Monitor reference table copy progress during master_add_node

When there are large reference tables, master_add_node can take a long time and it's a black box. We should have some way of monitoring its progress from the coordinator.

The only named entry point is master_add_node; trace how reference-table copies are initiated and what coordinator-visible state exists. Define the progress signal and its completion semantics before identifying the implementation and tests; the issue names no files or tests.

Written by the indexing model from the issue text.