Re: [GENERAL] how to get accurate values in pg_statistic (continued)

От: Tom Lane
Тема: Re: [GENERAL] how to get accurate values in pg_statistic (continued)
Дата: ,
Msg-id: 7126.1062787095@sss.pgh.pa.us
(см: обсуждение, исходный текст)
Ответ на: [GENERAL] how to get accurate values in pg_statistic (continued)  (Mary Edie Meredith)
Список: pgsql-performance

Скрыть дерево обсуждения

[GENERAL] how to get accurate values in pg_statistic (continued)  (Mary Edie Meredith, )
 Re: [GENERAL] how to get accurate values in pg_statistic  (Rod Taylor, )
  Re: [GENERAL] how to get accurate values in pg_statistic  (Neil Conway, )
   Re: [GENERAL] how to get accurate values in pg_statistic  (Mary Edie Meredith, )
   Re: [GENERAL] how to get accurate values in pg_statistic  (Rod Taylor, )
 Re: [GENERAL] how to get accurate values in pg_statistic (continued)  (Tom Lane, )
  Re: [GENERAL] how to get accurate values in pg_statistic  (Mary Edie Meredith, )
   Re: [GENERAL] how to get accurate values in pg_statistic  (Bruce Momjian, )
    Re: [GENERAL] how to get accurate values in pg_statistic  (Tom Lane, )
   Re: [GENERAL] how to get accurate values in pg_statistic  (Bruce Momjian, )
   Re: [GENERAL] how to get accurate values in pg_statistic  (Greg Stark, )
    Re: [GENERAL] how to get accurate values in pg_statistic  (Tom Lane, )
  Re: [GENERAL] how to get accurate values in pg_statistic (continued)  (Bruce Momjian, )
   Re: [osdldbt-general] Re: [GENERAL] how to get accurate  (Mary Edie Meredith, )
   Re: [osdldbt-general] Re: [GENERAL] how to get accurate  (Christopher Browne, )
  Re: [GENERAL] how to get accurate values in pg_statistic (continued)  (Christopher Browne, )
   Re: [GENERAL] how to get accurate values in pg_statistic (continued)  (Tom Lane, )
    Re: [GENERAL] how to get accurate values in pg_statistic  ("scott.marlowe", )
    Re: [GENERAL] how to get accurate values in pg_statistic  (Christopher Browne, )
     Re: [GENERAL] how to get accurate values in pg_statistic  ("scott.marlowe", )
 Re: [GENERAL] how to get accurate values in pg_statistic (continued)  (Tom Lane, )

Mary Edie Meredith <> writes:
> For our testing purposes, however, consistent results are extremely
> important. We have observed that small difference in one plan for one of
> 22 queries can cause a difference in the DBT-3 results.  If this
> happens, a small change in performance runs between two Linux kernels
> may appear to be due to the kernels, when in fact it is due to the plan
> change.

Fair enough.  If you are trying to force exactly repeatable results,
why don't you just "set seed = 0" before you ANALYZE?  There's only
one random-number generator, so that should force ANALYZE to make the
same random sampling every time.

Also, it'd be a good idea to ANALYZE the needed tables by name,
explicitly, to ensure that they are analyzed in a known order
rather than whatever order ANALYZE happens to find them in pg_class.

            regards, tom lane


В списке pgsql-performance по дате сообщения:

От: "Relaxin"
Дата:
Сообщение: Re: SELECT's take a long time compared to other DBMS
От: Hannu Krosing
Дата:
Сообщение: Re: Serious issues with CPU usage