Book IDs are not safe with concurrency #16
Description
The book creation process is not atomic, it first gets the new ID and then creates the book. Multiple processes could get the same ID which could lead to wrong data or data loss.
The proposed solution is to create a database table for books with a sub-table for pages. Creating a book can be done in a single operation returning the ID from an auto-increment.
What is required for this to work:
- Migration of current book structure.
- Updating the current code logic to implement this new structure.