Struct WalShipper 

pub struct WalShipper { /* private fields */ }

WAL Shipper manages the replication TCP server and follower connections.

Implementations

impl WalShipper

pub fn new() -> (Self, Sender<WALEntry>)

Create a new WAL shipper. Returns the shipper and a broadcast sender that should be given to the WAL for publishing new entries.

pub fn set_replication_mode( &mut self, mode: ReplicationMode, ack_timeout: Duration, )

Set the replication mode and ACK timeout.

pub fn replication_mode(&self) -> ReplicationMode

Get the current replication mode.

pub fn current_leader_offset(&self) -> u64

Get the current leader WAL offset.

pub async fn wait_for_ack(&self, target_offset: u64) -> bool

Wait for follower ACK(s) up to the given WAL offset.

Behavior depends on replication mode:

• Async: returns immediately.
• SemiSync: waits until at least 1 follower has ACKed target_offset, or the timeout expires (returns true on success, false on timeout).
• Sync: waits until ALL connected followers have ACKed target_offset, or the timeout expires.

pub fn set_metrics(&mut self, metrics: Arc<MetricsRegistry>)

Set the Prometheus metrics registry for replication metrics.

pub fn set_store(&mut self, store: Arc<EventStore>)

Attach the EventStore reference for catch-up protocol support.

When set, the shipper can stream Parquet snapshot files to followers that are too far behind for WAL-only catch-up. Must be called before serve().

pub fn status(&self) -> ReplicationStatus

Get current replication status for the health endpoint.

pub async fn serve(self: Arc<Self>, port: u16) -> Result<()>

Start the replication TCP server. This runs until the process shuts down.