Update compression method and expose badger options

go.mod

@@ -16,14 +16,16 @@ require (
 	github.com/armon/go-metrics v0.3.3 // indirect
 	github.com/aws/aws-sdk-go v1.30.25
 	github.com/crphang/orc v0.0.3
-	github.com/dgraph-io/badger/v2 v2.0.3
+	github.com/dgraph-io/badger/v2 v2.0.1-rc1.0.20200626160443-3042e3776798
 	github.com/dgraph-io/ristretto v0.0.2 // indirect
 	github.com/dgryski/go-farm v0.0.0-20200201041132-a6ae2369ad13 // indirect
 	github.com/dnaeon/go-vcr v1.0.1 // indirect
 	github.com/emitter-io/address v1.0.0
 	github.com/gogo/protobuf v1.3.1
+	github.com/golang/groupcache v0.0.0-20200121045136-8c9f03a8e57e // indirect
 	github.com/golang/protobuf v1.4.1 // indirect
 	github.com/golang/snappy v0.0.1
+	github.com/google/btree v1.0.0 // indirect
 	github.com/gopherjs/gopherjs v0.0.0-20200209183636-89e6cbcd0b6d // indirect
 	github.com/gorilla/mux v1.7.4
 	github.com/grab/async v0.0.5
@@ -39,6 +41,8 @@ require (
 	github.com/kelindar/loader v0.0.10
 	github.com/kelindar/lua v0.0.6
 	github.com/miekg/dns v1.1.29 // indirect
+	github.com/modern-go/concurrent v0.0.0-20180306012644-bacd9c7ef1dd // indirect
+	github.com/modern-go/reflect2 v1.0.1 // indirect
 	github.com/myteksi/hystrix-go v1.1.3
 	github.com/samuel/go-thrift v0.0.0-20191111193933-5165175b40af
 	github.com/satori/go.uuid v1.2.0 // indirect

go.sum

@@ -114,8 +114,8 @@ github.com/crphang/orc v0.0.3/go.mod h1:+siY09J77eYa8+0+UXZxeyv8nrmhvOY40DbqB7gR
 github.com/davecgh/go-spew v1.1.0/go.mod h1:J7Y8YcW2NihsgmVo/mv3lAwl/skON4iLHjSsI+c5H38=
 github.com/davecgh/go-spew v1.1.1 h1:vj9j/u1bqnvCEfJOwUhtlOARqs3+rkHYY13jYWTU97c=
 github.com/davecgh/go-spew v1.1.1/go.mod h1:J7Y8YcW2NihsgmVo/mv3lAwl/skON4iLHjSsI+c5H38=
-github.com/dgraph-io/badger/v2 v2.0.3 h1:inzdf6VF/NZ+tJ8RwwYMjJMvsOALTHYdozn0qSl6XJI=
-github.com/dgraph-io/badger/v2 v2.0.3/go.mod h1:3KY8+bsP8wI0OEnQJAKpd4wIJW/Mm32yw2j/9FUVnIM=
+github.com/dgraph-io/badger/v2 v2.0.1-rc1.0.20200626160443-3042e3776798 h1:PcjL/wjye6pw6IZK1SEKqp+wPpBm+YFx8gO7lBFArW8=
+github.com/dgraph-io/badger/v2 v2.0.1-rc1.0.20200626160443-3042e3776798/go.mod h1:3KY8+bsP8wI0OEnQJAKpd4wIJW/Mm32yw2j/9FUVnIM=
 github.com/dgraph-io/ristretto v0.0.2-0.20200115201040-8f368f2f2ab3 h1:MQLRM35Pp0yAyBYksjbj1nZI/w6eyRY/mWoM1sFf4kU=
 github.com/dgraph-io/ristretto v0.0.2-0.20200115201040-8f368f2f2ab3/go.mod h1:KPxhHT9ZxKefz+PCeOGsrHpl1qZ7i70dGTu2u+Ahh6E=
 github.com/dgraph-io/ristretto v0.0.2 h1:a5WaUrDa0qm0YrAAS1tUykT5El3kt62KNZZeMxQn3po=

