"Matthew T. O'Connor" <matthew@zeut.net> writes:
> Another quick thought for handling FSM contention problems. A backend could
> give up waiting for access to the FSM after a short period of time, and just
> append it's data to the end of the file the same way it's done now. Dunno
> if that is feasable but it seemed like an idea to me.
Mmm ... maybe, but I doubt it'd help much. Appending a page to the file
requires grabbing some kind of lock anyway (since you can't have two
backends doing it at the same instant). With any luck, that locking can
be merged with the locking involved in accessing the FSM.
regards, tom lane