For any of you with those failover clusters, do you know if "pg_ctl reload" works (for compatible config file changes), or must we bounce the database using "hares -offline" then "hares -online"?
pg_ctl reload does work in this case. HOWEVER, if you do something that could cause trouble to the cluster (i.e. a pg_hba.conf change that breaks connectivity), this could cause veritas to try and failover. It's recommended that you test your changes to avoid these scenarios.
Technically however, pg_ctl reload works just fine, just don't break anything :)