On 04/24/2013 11:20 AM, Daniel Cristian Cruz wrote:
I've done an explain analyze under the test environment, and there is no aggressive memory usage.
So I dropped the database in the new cluster and restored a fresh dump from production (in theory, that's the difference between the two environments).
Some minutes after I got an answer: after a dump / restore, there is no problem. The same cluster just a dump/restore.
Since I had no idea on what is the problem, and the structure from the dump of the bad database is equal to the dump from the production (it differs only in check constraints where "((turma.situacao)::text = ANY ((ARRAY['Aguardando Atualização'::character varying, 'Em Andamento'::character varying])::text[])))" became "((turma.situacao)::text = ANY (ARRAY[('Aguardando Atualização'::character varying)::text, ('Em Andamento'::character varying)::text])))"), how can I report a issue in pg_upgrade?
Thanks, Adrian, Tomas and Tom. It's a specific issue, it's on backup until May, 1st. Until there, if someone would like to know something about it, just ask me, I would like to help removing an issue.