Using MarkLogic Content Pump (mlcp)

When the input type is archive, mlcp is not able to retry loading metadata/naked properties when commit fails, since by design the client does not cache these inputs.

Loading temporal documents may have issues. When mlcp commit fails and catches exceptions, it tries rolling back before retry loading the whole batch. However, the previous transaction may have made it to the server and mlcp will get NOTXN exception. This may create issues for temporal documents, since they may be inserted multiple times.