Paul Ramsey <pramsey@refractions.net> writes:
> ... So the operational benefit of adding the complexity of a
> pre-fork system is not very high.
In particular, most of the connection startup overhead work cannot be
performed until we've identified which database to connect to (since
it largely consists of bootstrapping access to the system catalogs
in that database). If you want that work to be done in advance of
receiving a client connection request, life is much more complicated
than it would be for something like Apache.
There is considerable discussion of this point in the pgsql-hackers
archives. I'd suggest reading the past threads before trying to do
anything yourself.
regards, tom lane