APFSDS Raft - Distributed consensus using async-raft

Implements Raft consensus for connection state synchronization.

apfsds-raft

Raft consensus implementation for APFSDS distributed cluster.

Features

• async-raft Integration: Built on the async-raft crate
• Persistent Storage: WAL-backed log with HardState persistence
• ClickHouse Backup: Async export of committed entries
• Dynamic Membership: Add/remove nodes at runtime

Usage

use apfsds_raft::{RaftNode, RaftConfig};

let config = RaftConfig {
    node_id: 1,
    peers: vec!["192.168.1.2:25347", "192.168.1.3:25347"],
    data_dir: "/var/lib/apfsds/raft".into(),
};

let node = RaftNode::new(config).await?;

// Propose a command (leader only)
node.propose(command).await?;

// Query cluster state
let leader = node.current_leader().await;

Cluster Setup

[raft]
node_id = 1
peers = ["192.168.1.2:25347", "192.168.1.3:25347"]
data_dir = "/var/lib/apfsds/raft"

License

MIT