Is it possible to reuse the same transaction between multiple processes with Postgres? Or any other RDBMS? Is there name for such concept and does it make any sense?
One possible use case would be two service instances (different processes) modifying related set of entities within one distributed transaction.
For example, imagine a website that displays music playlists. Plain users can see playlists, but editors can collaborate on playlist editing. Editors would modify lists, but viewers wouldn't see latest changes until one of editors saves (commits) changes.
This is just illustrative example, and would take a different approach to the design, without relying on distributed transactions. I'm just interested in the concept of distributed transactions and it's plausability, since I'm not expert in the field.