Meta Service Configuration

This document describes all configuration items related to Meta service (Metadata Service/Placement Center). For general configuration information, please refer to COMMON.md.

Meta Service Overview

Meta service (also known as Placement Center) is RobustMQ's metadata management service, responsible for:

• Cluster node management
• Metadata storage and synchronization
• Service discovery
• Configuration management

---

Meta Runtime Configuration

Runtime Configuration

[place.runtime]
heartbeat_timeout_ms = 30000      # Heartbeat timeout (milliseconds)
heartbeat_check_time_ms = 1000    # Heartbeat check interval (milliseconds)

Configuration Description

Configuration	Type	Default	Description
heartbeat_timeout_ms	u64	30000	Node heartbeat timeout (milliseconds)
heartbeat_check_time_ms	u64	1000	Heartbeat check interval (milliseconds)

Heartbeat Mechanism Description

• Heartbeat Timeout: Nodes that don't send heartbeat within specified time will be marked as unavailable
• Heartbeat Check: Periodically check heartbeat status of all nodes
• Failover: Automatically elect new primary node when primary node fails

---

Meta Storage Configuration

RocksDB Storage Configuration

[rocksdb]
data_path = "./data/placement-center/data"  # Data storage path
max_open_files = 10000                      # Maximum open files

Configuration Description

Configuration	Type	Default	Description
data_path	string	"./data"	Meta service data storage directory
max_open_files	i32	10000	RocksDB maximum simultaneously open files

---

Meta Cluster Configuration

Cluster Node Configuration

# Metadata center node address configuration
[placement_center]
1 = "127.0.0.1:1228"
2 = "127.0.0.1:1229"  
3 = "127.0.0.1:1230"

Configuration Description

Configuration	Type	Description
placement_center	table	Metadata center node address mapping table
Key	string	Node ID (usually corresponds to broker_id)
Value	string	Node address in format "IP:port"

Cluster Deployment Modes

1. Single Node Mode: Configure only one node, suitable for testing and development
2. Multi-Node Mode: Configure multiple nodes for high availability
3. Recommended Configuration: Production environment recommends odd number of nodes (3, 5, 7)

---

Meta Service Role Configuration

Role Configuration

roles = ["meta", "broker"]    # Node role list

Role Description

• meta: Metadata service role
• broker: MQTT Broker role
• journal: Journal engine role

Deployment Modes

1. Integrated Deployment: roles = ["meta", "broker", "journal"]
2. Separated Deployment:
   • Meta nodes: roles = ["meta"]
   • Broker nodes: roles = ["broker"]
   • Journal nodes: roles = ["journal"]

---

Complete Meta Service Configuration Examples

Single Node Configuration

# Basic configuration
cluster_name = "dev-cluster"
broker_id = 1
roles = ["meta", "broker"]
grpc_port = 1228

# Metadata center configuration
[placement_center]
1 = "127.0.0.1:1228"

# Meta runtime configuration
[place.runtime]
heartbeat_timeout_ms = 10000
heartbeat_check_time_ms = 2000

# Storage configuration
[rocksdb]
data_path = "./data/meta"
max_open_files = 5000

High Availability Cluster Configuration

# Basic configuration
cluster_name = "production-cluster"
broker_id = 1
roles = ["meta"]
grpc_port = 1228

# Metadata center cluster configuration
[placement_center]
1 = "192.168.1.10:1228"
2 = "192.168.1.11:1228"
3 = "192.168.1.12:1228"

# Meta runtime configuration
[place.runtime]
heartbeat_timeout_ms = 30000
heartbeat_check_time_ms = 5000

# Storage configuration
[rocksdb]
data_path = "/data/robustmq/meta"
max_open_files = 20000

---

Environment Variable Override Examples

Meta Related Environment Variables

# Cluster basic configuration
export ROBUSTMQ_CLUSTER_NAME="production-cluster"
export ROBUSTMQ_BROKER_ID=1
export ROBUSTMQ_GRPC_PORT=1228

# Meta runtime configuration
export ROBUSTMQ_PLACE_RUNTIME_HEARTBEAT_TIMEOUT_MS=30000
export ROBUSTMQ_PLACE_RUNTIME_HEARTBEAT_CHECK_TIME_MS=5000

# Storage configuration
export ROBUSTMQ_ROCKSDB_DATA_PATH="/data/robustmq/meta"
export ROBUSTMQ_ROCKSDB_MAX_OPEN_FILES=20000

---

Performance Tuning Recommendations

High Concurrency Scenarios

[place.runtime]
heartbeat_timeout_ms = 15000      # Reduce heartbeat timeout
heartbeat_check_time_ms = 3000    # Increase heartbeat check frequency

[rocksdb]
max_open_files = 50000           # Increase file handles

[network]
handler_thread_num = 16          # Increase handler threads
queue_size = 10000              # Increase queue size

Low Latency Scenarios

[place.runtime]
heartbeat_check_time_ms = 500    # More frequent heartbeat checks

[network]
lock_max_try_mut_times = 10      # Reduce lock retry times
lock_try_mut_sleep_time_ms = 5   # Reduce lock sleep time

High Reliability Scenarios

[place.runtime]
heartbeat_timeout_ms = 60000     # Increase heartbeat tolerance
heartbeat_check_time_ms = 10000  # Moderate check frequency

[rocksdb]
data_path = "/data/robustmq/meta"  # Use persistent storage path
max_open_files = 30000             # Sufficient file handles

---

Cluster Management

Node Discovery

Meta service uses placement_center configuration for node discovery:

1. Each node connects to all configured Meta nodes on startup
2. Maintain node status through heartbeat mechanism
3. Automatically handle node failures and recovery

Data Consistency

• Use Raft protocol to ensure data consistency
• Support multi-replica data storage
• Automatically handle split-brain and network partitions

Failure Recovery

• Automatically detect node failures
• Support dynamic node joining and leaving
• Automatic data synchronization and recovery

---

Troubleshooting

Common Issues

1. Node Cannot Join Cluster

   • Check placement_center configuration
   • Verify network connectivity
   • Confirm ports are not occupied

2. Heartbeat Timeout

   • Adjust heartbeat_timeout_ms
   • Check network latency
   • Optimize system load

3. Data Storage Issues

   • Check data_path directory permissions
   • Verify disk space
   • Adjust max_open_files

4. Election Failure

   • Ensure odd number of cluster nodes
   • Check node time synchronization
   • Verify network partition situations

Debugging Methods

# Enable verbose logging
[log]
log_config = "./config/debug-tracing.toml"

# Shorten heartbeat interval for debugging
[place.runtime]
heartbeat_check_time_ms = 1000

---

Monitoring Metrics

Meta service provides the following monitoring metrics:

• Node health status
• Heartbeat latency statistics
• Data synchronization status
• Election status changes
• Storage usage

---

Documentation Version: v1.0
Last Updated: 2024-01-01
Based on Code Version: RobustMQ v0.1.31