MerkleDB -- add eviction batch size config (ava-labs#1586)

Author: Dan Laine

x/merkledb/db.go

@@ -28,8 +28,7 @@ import (
 )
 
 const (
-	RootPath          = EmptyPath
-	evictionBatchSize = 100
+	RootPath = EmptyPath
 	// TODO: name better
 	rebuildViewSizeFractionOfCacheSize = 50
 	minRebuildViewSizePerCommit        = 1000
@@ -116,6 +115,9 @@ type MerkleDB interface {
 }
 
 type Config struct {
+	// The number of nodes that are evicted from the cache and written to
+	// disk at a time.
+	EvictionBatchSize int
 	// The number of changes to the database that we store in memory in order to
 	// serve change proofs.
 	HistoryLength int
@@ -144,8 +146,9 @@ type merkleDB struct {
 	metadataDB database.Database
 
 	// If a value is nil, the corresponding key isn't in the trie.
-	nodeCache     onEvictCache[path, *node]
-	onEvictionErr utils.Atomic[error]
+	nodeCache         onEvictCache[path, *node]
+	onEvictionErr     utils.Atomic[error]
+	evictionBatchSize int
 
 	// Stores change lists. Used to serve change proofs and construct
 	// historical views of the trie.
@@ -172,12 +175,13 @@ func newDatabase(
 	metrics merkleMetrics,
 ) (*merkleDB, error) {
 	trieDB := &merkleDB{
-		metrics:    metrics,
-		nodeDB:     prefixdb.New(nodePrefix, db),
-		metadataDB: prefixdb.New(metadataPrefix, db),
-		history:    newTrieHistory(config.HistoryLength),
-		tracer:     config.Tracer,
-		childViews: make([]*trieView, 0, defaultPreallocationSize),
+		metrics:           metrics,
+		nodeDB:            prefixdb.New(nodePrefix, db),
+		metadataDB:        prefixdb.New(metadataPrefix, db),
+		history:           newTrieHistory(config.HistoryLength),
+		tracer:            config.Tracer,
+		childViews:        make([]*trieView, 0, defaultPreallocationSize),
+		evictionBatchSize: config.EvictionBatchSize,
 	}
 
 	// Note: trieDB.OnEviction is responsible for writing intermediary nodes to
@@ -757,7 +761,7 @@ func (db *merkleDB) onEviction(n *node) error {
 	// just [n], so that we don't immediately evict and write another
 	// node, because each time this method is called we do a disk write.
 	var err error
-	for removedCount := 0; removedCount < evictionBatchSize; removedCount++ {
+	for removedCount := 0; removedCount < db.evictionBatchSize; removedCount++ {
 		_, n, exists := db.nodeCache.removeOldest()
 		if !exists {
 			// The cache is empty.

x/merkledb/db_test.go

@@ -10,6 +10,7 @@ import (
 	"strconv"
 	"testing"
 
+	"github.com/prometheus/client_golang/prometheus"
 	"github.com/stretchr/testify/require"
 
 	"github.com/ava-labs/avalanchego/database"
@@ -26,6 +27,16 @@ func newNoopTracer() trace.Tracer {
 	return tracer
 }
 
+func newDefaultConfig() Config {
+	return Config{
+		EvictionBatchSize: 100,
+		HistoryLength:     100,
+		NodeCacheSize:     1_000,
+		Reg:               prometheus.NewRegistry(),
+		Tracer:            newNoopTracer(),
+	}
+}
+
 func Test_MerkleDB_Get_Safety(t *testing.T) {
 	db, err := getBasicDB()
 	require.NoError(t, err)
@@ -86,11 +97,7 @@ func Test_MerkleDB_DB_Load_Root_From_DB(t *testing.T) {
 	db, err := New(
 		context.Background(),
 		rdb,
-		Config{
-			Tracer:        newNoopTracer(),
-			HistoryLength: 100,
-			NodeCacheSize: 100,
-		},
+		newDefaultConfig(),
 	)
 	require.NoError(err)
 
