Joachim Wieland wrote:
> On Fri, May 25, 2007 at 10:33:41AM -0400, Tom Lane wrote:
> > We should find out why that's happening rather than just throwing an
> > ORDER BY at it. Considering the number of buildfarm machines that
> > aren't showing any such problem, there must be something odd about
> > yours. What's the platform? What plan is being chosen for that SELECT?
>
> It's regular Debian Linux 2.6 on ix86.
>
> EXPLAIN SELECT conname FROM pg_constraint WHERE conrelid = 'clstr_tst'::regclass;
> QUERY PLAN
> ---------------------------------------------------------------------------------------------------
> Index Scan using pg_constraint_conrelid_index on pg_constraint (cost=0.00..8.27 rows=1 width=64)
> Index Cond: (conrelid = 170982::oid)
> (2 rows)
>
>
> SELECT conname FROM pg_constraint WHERE conrelid = 'clstr_tst'::regclass;
> conname
> ----------------
> clstr_tst_con
> clstr_tst_pkey
> (2 rows)
>
> As said before, it only happens with "make installcheck", not "make check".
Maybe there's an autovacuum run just before the test that causes
pg_constraint entries to be reordered?
--
Alvaro Herrera http://www.amazon.com/gp/registry/CTMLCN8V17R4
Oh, oh, las chicas galacianas, lo harán por las perlas,
¡Y las de Arrakis por el agua! Pero si buscas damas
Que se consuman como llamas, ¡Prueba una hija de Caladan! (Gurney Halleck)