Date: Tue, 28 Aug 2018 12:00:45 -0300
From: Alvaro Herrera <alvherre@2ndquadrant.com>
To: wambacher@posteo.de
Cc: pgsql-admin@lists.postgresql.org
Subject: Re: tuple concurrently updated
Message-ID: <20180828150045.4jp4x2jvzod3ooik@alvherre.pgsql>
MIME-Version: 1.0
Content-Type: text/plain; charset=iso-8859-1
Content-Disposition: inline
Content-Transfer-Encoding: 8bit
In-Reply-To: <fc354354-bf6a-622c-e045-d554c58d9b31@posteo.de>
User-Agent: NeoMutt/20170306-137-4415bd-dirty (1.8.0)
Precedence: bulk

On 2018-Aug-28, wambacher@posteo.de wrote:

> no, i respect your questions, but i can't answer them. because i don't
> know in detail what osm2pgsql is doing.
> 
> planet_osm_line it postgis

> >              Spalte              |            Typ            |
> > Sortierfolge | NULL erlaubt? | Vorgabewert
> > ---------------------------------+---------------------------+--------------+---------------+-------------

> >  way                             | geometry(LineString,4326)
> > |              |               |
> > Indexe:
> >     "idx_planet_osm_line_highway" btree (highway)
> >     "planet_osm_line_index" gist (way), Tablespace »planet3_is1«
> >     "planet_osm_line_pkey" btree (osm_id), Tablespace »planet3_is1«
> >     "planet_osm_line_tags_index" gin (tags), Tablespace »planet3_is1«
> > Tablespace: »planet3_ts1«

I bet that something is happening to the special Postgis catalog table
for the geometry type when you run this update.  Maybe, say, if you have
two such updates in a transaction, things break.  The innards of
geometry_in are, um, fascinating (though I can't tell for sure whether
it tries to modify anything).

Also, I wouldn't put my hands to the fire for the gist support for it.

I further bet you'd have more luck asking a Postgis mailing list rather
than pgsql-admin.

-- 
Álvaro Herrera                https://www.2ndQuadrant.com/
PostgreSQL Development, 24x7 Support, Remote DBA, Training & Services