Re: The suggestion of reducing autovacuum_vacuum_cost_delay should be documented

On Tue, May 12, 2020 at 4:30 PM Bruce Momjian <bruce(at)momjian(dot)us> wrote:

> On Tue, May 12, 2020 at 07:26:07PM -0400, Tom Lane wrote:
> > Bruce Momjian <bruce(at)momjian(dot)us> writes:
> > > On Tue, May 12, 2020 at 03:38:16PM -0700, David G. Johnston wrote:
> > >> The proposal is to document in versions 9.4 to 11 that the
> recommended value
> > >> for the setting is 2ms while for reasons of continuity the default in
> these
> > >> versions is 20ms.
> > >> I don't really see any harm in it. Its not like the choice to reduce
> the value
> > >> was made because of new features introduced in 12 - it was a
> re-evaluation of a
> > >> 15 year old default.
> >
> > > Well, we really need to have some general discussion about whether
> > > changing defaults in major releases should trigger a mention to change
> > > the defaults in back branches. This is something that would have to be
> > > discussed on the hackers list.
> >
> > It's not immediately obvious that the new default value established in
> > version N is appropriate for version N-minus-several. Certainly,
> whatever
> > testing was done to justify the new default wouldn't have been done on
> old
> > versions; and there might have been relevant changes.
> >
> > In short: nope, I'm not on board with blindly back-patching such
> > recommendations.
>
> That was my analysis too.
>

To be clear, because my cursory reading of the thread that was linked from
the commit suggested that this specific situation was more "lets catch up
to modern times", my position isn't that such documentation changes should
be done as a rule, I am suggesting that we give a yes/no decision on this
specific change (in advance of bike-shedding the wording). IMO neither a
blanket rule allowing or prohibiting such a change to the documentation
makes sense given the rarity of the event.

David J.