internal/config/config.go

@@ -23,6 +23,16 @@ type Config struct {
 	Statsd   *StatsD    `json:"statsd,omitempty" yaml:"statsd" env:"STATSD"`
 	Computed []Computed `json:"computed" yaml:"computed" env:"COMPUTED"`
 	K8s      *K8s       `json:"k8s,omitempty" yaml:"k8s" env:"K8S"`
+	Badger   Badger     `json:"badger,omitempty" yaml:"badger" env:"badger"`
+}
+
+type Badger struct {
+	SyncWrites          *bool   `json:"syncWrites" yaml:"syncWrites" env:"SYNCWRITES"`                            // Whether to sync writes to disk before ack. defaults to true
+	ValueLogMaxEntries  *uint32 `json:"valueLogMaxEntries" yaml:"valueLogMaxEntries" env:"VALUELOGMAXENTRIES"`    // maximum number of entries a value log file can hold approximately. defaults to 5000
+	MaxTableSize        *int64  `json:"maxTableSize" yaml:"maxTableSize" env:"MAXTABLESIZE"`                      // maximum size in bytes for each LSM table or file.
+	LevelOneSize        *int64  `json:"levelOneSize" yaml:"levelOneSize" env:"LEVELONESIZE"`                      // maximum total size in bytes for Level 1. defaults to 1 Million
+	LevelSizeMultiplier *int    `json:"levelSizeMultiplier" yaml:"levelSizeMultiplier" env:"LEVELSIZEMULTIPLIER"` // the ratio between the maximum sizes of contiguous levels in the LSM. defaults to 10
+	MaxLevels           *int    `json:"maxLevels" yaml:"maxLevels" env:"MAXLEVELS"`                               // maximum number of levels of compaction allowed in the LSM. defaults to 7
 }
 
 type K8s struct {

internal/server/server_ingest.go

@@ -40,5 +40,6 @@ func (s *Server) Ingest(ctx context.Context, request *talaria.IngestRequest) (*t
 		}
 	}
 
+	s.monitor.Count("server", "ingestCount", int64(len(blocks)))
 	return nil, nil
 }

internal/storage/compact/compact.go

@@ -82,6 +82,7 @@ func (s *Storage) Delete(keys ...key.Key) error {
 
 // Compact runs the compaction on the storage
 func (s *Storage) Compact(ctx context.Context) (interface{}, error) {
+	st := time.Now()
 	var hash uint32
 	var blocks []block.Block
 	var merged []key.Key
@@ -134,7 +135,9 @@ func (s *Storage) Compact(ctx context.Context) (interface{}, error) {
 
 	// Wait for the pool to be close
 	close(queue)
-	return wpool.Outcome()
+	out, err := wpool.Outcome()
+	s.monitor.Histogram(ctxTag, "compactlatency", float64(time.Since(st)))
+	return out, err
 }
 
 // merge adds an key-value pair to the underlying database
@@ -163,11 +166,14 @@ func (s *Storage) merge(keys []key.Key, blocks []block.Block, schema typeof.Sche
 			}
 		}
 
+		start := time.Now()
 		//  Delete all of the keys that we have appended
 		if err = s.buffer.Delete(keys...); err != nil {
 			s.monitor.Count1(ctxTag, "error", "type:delete")
-			s.monitor.Error(err)
+			s.monitor.Error(errors.Internal("merge error %s", err))
 		}
+		s.monitor.Histogram(ctxTag, "deletelatency", float64(time.Since(start)))
+		s.monitor.Count(ctxTag, "deleteCount", int64(len(keys)))
 		return
 	})
 }

internal/storage/compact/compact_test.go

@@ -11,6 +11,8 @@ import (
 	"testing"
 	"time"
 
+	"github.com/kelindar/talaria/internal/config"
+
 	"github.com/kelindar/talaria/internal/encoding/block"
 	"github.com/kelindar/talaria/internal/encoding/key"
 	"github.com/kelindar/talaria/internal/encoding/typeof"
@@ -60,7 +62,10 @@ func runTest(t *testing.T, test func(store *disk.Storage)) {
 func run(f func(store *disk.Storage)) {
 	dir, _ := ioutil.TempDir("", "test")
 	store := disk.New(monitor.NewNoop())
-	_ = store.Open(dir)
+	syncWrite := false
+	_ = store.Open(dir, config.Badger{
+		SyncWrites: &syncWrite,
+	})
 
 	// Close once we're done and delete data
 	defer func() { _ = os.RemoveAll(dir) }()

internal/storage/disk/disk.go

@@ -16,6 +16,7 @@ import (
 
 	"github.com/dgraph-io/badger/v2"
 	"github.com/grab/async"
+	"github.com/kelindar/talaria/internal/config"
 	"github.com/kelindar/talaria/internal/encoding/key"
 	"github.com/kelindar/talaria/internal/monitor"
 	"github.com/kelindar/talaria/internal/monitor/errors"
@@ -46,18 +47,18 @@ func New(m monitor.Monitor) *Storage {
 }
 
