Re: partitioned indexes and tablespaces

On Wed, Nov 7, 2018 at 10:46 AM Alvaro Herrera <alvherre(at)2ndquadrant(dot)com> wrote:
> I think 11.0 is ready for testing that a migration from a production
> running 10.x, but not for just blindly migrating. If you wanted to take
> such a leap of faith, surely you'd wait for 11.1 at the very least.

I think that's an irresponsible attitude for a committer to take. In
practice, you are probably right, but we shouldn't treat our
supposedly-stable releases as if they don't really need to be kept
stable. That's a recipe for anybody back-patching anything they feel
like whenever they like, which is a recipe for disaster.

But maybe you've adopted that policy already. You back-patched a
behavior change 2 days before a minor release when the vote was 2-3
against the change. If it matters, the three people opposing it all
work for different companies, wheres your only concurring vote came
from someone with whom you share an employer. I though the principle
here was that we operate by consensus; that is not a consensus to
proceed at all, let alone to back-patch on short notice.

> > Seeing this stuff discussed and committed in haste gives me the sad
> > face.
>
> Nah.

Sorry, but you don't get to decide what makes other people sad. I'm
with Michael on this one.

--
Robert Haas
EnterpriseDB: http://www.enterprisedb.com
The Enterprise PostgreSQL Company