> It sounds like the behavior change would be desirable or at least > neutral. I will have to try it out and see if the refactoring is a net > improvement or turns into a mess.
I think this is a good operational improvement, particularly if someone wants to change work_mem in a pinch, and the only option now they have it to somehow get the application to re-prepare; deallocating all prepared statements or reconnecting. This is even worse with extended query protocol prepared statements in which there is no visibility in pg_prepared_statements. So one may be forced to use DEALLOCATE ALL. However, I think any GUC that can influence the planner should be considered for consistency in behavior. It was mentioned above with the enable_* GUCs, but another one I can think of is the max_parallel_workers_per_gather which should then force a re-plan if changed. I have seen users need to turn that off in a hurry when it impacts their oltp workload. -- Sami Imseih Amazon Web Services (AWS)
