Currently, the test expects the recipients to get to steady-state before the abort command is sent. However, it is possible the coordinator document can exist before the reshardingFields.recipientShards have been added when the coordinator transitions to kPreparingToDonate. Thus, the executeAtStartOfReshardingFn in the test case may return early - causing unintended behavior.

A potential fix would be to return false if the coordinator exists but the coordinatorDoc.recipientShards do not before attempting to check each recipient shard is in steady-state.