Teodor Sigaev wrote:
>> current default of 1000 pages (8Mb) seems really pretty silly for modern
>> machines; we could certainly set it to 10 times that without problems,
>> and maybe much more. Thoughts?
>
> May be, set by default effective_cache_size equal to number of shared
> buffers?
> If pgsql is configured to use quarter or half of total memory for
> shared buffer, then effective_cache_size will have good approximation...
>
>
Initdb does not currently make any attempt to discover the extent of
physical or virtual memory, it simply tries to start postgres with
certain shared_buffer settings, starting at 4000, and going down until
we get a success.
max_fsm_pages is now fixed proportionally with shared_buffers, and I
guess we could do something similar with effective_cache_size, but since
IIRC this doesn't involve shared memory I'm inclined to agree with Tom
that it should just be fixed at some substantially higher level.
cheers
andrew