Re: vacuum_cost_page_miss default value and modern hardware

On Thu, Jan 21, 2021 at 5:12 PM Peter Geoghegan <pg(at)bowt(dot)ie> wrote:
> I'm going to go ahead with committing my patch to lower the default
> next week. If anybody has any objections to that plan, please speak
> up.

Just pushed a commit that reduced the default for vacuum_cost_page_miss to 2.

One more thing on this: I noticed that the docs for these settings say
"There are many situations where it is not important that maintenance
commands like VACUUM and ANALYZE finish quickly". That now seems a
little dubious to me -- I wonder if we should refresh it.

There are not that many problems that can be solved by making VACUUM
slower. This is due to the visibility map, and to a lesser degree
certain key improvements in index AMs. The text that I quoted was
written in 2005, a time when the delay stuff was still very new, and
even the earliest visibility map design was still a few years away.

Thanks
--
Peter Geoghegan