fix #287. make the lru cache size configurable

config.toml.sample

@@ -74,6 +74,11 @@ query-shard-buffer-size = 1000
 # the process
 max-open-files = 40
 
+# LRU cache size, LRU is used by leveldb to store contents of the
+# uncompressed sstables. You can use `m` or `g` prefix for megabytes
+# and gigabytes, respectively.
+lru-cache-size = "200m"
+
 # These options specify how data is sharded across the cluster. There are two 
 # shard configurations that have the same knobs: short term and long term.
 # Any series that begins with a capital letter like Exceptions will be written

src/configuration/config.toml

@@ -70,6 +70,7 @@ query-shard-buffer-size = 1000
 # Maximum mmap open files, this will affect the virtual memory used by
 # the process
 # max-open-files = 40
+# lru-cache-size = "200m"
 
 # These options specify how data is sharded across the cluster. There are two 
 # shard configurations that have the same knobs: short term and long term.

src/configuration/configuration.go

@@ -9,9 +9,38 @@ import (
 	"io/ioutil"
 	"os"
 	"regexp"
+	"strconv"
 	"time"
 )
 
+type size struct {
+	int
+}
+
+const (
+	ONE_MEGABYTE = 1024 * 1024
+	ONE_GIGABYTE = 1024 * ONE_MEGABYTE
+)
+
+func (d *size) UnmarshalText(text []byte) error {
+	str := string(text)
+	length := len(str)
+	size, err := strconv.ParseInt(string(text[:length-1]), 10, 64)
+	if err != nil {
+		return err
+	}
+	switch suffix := text[len(text)-1]; suffix {
+	case 'm':
+		size *= ONE_MEGABYTE
+	case 'g':
+		size *= ONE_GIGABYTE
+	default:
+		return fmt.Errorf("Unknown size suffix %s", suffix)
+	}
+	d.int = int(size)
+	return nil
+}
+
 type duration struct {
 	time.Duration
 }
@@ -59,6 +88,7 @@ type LoggingConfig struct {
 
 type LevelDbConfiguration struct {
 	MaxOpenFiles int `toml:"max-open-files"`
+	LruCacheSize int `toml:"lru-cache-size"`
 }
 
 type ShardingDefinition struct {
@@ -154,6 +184,7 @@ type Configuration struct {
 	LogLevel                  string
 	BindAddress               string
 	LevelDbMaxOpenFiles       int
+	LevelDbLruCacheSize       int
 	ShortTermShard            *ShardConfiguration
 	LongTermShard             *ShardConfiguration
 	ReplicationFactor         int
@@ -226,6 +257,7 @@ func parseTomlConfiguration(filename string) (*Configuration, error) {
 		Hostname:                  tomlConfiguration.Hostname,
 		BindAddress:               tomlConfiguration.BindAddress,
 		LevelDbMaxOpenFiles:       tomlConfiguration.LevelDb.MaxOpenFiles,
+		LevelDbLruCacheSize:       tomlConfiguration.LevelDb.LruCacheSize,
 		LongTermShard:             &tomlConfiguration.Sharding.LongTerm,
 		ShortTermShard:            &tomlConfiguration.Sharding.ShortTerm,
 		ReplicationFactor:         tomlConfiguration.Sharding.ReplicationFactor,
@@ -251,6 +283,11 @@ func parseTomlConfiguration(filename string) (*Configuration, error) {
 		config.LevelDbMaxOpenFiles = 100
 	}
 
+	// if it wasn't set, set it to 100
+	if config.LevelDbLruCacheSize == 0 {
+		config.LevelDbLruCacheSize = 200 * ONE_MEGABYTE
+	}
+
 	return config, nil
 }
 

src/configuration/configuration_test.go

@@ -50,3 +50,11 @@ func (self *LoadConfigurationSuite) TestConfig(c *C) {
 	c.Assert(config.WalIndexAfterRequests, Equals, 1000)
 	c.Assert(config.WalRequestsPerLogFile, Equals, 10000)
 }
+
+func (self *LoadConfigurationSuite) TestSizeParsing(c *C) {
+	var s size
+	c.Assert(s.UnmarshalText([]byte("200m")), IsNil)
+	c.Assert(s.int, Equals, 200*ONE_MEGABYTE)
+	c.Assert(s.UnmarshalText([]byte("10g")), IsNil)
+	c.Assert(s.int, Equals, 10*ONE_GIGABYTE)
+}

src/datastore/leveldb_shard_datastore.go

@@ -69,7 +69,7 @@ func NewLevelDbShardDatastore(config *configuration.Configuration) (*LevelDbShar
 		return nil, err
 	}
 	opts := levigo.NewOptions()
-	opts.SetCache(levigo.NewLRUCache(ONE_MEGABYTE))
+	opts.SetCache(levigo.NewLRUCache(config.LevelDbLruCacheSize))
 	opts.SetCreateIfMissing(true)
 	opts.SetBlockSize(64 * ONE_KILOBYTE)
 	filter := levigo.NewBloomFilter(SHARD_BLOOM_FILTER_BITS_PER_KEY)