 // Open creates a disk storage and open the directory
-func Open(dir string, name string, monitor monitor.Monitor) *Storage {
+func Open(dir string, name string, monitor monitor.Monitor, options config.Badger) *Storage {
 	diskStorage := New(monitor)
 	tableDir := path.Join(dir, name)
-	err := diskStorage.Open(tableDir)
+	err := diskStorage.Open(tableDir, options)
 	if err != nil {
 		panic(err)
 	}
 	return diskStorage
 }
 
 // Open opens a directory.
-func (s *Storage) Open(dir string) error {
+func (s *Storage) Open(dir string, options config.Badger) error {
 
 	// Default to a /data directory
 	if dir == "" {
@@ -71,14 +72,51 @@ func (s *Storage) Open(dir string) error {
 
 	// Create the options
 	opts := badger.DefaultOptions(dir)
-	opts.SyncWrites = false
-	opts.MaxTableSize = 64 << 15
-	opts.ValueLogMaxEntries = 5000
-	opts.LevelOneSize = 1 << 16
-	opts.LevelSizeMultiplier = 3
-	opts.MaxLevels = 25
-	opts.Truncate = true
-	opts.Logger = &logger{s.monitor}
+	if options.SyncWrites != nil {
+		opts = opts.WithSyncWrites(*options.SyncWrites)
+	}
+
+	// max size of lsm tree in bytes after which data is propagated to disk.
+	// The default is 64 MB.
+	if options.MaxTableSize != nil {
+		opts = opts.WithMaxTableSize(*options.MaxTableSize)
+	}
+
+	// ValueLogMaxEntries sets the maximum number of entries a value log file can hold approximately.
+	// A actual size limit of a value log file is the minimum of ValueLogFileSize and
+	// ValueLogMaxEntries.
+	// The default value of ValueLogMaxEntries is 1 Million
+	if options.ValueLogMaxEntries != nil {
+		opts = opts.WithValueLogMaxEntries(*options.ValueLogMaxEntries)
+	}
+
+	// LevelOneSize sets the maximum total size for Level 1.
+	// The default value of LevelOneSize is 20MB.
+	if options.LevelOneSize != nil {
+		opts = opts.WithLevelOneSize(*options.LevelOneSize)
+	}
+
+	// Maximum number of levels of compaction allowed in the LSM.
+	// The default value of MaxLevels is 7.
+	if options.MaxLevels != nil {
+		opts = opts.WithMaxLevels(*options.MaxLevels)
+	}
+
+	// LevelSizeMultiplier sets the ratio between the maximum sizes of contiguous levels in the LSM.
+	// Once a level grows to be larger than this ratio allowed, the compaction process will be
+	//  triggered.
+	// The default value of LevelSizeMultiplier is 10.
+	if options.LevelSizeMultiplier != nil {
+		opts = opts.WithLevelSizeMultiplier(*options.LevelSizeMultiplier)
+	}
+
+	// Truncate indicates whether value log files should be truncated to delete corrupt data, if any.
+	// This option is ignored when ReadOnly is true.
+	// The default value of Truncate is false.
+	opts = opts.WithTruncate(true)
+
+	opts = opts.WithLogger(&logger{s.monitor})
+	s.monitor.Info("opening badger with options %+v", opts)
 
 	// Attempt to open the database
 	db, err := badger.Open(opts)

internal/storage/disk/disk_test.go

@@ -12,6 +12,8 @@ import (
 	"testing"
 	"time"
 
+	"github.com/kelindar/talaria/internal/config"
+
 	"github.com/kelindar/talaria/internal/encoding/key"
 	"github.com/kelindar/talaria/internal/monitor"
 	"github.com/stretchr/testify/assert"
@@ -28,7 +30,7 @@ func runTest(t *testing.T, test func(store *Storage)) {
 func run(f func(store *Storage)) {
 	dir, _ := ioutil.TempDir("", "test")
 	store := New(monitor.NewNoop())
-	_ = store.Open(dir)
+	_ = store.Open(dir, config.Badger{})
 
 	// Close once we're done and delete data
 	defer func() { _ = os.RemoveAll(dir) }()
@@ -250,7 +252,10 @@ func populate(store *Storage) {
 
 func TestOpen(t *testing.T) {
 	assert.NotPanicsf(t, func() {
-		disk := Open(".", "test-table", monitor.NewNoop())
+		syncWrites := false
+		disk := Open(".", "test-table", monitor.NewNoop(), config.Badger{
+			SyncWrites: &syncWrites,
+		})
 		assert.NotNil(t, disk)
 		disk.Close()
 		os.RemoveAll("test-table")

internal/storage/flush/flush.go

@@ -5,6 +5,7 @@ package flush
 
 import (
 	"bytes"
+	"compress/flate"
 	"sync"
 	"time"
 
