In the current oplog rollover check, we check to see if the oplog has rolled over (the latest timestamp that existed when mongodump started is still in the oplog collection) then dump the new oplog entries, then check again to make sure the oplog didn't roll over during the oplog dump.

We can make this cleaner and more correct by checking once and then dumping the collection while holding on to the oldest oplog entry, checking that entry at the end to make sure it is the cutoff.