Struct ReplicationGroup 

pub struct ReplicationGroup { /* private fields */ }

A simulated Raft cluster of TsdbRaftNodes.

All message delivery, election timeout tracking, and commit-index advancement are driven by calls to ReplicationGroup::tick.

Implementations

impl ReplicationGroup

pub fn new(ids: &[&str], timeout_override: Option<&[u32]>) -> Self

Create a group with the given node IDs and (optionally distinct) election timeouts.

Pass timeout_override = None to assign uniform timeouts (10 ticks). Pass a slice of per-node timeouts to make elections deterministic.

pub fn cluster_size(&self) -> usize

Return the number of nodes in the group.

pub fn node(&self, id: &str) -> Option<&TsdbRaftNode>

Return a reference to a node by ID.

pub fn node_mut(&mut self, id: &str) -> Option<&mut TsdbRaftNode>

Return a mutable reference to a node by ID.

pub fn leader_id(&self) -> Option<String>

Find the current leader, if any.

Returns None if no node is in the Leader role.

pub fn leader_count(&self) -> usize

Return the number of nodes currently acting as leader.

In a healthy cluster this should always be <= 1.

pub fn quorum_size(&self) -> usize

Return the quorum size (majority) for this cluster.

pub fn partition(&mut self, id: &str)

Partition a node (simulates network isolation).

pub fn heal(&mut self, id: &str)

Heal a partitioned node.

pub fn propose(&mut self, cmd: TsdbCommand) -> TsdbResult<(String, u64)>

Propose a write command on the current leader.

Returns (leader_id, log_index) or an error if there is no leader.

pub fn propose_write_entry( &mut self, entry: &WriteEntry, ) -> TsdbResult<(String, u64)>

Propose a WriteEntry via the current leader (convenience wrapper).

Returns (leader_id, log_index) or an error.

pub fn propose_and_commit( &mut self, entry: &WriteEntry, max_ticks: u32, ) -> TsdbResult<(String, u64)>

Propose a WriteEntry and wait for quorum commit by ticking.

Returns (leader_id, log_index) after up to max_ticks ticks. Returns an error if the entry is not committed within max_ticks.

pub fn tick(&mut self)

Advance the simulation by one logical tick.

Each tick:

1. Sends leader heartbeats first (so they enter the bus).
2. Delivers all pending messages from the bus.
3. Increments election timers on followers/candidates.
4. Advances commit indices on the leader.

The order matters: heartbeats are sent first so they can be delivered in the same tick, preventing unnecessary election timeouts.

pub fn tick_n(&mut self, n: u32)

Run n ticks in sequence.

pub fn node_ids(&self) -> Vec<String>

Return all node IDs in the cluster.

pub fn pending_messages(&self) -> usize

Return the number of pending messages on the bus.