Re: Binary upgrade from <12 to 12 creates toast table for partitioned tables

Hi,

On 2019-03-07 13:08:35 -0500, Robert Haas wrote:
> On Wed, Mar 6, 2019 at 3:41 PM Andres Freund <andres(at)anarazel(dot)de> wrote:
> > I think we probably should have pg_dump suppress emitting information
> > about the toast table of partitioned tables?
>
> +1. That seems like the right fix.

Cool. Alvaro, Kyatoro, Michael, are either of you planning to tackle
that? Afaict it's caused by

commit 807ae415c54628ade937cb209f0fc9913e6b0cf5
Author: Alvaro Herrera <alvherre(at)alvh(dot)no-ip(dot)org>
Date: 2019-01-04 14:51:17 -0300

Don't create relfilenode for relations without storage

Some relation kinds had relfilenode set to some non-zero value, but
apparently the actual files did not really exist because creation was
prevented elsewhere. Get rid of the phony pg_class.relfilenode values.

Catversion bumped, but only because the sanity_test check will fail if
run in a system initdb'd with the previous version.

Reviewed-by: Kyotaro HORIGUCHI, Michael Paquier
Discussion: https://postgr.es/m/20181206215552.fm2ypuxq6nhpwjuc@alvherre.pgsql

> > While I'm not hugely bothered by binary upgrade mode creating
> > inconsistent states - there's plenty of ways to crash the server that
> > way - it probably also would be a good idea to have heap_create()
> > elog(ERROR) when accessmtd is invalid.
>
> Not sure about this part.

As in, we shouldn't elog out? Or we should have an ereport with a proper
error, or ...?

Greetings,

Andres Freund