Skip to main content

Complete membuss.yaml Configuration Schema

Membuss node configuration is managed via a single YAML manifest (membuss.yaml). Every setting has an explicit fallback default inside core/config/config.go.


Complete membuss.yaml Manifest

# libp2p Transport Listen Addresses
listen_addrs:
- /ip4/0.0.0.0/tcp/4001
- /ip4/0.0.0.0/udp/4001/quic-v1
- /ip4/0.0.0.0/tcp/4002/ws

# Directory Paths
data_dir: ./data
blocks_path: ./data/blocks # Optional flat file blob storage path for large objects (>= 1MB)

# Service Bindings
gateway_addr: 127.0.0.1:8080
api_addr: 127.0.0.1:5001
grpc_addr: 127.0.0.1:50051

# Subsystem Enable Toggles
servers:
gateway:
enabled: true # Mem-Gate HTTP gateway & Web Explorer UI
node_api:
enabled: true # Local REST Control API (/api/v1)
grpc:
enabled: true # Daemon gRPC service

# Multihash & Storage Engine Settings
storage:
default_hash: "blake3" # Multihash codec: "blake3" (0x1e), "sha256" (0x12), "sha512" (0x13)
bloom_capacity: 10000000 # Counting Bloom Filter expected items
bloom_fp_rate: 0.001 # Target false positive rate (0.1%)

# Persistence & Anchor Engine
anchor_mode: false # true = automatically discover and mirror announced network content
anchor:
max_storage: 100GB # Bounded storage quota (e.g. 500MB, 100GB, 1TB, 0 = unlimited)
fetch_concurrency: 8 # Parallel worker pool size for backlog mirror fetching
reacquire_batch_size: 32 # Sampling batch size per discovery round

# Modular Plugin Extensions
plugins:
enabled: true
active:
- echo-inspector
config:
echo-inspector:
log_level: info

# Observability
metrics_enabled: true # Exposes Prometheus metrics at /metrics
log_level: info

Detailed Parameter Reference

storage.default_hash

  • Type: string
  • Default: "blake3"
  • Options: "blake3", "sha256", "sha512"
  • Description: Sets the default multihash algorithm used by BuildParallel during chunk ingestion.

anchor_mode & anchor.max_storage

  • Type: boolean & string
  • Default: false & "100GB"
  • Description: Configures whether the node operates as an Anchor Node, listening to network announcements and maintaining bounded storage with automated LRU eviction for auto-mirrored content.