"Kevin Grittner" <Kevin.Grittner@wicourts.gov> writes:
> Peter Eisentraut wrote:
>> The patch is pretty straightforward,
>> but does anyone else actually want this? Comments?
> I agree that the initdb name seems odd next to the other executable
> names, but the functionality seems a little out of place to me in
> pg_ctl. The other options all correspond (more or less) to LSB init
> script actions (and we've been talking about the desirability of
> making that a closer fit); while this is something which would *not
> be appropriate* in an init script.
Well, it's not appropriate or safe as a default action, but there
already is a nonstandard "service postgresql init" action in at least
the PGDG and Red Hat init scripts. In fact, I believe that Zdenek's
entire rationale for this is predicated on the assumption that he can
eventually make initdb's disappearance transparent, if he can get
people used to using such a thing instead of initdb'ing by hand.
regards, tom lane