Kevin Grittner wrote:
> This seems likely to overlap the review I was soon going to do of the
> differences between pg_ctl behavior and what is required for LSB
> conformance. I'll make sure to test this behavior along with others.
> One of my current complaints is that pg_ctl doesn't wait until it is
> actually ready to receive connections before returning an indication
> of success.
Maybe write the file as postmaster.ports.starting or some such and
rename it to its final name when recovery has finished?
--
Alvaro Herrera http://www.CommandPrompt.com/
The PostgreSQL Company - Command Prompt, Inc.