Crate atomr_cluster_sharding 

atomr-cluster-sharding.

Structs

CoordinatorState

Eventsourced coordinator state — kept separate from ShardCoordinator so callers can rebuild it from journal replay. The in-memory ShardCoordinator is the local projection; this struct mirrors it through the persistence layer.

DDataShardCoordinator

DData-backed allocation coordinator.

EntityRef

HandoffCoordinator

Per-shard handoff state machine.

InMemoryRememberStore

In-memory store — for tests and as a reference implementation.

LeastShardAllocationStrategy

Place new shards on the region with the fewest shards, breaking ties lexicographically. Rebalances if the difference between most- and least-loaded regions exceeds rebalance_threshold.

PassivationTracker

PersistentShardCoordinator

Wraps a ShardCoordinator with Eventsourced plumbing. Use recover on boot, then command for every allocation / rebalance / removal.

PinnedAllocationStrategy

Pin every shard to a specific region (useful for tests / static allocation). analog: a custom strategy returning a constant region.

RebalanceRunner

RememberedEntities

In-process registry of remembered entity ids. Wraps a RememberEntitiesStore and serves quick lookups from a local snapshot.

Shard

ShardCoordinator

ShardRegion

Enums

CoordinatorCommand

CoordinatorError

CoordinatorEvent

HandoffError

HandoffState

RebalanceAction

Action emitted by RebalanceRunner::step.

RememberError

Traits

MessageExtractor

RememberEntitiesStore

Pluggable persistence store for remembered entities.

ShardAllocationStrategy

Pluggable shard allocation policy.

Functions

project_into

Project a CoordinatorState (rebuilt from journal replay) onto a fresh ShardCoordinator. Useful right after recover.