# Storage jobs, cancellation, and recovery

Storage transfers are tracked as jobs with saved selections, progress, and outcome receipts. A submitted job can wait for a worker before starting; admission can take up to about a minute. Keep the distinction between accepted, running, and completed work clear.

## Read the result by item

A job can succeed for some items and fail or retain others. Inspect the receipt to find what was copied, moved, skipped, retained, or refused. Reasons such as a changed source revision or an unsafe folder cleanup require a different response from a temporary provider failure.

A partial outcome does not roll back every completed copy. Before retrying, inspect the destination to avoid producing unnecessary duplicates or overwriting a successful result.

## Cancel

Cancellation stops subsequent work. Files already copied or moved stay in their completed state. A cancellation request is therefore not an undo command. Review both source and destination after cancellation, particularly when the job had already started deleting verified sources.

## Retry and uncertainty

A retry can resume an eligible operation using its saved progress. An interrupted App import can finalize verified snapshots without creating a second ingestion event. Follow the existing job's recovery path rather than starting a new selection immediately.

An ambiguous mutation is different: the provider may have changed the item even though Strive did not receive a definite response. Inspect the current provider state and the job receipt before replaying it. Blindly repeating an uncertain move, delete, or replacement can affect the wrong later state.

If authorization expired, reconnect and test the provider before retrying. If the error concerns a revision conflict, refresh the item and make a new deliberate selection. Preserve the original receipt when investigating an incomplete transfer.