Re: The missing pg_get_*def functions

From: Andres Freund <andres(at)2ndquadrant(dot)com>
To: Joel Jacobson <joel(at)trustly(dot)com>
Cc: Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us>, Noah Misch <noah(at)leadboat(dot)com>, Pg Hackers <pgsql-hackers(at)postgresql(dot)org>
Subject: Re: The missing pg_get_*def functions
Date: 2013-04-30 15:18:16
Message-ID: 20130430151816.GE25261@alap2.anarazel.de
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-hackers

On 2013-04-30 15:57:02 +0200, Andres Freund wrote:
> On 2013-04-30 05:14:15 +0100, Joel Jacobson wrote:
> > On Tue, Apr 30, 2013 at 12:58 AM, Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us> wrote:
> > > Noah Misch <noah(at)leadboat(dot)com> writes:
> > >> Note also that minor releases can readily fix bugs in C-language functions,
> > >> but we have no infrastructure to update sql-language functions after initdb.
> > >> That flexibility is unfortunate to lose, particularly for something that
> > >> pg_dump depends on.
> > >
> > > That alone would probably be sufficient reason why we would never allow
> > > pg_dump to depend on any such thing (not that I see a compelling
> > > argument for it to do so anyway...).
> >
> > It would be better to find a way to update sql-language functions in
> > minor upgrades, instead of shutting that door entirely for all future
> > implementation ideas involving sql-language functions in the
> > pg_catalog.
>
> I'd be very careful with jumping on this task. I am pretty sure its a
> very good way to get very, very frustrated if you don't present a widely
> accepted design beforehand. Doing this correctly is *far far* from easy.
>
> Just a little collection of problems:
> * You need to connect to all databases, not just one. There's no
> infrastructure for this.
> * You need to do the update *before* allowing any external
> connections. Otherwise the feature won't be useful to fix actual
> problems. Again, there is no infrastructure for this.
> * You need to do it in a way that a) doesn't slow down normal startup b)
> doesn't break if the update has only been applied to 9999 of 10000
> databases.

Another rather fundamental problem:

This obviously cannot be done directly on a standby. So either we cannot
rely on those updates having been performed or you need to update the
standby in lockstep with the primary. Neither seems acceptable.

Greetings,

Andres Freund

--
Andres Freund http://www.2ndQuadrant.com/
PostgreSQL Development, 24x7 Support, Training & Services

In response to

Browse pgsql-hackers by date

  From Date Subject
Next Message Kevin Grittner 2013-04-30 15:35:32 Re: Remaining beta blockers
Previous Message Andres Freund 2013-04-30 15:04:28 Re: Remaining beta blockers