There doesn't seem to be a test or any comments documenting what happens when an oplog write occurs while a node is scanning the oplog and generating the initial markers (during step-up). e.g. does the write get factored into the size for the marker spacing? is the size updated later after step-up finishes?

The intent of this ticket is to just figure out what happens in this scenario. It doesn't seem particularly urgent to fix if we do not factor in this write, worst case we truncate slightly later than expected.