@@ -68,7 +69,7 @@ func (s *Storage) Merge(blocks []block.Block, schema typeof.Schema) ([]byte, []b
 	buffer := s.memoryPool.Get().(*bytes.Buffer)
 	writer, err := eorc.NewWriter(buffer,
 		eorc.SetSchema(orcSchema),
-		eorc.SetCompression(eorc.CompressionSnappy{}))
+		eorc.SetCompression(eorc.CompressionZlib{Level: flate.DefaultCompression}))
 
 	for _, blk := range blocks {
 		rows, err := blk.Select(blk.Schema())

internal/storage/flush/flush_test.go

@@ -5,6 +5,7 @@ package flush
 
 import (
 	"bytes"
+	"compress/flate"
 	"io/ioutil"
 	"testing"
 
@@ -74,7 +75,7 @@ func TestMerge(t *testing.T) {
 	orcBuffer := &bytes.Buffer{}
 	writer, _ = eorc.NewWriter(orcBuffer,
 		eorc.SetSchema(orcSchema),
-		eorc.SetCompression(eorc.CompressionSnappy{}))
+		eorc.SetCompression(eorc.CompressionZlib{Level: flate.DefaultCompression}))
 	_ = writer.Write("eventName", 1, 1.0)
 	_ = writer.Write("eventName", 2, 2.0)
 	_ = writer.Close()
@@ -152,7 +153,7 @@ func TestMerge_DifferentSchema(t *testing.T) {
 	orcBuffer := &bytes.Buffer{}
 	writer, _ = eorc.NewWriter(orcBuffer,
 		eorc.SetSchema(orcSchema2),
-		eorc.SetCompression(eorc.CompressionSnappy{}))
+		eorc.SetCompression(eorc.CompressionZlib{Level: flate.DefaultCompression}))
 	_ = writer.Write("eventName", 1, 1.0, nil)
 	_ = writer.Write("eventName", 2, 2.0, "s")
 	_ = writer.Close()

internal/table/log/log.go

@@ -31,7 +31,7 @@ type Table struct {
 
 // New creates a new table implementation.
 func New(cfg config.Func, cluster Membership, monitor monitor.Monitor) *Table {
-	store := disk.Open(cfg().Storage.Directory, cfg().Tables.Log.Name, monitor)
+	store := disk.Open(cfg().Storage.Directory, cfg().Tables.Log.Name, monitor, cfg().Badger)
 	base := timeseries.New(cluster, monitor, store, timeseries.Config{
 		Name:   cfg().Tables.Log.Name,
 		TTL:    cfg().Tables.Log.TTL,

internal/table/timeseries/timeseries_test.go

@@ -4,6 +4,10 @@
 package timeseries_test
 
 import (
+	"io/ioutil"
+	"os"
+	"testing"
+
 	"github.com/kelindar/talaria/internal/config"
 	"github.com/kelindar/talaria/internal/encoding/block"
 	"github.com/kelindar/talaria/internal/encoding/typeof"
@@ -12,9 +16,6 @@ import (
 	"github.com/kelindar/talaria/internal/storage/disk"
 	"github.com/kelindar/talaria/internal/table/timeseries"
 	"github.com/stretchr/testify/assert"
-	"io/ioutil"
-	"os"
-	"testing"
 )
 
 const testFile2 = "../../../test/test2.orc"
@@ -56,7 +57,7 @@ func TestTimeseries_DynamicSchema(t *testing.T) {
 	}
 
 	monitor := monitor2.NewNoop()
-	store := disk.Open(dir, timeseriesCfg.Name, monitor)
+	store := disk.Open(dir, timeseriesCfg.Name, monitor, config.Badger{})
 
 	// Start the server and open the database
 	eventlog := timeseries.New(new(noopMembership), monitor, store, timeseriesCfg)
@@ -147,7 +148,7 @@ int1: int64
 	}
 
 	monitor := monitor2.NewNoop()
-	store := disk.Open(dir, timeseriesCfg.Name, monitor)
+	store := disk.Open(dir, timeseriesCfg.Name, monitor, config.Badger{})
 
 	// Start the server and open the database
 	eventlog := timeseries.New(new(noopMembership), monitor, store, timeseriesCfg)
@@ -157,12 +158,11 @@ int1: int64
 	assert.Nil(t, err)
 	expectedSchema := typeof.Schema{
 		"string1": typeof.String,
-		"int1": typeof.Int64,
+		"int1":    typeof.Int64,
 	}
 	assert.Equal(t, expectedSchema, actualSchema)
 }
 
-
 func newSplitQuery(eventName, colName string) *presto.PrestoThriftTupleDomain {
 	return &presto.PrestoThriftTupleDomain{
 		Domains: map[string]*presto.PrestoThriftDomain{