The theory behind a sql file like this is you can run it against an empty database and get an equivalent production schema. We have inserted new lines out of order, and branch names in the file don't represent branch names in the database. Staging is out of sync with production, as well. I think we have to start over. Ideally we retire graphserver before we have to solve this, but I'm not sure how reliable that would be. One way to solve this without a ton of human intervention would be to dump the applicable schemas + branches/machines tables from production to a new data.sql file, and check that in. Then blow away staging graphs and copy over production, and voila, we're in sync again.

graph server is dead