Module recovery 

Distributed recovery manager. Distributed recovery manager.

Complements the single-node RecoveryManager in laminar-db with distributed checkpoint discovery, incremental delta application, source seeking, and integrity verification.

Recovery sequence

1. Discover: Try read_latest(), fall back to list_checkpoints() (C9: dual-source discovery catches partial Raft commit failures)
2. Load manifest: load_manifest(id)
3. Validate: Check operator names match the current deployment
4. Restore state: For each operator partition — full snapshot or incremental delta (C1: incremental support)
5. Seek sources: Reset source offsets to checkpoint positions
6. Return RecoveryResult

Structs

RecoveryConfig

Configuration for the distributed recovery manager.

RecoveryManager

Distributed recovery manager.

RecoveryResult

Successful recovery result.

Enums

RecoveryError

Errors from the recovery process.

Traits

Restorable

Trait for operator state that can be restored from a checkpoint.

Seekable

Trait for source connectors that can seek to a checkpoint offset.

TypedSeekable

Trait for source connectors that can seek using a typed SourcePosition.