Le 05/26/2011 01:18 AM, Tatsuo Ishii a écrit : >>>> There shouldn't be any harm to do so. Except you won't be able to create >>>> database with the default template database (template1) while pgpool is >>>> connected to the template1 database. But it connects to it every >>>> health_check_period (or every 30 sec if health_check_period<=0). >>>> >>>> Anyway, this is probably something we should fix. You're not required to >>>> have a postgres database. >>> >>> Are you sure? Many of PostgreSQL tools require to have postgres >>> database. For example, createdb and createuser. So I think a >>> PostgreSQL configuration without postgres database is broken anyway >>> and I think pgpool-Ii do not need to take care of the case. >>> >>> BTW for #3, probably you are talking about >>> establish_persistent_connection(void) rather than >>> make_persistent_db_connection() takes database parameter, not fixed >>> usage of postgres database. establish_persistent_connection() is used >>> for streaming replication lag check, which is only available 9.0 or >>> later, which definitely has postgres database if you do not broke. >> >> You can drop the postgres database, and use another one as your postgres >> database. You only need to use the PGDATABASE env variable. It makes it >> really cumbersome to use createdb, createuser, and other tools, and I >> see no reason for doing it, but it works. > > Ok. IMO think this kind of usecase is really rare, and fixing it is > not worth the trouble.
If Asif makes the patch himself, I don't think we should reject it. But it's your call. -- Guillaume http://www.postgresql.fr http://dalibo.com _______________________________________________ Pgpool-hackers mailing list [email protected] http://pgfoundry.org/mailman/listinfo/pgpool-hackers
