mongos mapReduce cleanup doesn't obey request's writeConcern

https://github.com/mongodb/mongo/blob/r4.1.5/src/mongo/s/commands/cluster_map_reduce_cmd.cpp#L731

Note: this might be a minor issue in the real world, but it causes some issue in transaction testing because of 2 phase drops. The drop can acquire the database X lock later after the mapReduce has already finished and conflict with the transactions trying to get the same lock in the tests.