Re: Auto-tuning work_mem and maintenance_work_mem
От | Magnus Hagander |
---|---|
Тема | Re: Auto-tuning work_mem and maintenance_work_mem |
Дата | |
Msg-id | CABUevEwkNOpexCGabp=mU2ZhpoiQ086iHEdeJqxEz+gfLcpKpw@mail.gmail.com обсуждение исходный текст |
Ответ на | Re: Auto-tuning work_mem and maintenance_work_mem (Peter Geoghegan <pg@heroku.com>) |
Список | pgsql-hackers |
On Thu, Oct 10, 2013 at 5:35 AM, Peter Geoghegan <pg@heroku.com> wrote: > On Wed, Oct 9, 2013 at 8:20 PM, Bruce Momjian <bruce@momjian.us> wrote: >> I am not sure that having that external to the backend really makes >> sense because I am concerned people will not use it. We can certainly >> add it to change our defaults, of course. Also consider many installs >> are automated. > > Sure. > > I was imagining that we'd want to write the tool with the idea in mind > that it was usually run immediately after initdb. We'd reach out to > packagers to have them push it into the hands of users where that's > practical. > > If you think that sounds odd, consider that on at least one popular > Linux distro, installing MySQL will show a ncurses interface where the > mysql password is set. We wouldn't need anything as fancy as that. That's a packaging feature though, and not a MySQL feature. And of course, on other platforms, popping up somethjing like that is explicitly *forbidden* by the packaging standards. But it shows an important distinction - we really only need to provide an *infrastructure* that can be used on all platforms. The platform specific interfaces to it can go in the packaging. (And yes, in a lot of cases "we" are also the packagers, but the point being that this code is already 100% platform specific, making the problem easier) -- Magnus HaganderMe: http://www.hagander.net/Work: http://www.redpill-linpro.com/
В списке pgsql-hackers по дате отправления: