[postgis-users] An I/O error occured while sending to the backend

[Mark Cave-Ayland]

On Wed, 2007-08-15 at 15:45 +0200, Stefan Zweig wrote:
> hi mark, hi list,
> 
> i have a problem with running these querys:
>  
>  SELECT
>  AddGeometryColumn('public','_g2994','the_geom_1','4326','MULTIPOLYGON',2);
>  ALTER TABLE _g2994 DROP CONSTRAINT enforce_geotype_the_geom_1;
>  ALTER TABLE _g2994 ADD CONSTRAINT enforce_geotype_the_geom_1 CHECK
>  (geometrytype(the_geom_1) = 'MULTIPOLYGON'::text OR geometrytype(the_geom_1)
>  = 'POLYGON'::text OR the_geom_1 IS NULL);
>  
>  everthing is alright up to here...
>  
>  then i run
>  UPDATE _g2994 SET the_geom_1 = buffer(simplify(the_geom,0.0063775005),0)
>  WHERE true
> 
> unfortunately i am not an expert in reading postgres' server.log but this what happens after the above querys where executed:


Hi Stefan,

Which platform are you using - what is the output of "SELECT version()"
and "SELECT postgis_full_version()"? If you Google for PostgreSQL/Signal
11 then it can be caused by bad hardware, a process exceeding its ulimit
threshold or a bug in the backend (perhaps GEOS/PostGIS).

In some cases, the problem can be caused by a single geometry within a
table - perhaps the buffer() and/or simplify() calculations are causing
in error? In this case you could try and identify the geometry in
question doing something like:

CREATE TABLE test AS SELECT gid,
buffer(simplify(the_geom,0.0063775005),0) FROM _g2994;

If this crashes, then the first geometry that isn't in the test table is
causing it to crash, i.e. the geometry with the next highest gid.


HTH,

Mark.

-- 
ILande - Open Source Consultancy
http://www.ilande.co.uk