Struct ReplicaManager 

pub struct ReplicaManager { /* private fields */ }

Central manager for shard-level replica tracking and failover coordination.

Implementations

impl ReplicaManager

pub fn new(replication_factor: usize) -> Self

Create a new manager with the given target replication_factor.

replication_factor must be at least 1 (1 = primary only, no replicas).

pub fn register_replica( &mut self, replica: ShardReplica, ) -> Result<(), VectorError>

Register a new replica.

Errors

Returns VectorError::InvalidData when:

• A replica with the same (shard_id, replica_id) already exists.
• replica.state is Primary but a primary is already registered for that shard.

pub fn unregister_replica(&mut self, shard_id: u64, replica_id: &str) -> bool

Unregister a replica by (shard_id, replica_id).

Returns true if the replica was found and removed.

pub fn promote_to_primary( &mut self, shard_id: u64, replica_id: &str, ) -> Result<(), VectorError>

Promote replica_id to Primary for shard_id.

Any existing primary is demoted to Replica first.

Errors

Returns VectorError::InvalidData when the target replica is not found or is in the Failed state.

pub fn mark_failed(&mut self, shard_id: u64, replica_id: &str)

Mark a replica as Failed.

If the failed replica was the primary, no automatic failover occurs here — call promote_to_primary on a healthy replica to elect a new one.

A no-op when the (shard_id, replica_id) pair does not exist.

pub fn auto_failover(&mut self, shard_id: u64) -> Result<String, VectorError>

Attempt automatic failover for shard_id when the current primary fails.

Selects the healthy replica with the largest vector_count (most up-to-date) and promotes it. Returns the promoted replica’s ID on success.

Errors

Returns VectorError::InvalidData when no healthy replica exists for the shard.

pub fn update_sync_progress( &mut self, shard_id: u64, replica_id: &str, progress: f64, )

Update the progress of a CatchingUp replica.

When progress reaches 1.0, the replica is automatically promoted to Replica.

pub fn get_primary(&self, shard_id: u64) -> Option<&ShardReplica>

Return the primary replica for shard_id, if any.

pub fn get_replicas(&self, shard_id: u64) -> Vec<&ShardReplica>

Return all replicas for shard_id (including primary).

pub fn get_healthy_replicas(&self, shard_id: u64) -> Vec<&ShardReplica>

Return all healthy (non-failed) replicas for shard_id.

pub fn shard_ids(&self) -> Vec<u64>

Return all tracked shard IDs.

pub fn replication_factor(&self) -> usize

The configured replication factor.

pub fn needs_rebalancing(&self) -> bool

Returns true when any shard is under- or over-replicated relative to the configured replication_factor.

pub fn replication_status(&self) -> ReplicationStatus

Produce a ReplicationStatus summary of the cluster’s health.