We've noticed on our CentOS PostgreSQL database servers that every postgres process is no longer running at the default nice priority of '0'. They are running at a lower priority (2). Is Linux lowering the priority of these processes over time, or does a change in niceness have to be caused by a more manual process (renice, from within postgres, etc.)? My assumption was the latter.
Restarting PostgreSQL returns all 'postmaster' processes to the default nice value of '0'.
This is occuring on CentOS 5 with PostgreSQL 9.0