От: Tom Lane
Тема: Re: Slow query problem
Дата: ,
Msg-id: 20723.1073625127@sss.pgh.pa.us
(см: обсуждение, исходный текст)
Ответ на: Re: Slow query problem  (Mike Glover)
Список: pgsql-performance

Скрыть дерево обсуждения

Slow query problem  (Bradley Tate, )
 Re: Slow query problem  (Mike Glover, )
  Re: Slow query problem  (Bruno Wolff III, )
  Re: Slow query problem  (Tom Lane, )
 Re: Slow query problem  (Dennis Björklund, )
  Re: Slow query problem  (Richard Huxton, )
   Re: Slow query problem  (Dennis Björklund, )
    Re: Slow query problem  (Richard Huxton, )
     Re: Slow query problem  (Tom Lane, )
     Re: Slow query problem  (Stephan Szabo, )
    Re: Slow query problem  (Bradley Tate, )

Mike Glover <> writes:
> You should bump sort_mem as high as you can stand.  with only 8MB sort
> memory available, you're swapping intermediate sort pages to disk --
> a lot. Try the query with sort_mem set to 75MB (to do the entire sort in
> memory).

7.4 will probably flip over to a hash-based aggregation method, and not
sort at all, once you make sort_mem large enough that it thinks the hash
table will fit in sort_mem.

            regards, tom lane


В списке pgsql-performance по дате сообщения:

От: Richard Huxton
Дата:
Сообщение: Re: Slow query problem
От: Bradley Tate
Дата:
Сообщение: Re: Slow query problem