On Thu, Aug 06, 2020 at 11:50:06AM +0900, Michael Paquier wrote:
> True that you need an extra GRANT USAGE ON pg_toast to achieve that
> for users with no privileges, but that's not impossible now either. I
> am not sure that this use-case justifies a new option and more
> complications in the code paths of vacuum though. So let's see first
> if others have an opinion to offer.
Seeing nothing happening here, I am marking the CF entry as returned
with feedback. FWIW, I still tend to think that we could call this
stuff a day if we had an option to skip a toast relation when willing
to vacuum the parent relation.
--
Michael
