Tom Lane wrote:
> I don't see any very good argument for allowing this mechanism to set
> minxid = FrozenXid in the first place. If there are only frozenXid in
> the table, set minxid = current XID. That eliminates the entire problem
> at a stroke.
Ok, so I shall go back to the original patch, which did exactly this.
Is it OK for applying?
(I'm using RecentXmin instead of current XID though, because a
currently-running transaction could insert tuples in the table I just
vacuumed.)
--
Alvaro Herrera http://www.CommandPrompt.com/
The PostgreSQL Company - Command Prompt, Inc.