/unused.go

@@ -0,0 +1,10 @@
+package templates
+
+// {{if $i}},{{end}} adds a comma after the first element.
+// When $i == 0 (first element) {{ if $i }} returns false (0)
+
+var Unused = `
+Unused indexes since last restart
+{{ range . }}
+{{ .Spec.Namespace }}, index '{{ .Name }}' with fields { {{- range $i, $val := .Key }}{{if $i}}, {{end}}{{ $val.Key }}:{{ $val.Value }} }{{ end }}{{ end}}
+`