In case if one Distributed has multiple shards, and underlying Distributed has only one, there can be the case when the query will be tried to process from Complete to WithMergeableStateAfterAggregation, which is obviously wrong.