Re: Optimizer not using index on 120M row table

"Jim C. Nasby" <jim@nasby.net> writes:
> Should effective_cache_size include the size of shared_buffers?

Yes ... although IMHO, if shared_buffers is large enough to materially
affect that number, it's too large ;-)

> FreeBSD doesn't seem to want to use more than about 300M for disk
> caching, so I currently have shared_buffers set to 90000 or about 700M
> (the box has 2G, but pgsql currently has to share with Sybase). Are
> there any issues with setting shared_buffers so high?

Plenty, see many past threads in pgsql-performance and other lists.
There are strong reasons to think that you should let the kernel do the
bulk of the caching work.

            regards, tom lane