"Tim Barnard" <tbarnard@povn.com> writes:
> <snip>
> If it does happen, you can just retry the failed object-creation
> command until you get a nonconflicting OID assignment.
> <snip>
> Tom, or anyone who'd know for that matter, what is the
> exact error I would need to look for?
It'd be a complaint about attempt to insert a duplicate key into
the unique index on one of several system catalogs' OID columns.
There wouldn't be any earlier detection of the problem than that.
regards, tom lane