Re: Experimental ARC implementation

Bruce Momjian <pgman@candle.pha.pa.us> writes:
> Have you considered having the background writer check the pages it is
> about to write to see if they can be added to the FSM, thereby reducing
> the need for vacuum?

The 7.4 rewrite of FSM depends on the assumption that all the free space
in a given relation is reported to FSM in one batch (ie, at the end of a
VACUUM pass).  This solves problems in both speed (page-at-a-time update
of FSM was horrendously expensive) and space allocation policy (we now
use the number of "interesting" pages in each relation as input
information for the allocation policy).  To do anything like the above,
you'd need to find different solutions to these problems.
        regards, tom lane