On Thu, Feb 6, 2020 at 1:52 PM Andres Freund <andres@anarazel.de> wrote:
> I admit it's awkward. I think we possibly could still just make the size
> displayed in bytes in either case, reducing that issue a *bit*?
That seems like it makes it even more confusing, honestly.
> > It'd sort of be nicer to have two separate GUCs,
> > one measured as a multiple and the other measured in bytes, but maybe
> > that's just exchanging one form of confusion for another.
>
> We don't really have a good way to deal with GUCs where setting one
> precludes the other, especially when those GUCs should be changable at
> runtime :(.
It can work if one of the GUCs is king, and the other one takes effect
only the first one is set to some value that means "ignore me". We
have a number of examples of that, e.g. autovacuum_work_mem,
autovacuum_vacuum_cost_limit.
--
Robert Haas
EnterpriseDB: http://www.enterprisedb.com
The Enterprise PostgreSQL Company