Re: 8.3beta3 ERROR: cached plan must not change result type

Hi,

I am seeing this error with 8.3beta3 on debian unstable:

2007-11-28 15:15:46 CET ERROR: cached plan must not change result type
2007-11-28 15:15:46 CET STATEMENT: SELECT s.*,
coalesce(d.firstname||' ','')||coalesce(d.lastname,'') as person_name,
person_url(d.id_person),
t.person_type,
t.id_person
from show s
left join person_to_show t using (id_show)
left join person d using (id_person)
where s.id_show = $1

The query is sent by a mod_perl app and a restart of apache2 fixes it.
Let me know if you need more info.

Cheers,

Louis-David Mitterrand <vindex+lists-pgsql-hackers(at)apartia(dot)org> writes:
> I am seeing this error with 8.3beta3 on debian unstable:

> 2007-11-28 15:15:46 CET ERROR: cached plan must not change result type

> Let me know if you need more info.

Yes.

regards, tom lane

On Wed, Nov 28, 2007 at 10:45:31AM -0500, Tom Lane wrote:
> Louis-David Mitterrand <vindex+lists-pgsql-hackers(at)apartia(dot)org> writes:
> > I am seeing this error with 8.3beta3 on debian unstable:
>
> > 2007-11-28 15:15:46 CET ERROR: cached plan must not change result type
>
> > Let me know if you need more info.
>
> Yes.

Okay :) I guess I had it coming.

The only thing I know is this error started showing up in the logs at
15:15:46 making our site unavailble until I restarted apache.

No postgres code has been modified recently on the app and it's the
first time I see that error.

Also I don't know, yet, how to reproduce it. Will try and report.

On Wed, Nov 28, 2007 at 04:51:17PM +0100, Louis-David Mitterrand wrote:
> On Wed, Nov 28, 2007 at 10:45:31AM -0500, Tom Lane wrote:
> > Louis-David Mitterrand <vindex+lists-pgsql-hackers(at)apartia(dot)org> writes:
> > > I am seeing this error with 8.3beta3 on debian unstable:
> >
> > > 2007-11-28 15:15:46 CET ERROR: cached plan must not change result type
> >
> > > Let me know if you need more info.
> >
> > Yes.
>
> Okay :) I guess I had it coming.
>
> The only thing I know is this error started showing up in the logs at
> 15:15:46 making our site unavailble until I restarted apache.
>
> No postgres code has been modified recently on the app and it's the
> first time I see that error.
>
> Also I don't know, yet, how to reproduce it. Will try and report.

Wow, this above email took three days to reach to list.

Received: from maia-2.hub.org (maia-2.hub.org [200.46.204.187])
by zenon.apartia.fr (Postfix) with ESMTP id 2B95E59C5B43C
for <vindex+lists-pgsql-hackers(at)apartia(dot)org>; Sat, 1 Dec 2007 02:14:00
+0100 (CET)
Received: from postgresql.org (postgresql.org [200.46.204.71])
by maia-2.hub.org (Postfix) with ESMTP id C5C1F2CA2E2
for <vindex+lists-pgsql-hackers(at)apartia(dot)org>; Wed, 28 Nov 2007 11:51:20
-0400 (AST)

Why did maia-2.hub.org keep it so long in its belly?

Louis-David Mitterrand wrote:

> Wow, this above email took three days to reach to list.
>
> Received: from maia-2.hub.org (maia-2.hub.org [200.46.204.187])
> by zenon.apartia.fr (Postfix) with ESMTP id 2B95E59C5B43C
> for <vindex+lists-pgsql-hackers(at)apartia(dot)org>; Sat, 1 Dec 2007 02:14:00
> +0100 (CET)
> Received: from postgresql.org (postgresql.org [200.46.204.71])
> by maia-2.hub.org (Postfix) with ESMTP id C5C1F2CA2E2
> for <vindex+lists-pgsql-hackers(at)apartia(dot)org>; Wed, 28 Nov 2007 11:51:20
> -0400 (AST)
>
> Why did maia-2.hub.org keep it so long in its belly?

I got it immediately:

Received: from maia-2.hub.org (maia-2.hub.org [200.46.204.187])
by perhan.alvh.no-ip.org (Postfix) with ESMTP id 0B76547BFF
for <alvherre(at)alvh(dot)no-ip(dot)org>; Wed, 28 Nov 2007 12:52:30 -0300 (CLST)
Received: from postgresql.org (postgresql.org [200.46.204.71])
by maia-2.hub.org (Postfix) with ESMTP id ED7032CA417
for <alvherre(at)alvh(dot)no-ip(dot)org>; Wed, 28 Nov 2007 11:52:21 -0400 (AST)

Perhaps your system refused mail from maia-2 for a while and then it got
queued. How could have 3 days passed for it to be delivered, I don't know.

--
Alvaro Herrera Developer, http://www.PostgreSQL.org/
"Prefiero omelette con amigos que caviar con tontos"
(Alain Nonnet)