| From: | Magnus Hagander <magnus(at)hagander(dot)net> |
|---|---|
| To: | Alvaro Herrera <alvherre(at)commandprompt(dot)com> |
| Cc: | Dawid Kuroczko <qnex42(at)gmail(dot)com>, josh(at)agliodbs(dot)com, pgsql-hackers(at)postgresql(dot)org |
| Subject: | Re: Permanent settings |
| Date: | 2008-02-20 13:29:52 |
| Message-ID: | 20080220132952.GN17768@svr2.hagander.net |
| Views: | Raw Message | Whole Thread | Download mbox | Resend email |
| Thread: | |
| Lists: | pgsql-hackers |
On Wed, Feb 20, 2008 at 10:20:55AM -0300, Alvaro Herrera wrote:
> Magnus Hagander escribió:
> > On Wed, Feb 20, 2008 at 09:36:43AM -0300, Alvaro Herrera wrote:
>
> > > IMO restart-only settings should not be changeable via the new SQL
> > > command. It's just too messy to deal with that.
> >
> > I respectfully disagree. It should be settable. You need a restart, sure,
> > and the GUI app should tell you that. But you shuld be able to change them.
> > (for example, pgadmin can restart the server just fine for you if you're on
> > Windows)
>
> What if the server doesn't come back up? Say, because you increased
> shared_buffers and now it doesn't fit on the kernel limits. If you
> haven't made arrangements to be able to edit the postgresql.conf file
> beforehand, you're hosed.
Right. So a warning in the GUI program is important, but that's no reason
to restrict the API.
Doing DROP TABLE can also be very dangerous. Or DROP TYPE CASCADE. Yet we
do support them, and rely on the user to think first, or the GUI pogram to
show warnings.
//Magnus
| From | Date | Subject | |
|---|---|---|---|
| Next Message | Dimitri Fontaine | 2008-02-20 14:00:49 | Re: Permanent settings |
| Previous Message | Andrew Sullivan | 2008-02-20 13:27:23 | Re: Including PL/PgSQL by default |