@@ -112,11 +119,7 @@ func Test_MerkleDB_DB_Load_Root_From_DB(t *testing.T) {
 	db, err = New(
 		context.Background(),
 		rdb,
-		Config{
-			Tracer:        newNoopTracer(),
-			HistoryLength: 100,
-			NodeCacheSize: 100,
-		},
+		newDefaultConfig(),
 	)
 	require.NoError(err)
 	reloadedRoot, err := db.GetMerkleRoot(context.Background())
@@ -132,14 +135,13 @@ func Test_MerkleDB_DB_Rebuild(t *testing.T) {
 
 	initialSize := 10_000
 
+	config := newDefaultConfig()
+	config.NodeCacheSize = initialSize
+
 	db, err := New(
 		context.Background(),
 		rdb,
-		Config{
-			Tracer:        newNoopTracer(),
-			HistoryLength: 100,
-			NodeCacheSize: initialSize,
-		},
+		config,
 	)
 	require.NoError(err)
 
@@ -167,10 +169,7 @@ func Test_MerkleDB_Failed_Batch_Commit(t *testing.T) {
 	db, err := New(
 		context.Background(),
 		memDB,
-		Config{
-			Tracer:        newNoopTracer(),
-			HistoryLength: 300,
-		},
+		newDefaultConfig(),
 	)
 	require.NoError(t, err)
 
@@ -190,11 +189,7 @@ func Test_MerkleDB_Value_Cache(t *testing.T) {
 	db, err := New(
 		context.Background(),
 		memDB,
-		Config{
-			Tracer:        newNoopTracer(),
-			HistoryLength: 300,
-			NodeCacheSize: minCacheSize,
-		},
+		newDefaultConfig(),
 	)
 	require.NoError(t, err)
 
@@ -815,11 +810,7 @@ func runRandDBTest(require *require.Assertions, r *rand.Rand, rt randTest) {
 			dbTrie, err := newDatabase(
 				context.Background(),
 				memdb.New(),
-				Config{
-					Tracer:        newNoopTracer(),
-					HistoryLength: 0,
-					NodeCacheSize: minCacheSize,
-				},
+				newDefaultConfig(),
 				&mockMetrics{},
 			)
 			require.NoError(err)

x/sync/client_test.go

@@ -11,6 +11,7 @@ import (
 	"time"
 
 	"github.com/golang/mock/gomock"
+	"github.com/prometheus/client_golang/prometheus"
 
 	"github.com/stretchr/testify/require"
 
@@ -27,6 +28,16 @@ import (
 	syncpb "github.com/ava-labs/avalanchego/proto/pb/sync"
 )
 
+func newDefaultDBConfig() merkledb.Config {
+	return merkledb.Config{
+		EvictionBatchSize: 100,
+		HistoryLength:     defaultRequestKeyLimit,
+		NodeCacheSize:     defaultRequestKeyLimit,
+		Reg:               prometheus.NewRegistry(),
+		Tracer:            newNoopTracer(),
+	}
+}
+
 func sendRangeRequest(
 	t *testing.T,
 	db SyncableDB,
@@ -377,21 +388,14 @@ func TestGetChangeProof(t *testing.T) {
 	trieDB, err := merkledb.New(
 		context.Background(),
 		memdb.New(),
-		merkledb.Config{
-			Tracer:        newNoopTracer(),
-			HistoryLength: defaultRequestKeyLimit,
-			NodeCacheSize: defaultRequestKeyLimit,
-		},
+		newDefaultDBConfig(),
 	)
 	require.NoError(t, err)
+
 	verificationDB, err := merkledb.New(
 		context.Background(),
 		memdb.New(),
-		merkledb.Config{
-			Tracer:        newNoopTracer(),
-			HistoryLength: defaultRequestKeyLimit,
-			NodeCacheSize: defaultRequestKeyLimit,
-		},
+		newDefaultDBConfig(),
 	)
 	require.NoError(t, err)
 	startRoot, err := trieDB.GetMerkleRoot(context.Background())