Lux Docs

Default Options

opts := badger.DefaultOptions("/path/to/data")

This provides sensible defaults for most use cases. Override individual fields as needed.

Key Options

Option	Default	Description
`Dir`	(required)	Directory for LSM tree (keys + metadata)
`ValueDir`	same as Dir	Directory for value log (can be separate SSD)
`MemTableSize`	64MB	Size of each in-memory table
`NumMemtables`	5	Number of memtables to keep
`NumLevelZeroTables`	5	Tables in level 0 before compaction
`ValueLogFileSize`	1GB	Size of each value log file
`ValueThreshold`	1MB	Values larger than this go to value log
`NumCompactors`	4	Number of concurrent compaction goroutines
`Compression`	Snappy	Compression algorithm (None, Snappy, ZSTD)
`DetectConflicts`	true	Enable transaction conflict detection

Encryption at Rest

opts := badger.DefaultOptions("/path/to/data")

// 16 or 32 byte key for AES-128 or AES-256
encryptionKey := []byte("0123456789abcdef0123456789abcdef")
opts.EncryptionKey = encryptionKey
opts.EncryptionKeyRotationDuration = 10 * 24 * time.Hour

db, err := badger.Open(opts)

When encryption is enabled, all data on disk (LSM tree and value log) is encrypted. The encryption key must be provided on every Open call.

In-Memory Mode

opts := badger.DefaultOptions("").WithInMemory(true)
db, err := badger.Open(opts)

Useful for testing or ephemeral caches. No data is persisted to disk.

Performance Tuning

For write-heavy workloads (blockchain node sync):

opts.NumMemtables = 10
opts.NumLevelZeroTables = 10
opts.NumLevelZeroTablesStall = 20
opts.NumCompactors = 8
opts.Compression = options.ZSTD

For read-heavy workloads (indexer queries):

opts.NumMemtables = 3
opts.BlockCacheSize = 256 << 20 // 256MB block cache
opts.IndexCacheSize = 128 << 20 // 128MB index cache

Platform Support

Platform	Locking	Notes
Linux/macOS/BSD	POSIX flock	Full support
Windows	LockFileEx	Full support
AIX	No fsync	Reduced durability guarantees
Plan9/WASM	No locking	Single-process only

Configuration