Itagaki Takahiro <itagaki.takahiro@oss.ntt.co.jp> writes:
> Is it reasonable to replace relation_open() calls to try_relation_open() ?
I don't think so; that's just papering over one form of the problem,
and who's to say that an error is not desirable when a wrong OID is
given? In general there's always a risk of concurrency problems.
regards, tom lane