Skip to main content

run_retained_execution_workflow

Function run_retained_execution_workflow 

Source
pub async fn run_retained_execution_workflow<D: ExecutionWorkflowStageDriver>(
    layout: &BackupLayoutGuard,
    expected_workflow: &ArtifactChecksumRecord,
    retained: &BTreeMap<u64, RetainedStageEvidence>,
    driver: &mut D,
) -> Result<ExecutionWorkflowSettlementView, ExecutionWorkflowRunError<D::Error>>
Expand description

Continue an existing allocation after replaying every occupied settled stage.

The caller supplies the complete occupied stage set and authenticates each stage’s learned evidence. Before any driver callback, the coordinator checks that set against the held layout and replays every exact all-Applied original checkpoint. Already bound successor edges must commit the qualified digest supplied for that exact edge; absent overrides use the stage checkpoint’s learned digest. Pending, partially prepared or omitted occupied stages refuse; none is reissued. Only absent stages may be prepared and executed once, with fresh authority and external-effect admission remaining with the driver. This is no terminal receipt, same-ID restore or reference-release permit.

§Errors

Rejects incomplete or changed retained evidence before new driver callbacks; later stage failures preserve every original journal and effect obligation.