Re: 答复: [PERFORM] Postgresql update op is very very slow

От: Tom Lane
Тема: Re: 答复: [PERFORM] Postgresql update op is very very slow
Дата: ,
Msg-id: 16213.1214492661@sss.pgh.pa.us
(см: обсуждение, исходный текст)
Ответ на: 答复: [PERFORM] Postgresql update op is very very slow  ("jay")
Ответы: Re: [PERFORM] 答复: [PERFORM] Postgresql update op is very very slow  ("Scott Marlowe")
Список: pgsql-performance

Скрыть дерево обсуждения

Postgresql is very slow  (bijayant kumar, )
 Re: Postgresql is very slow  (Jan de Visser, )
 Re: Postgresql is very slow  (Jeremy Harris, )
 Re: Postgresql is very slow  (, )
  Re: Postgresql is very slow  (bijayant kumar, )
   Re: Postgresql is very slow  ("Scott Marlowe", )
    Re: Postgresql is very slow  ("Scott Marlowe", )
    Re: Postgresql is very slow  (, )
     Re: Postgresql is very slow  (bijayant kumar, )
      Re: Postgresql is very slow  (, )
       Re: Postgresql is very slow  (bijayant kumar, )
       Re: [SOLVED] Postgresql is very slow  (bijayant kumar, )
      Re: Postgresql is very slow  (PFC, )
       Postgresql update op is very very slow  ("jay", )
        Re: Postgresql update op is very very slow  (Rusty Conover, )
         答复: [PERFORM] Postgresql update op is very very slow  ("jay", )
          PostgreSQL and Ruby on Rails - better accessibility  ("Amol Pujari", )
           Re: PostgreSQL and Ruby on Rails - better accessibility  (Nikhils, )
        Re: Postgresql update op is very very slow  ("Heikki Linnakangas", )
         答复: [PERFORM] Postgresql update op is very very slow  ("jay", )
          Re: [PERFORM] 答复: [PERFORM] Postgresql update op is very very slow  ("Pavan Deolasee", )
           Re: [PERFORM] Re: [PERFORM] 答复: [PERFORM] Postgresql update op is very very slow  ("Scott Marlowe", )
          Re: ??: Postgresql update op is very very slow  ("Heikki Linnakangas", )
          Re: 答复: [PERFORM] Postgresql update op is very very slow  (Tom Lane, )
           Re: [PERFORM] 答复: [PERFORM] Postgresql update op is very very slow  ("Scott Marlowe", )
        Re: Postgresql update op is very very slow  (Greg Smith, )
      Re: ??: Postgresql update op is very very slow  ("Holger Hoffstaette", )
       Re: ??: Postgresql update op is very very slow  (Craig Ringer, )
        Re: ??: Postgresql update op is very very slow  (Andrew Sullivan, )
        Re: ??: Postgresql update op is very very slow  (Greg Smith, )
       Re: ??: Postgresql update op is very very slow  (Andrew Sullivan, )
       Re: ??: Postgresql update op is very very slow  (Mark Mielke, )
       Re: ??: Postgresql update op is very very slow  (Greg Smith, )
    Re: Postgresql is very slow  ("Ian Barwick", )
 Re: Postgresql is very slow  ("Gregory S. Youngblood", )

"jay" <> writes:
>     I know the problem, because there are about 35 million rows , which
> cost about 12G disk space and checkpoint segments use 64, but update
> operation is in one transaction which lead fast fill up the checkpoint
> segments and lead do checkpoints frequently, but checkpoints will cost lots
> resources, so update operation become slowly and slowly and bgwrite won't
> write because it's not commit yet.
> Create a new table maybe a quick solution, but it's not appropriated in some
> cases.
>     If we can do commit very 1000 row per round, it may resolve the
> problem.

No, that's utterly unrelated.  Transaction boundaries have nothing to do
with checkpoints.

            regards, tom lane


В списке pgsql-performance по дате сообщения:

От: "Jonah H. Harris"
Дата:
Сообщение: Re: Federated Postgresql architecture ?
От: Josh Berkus
Дата:
Сообщение: Re: Federated Postgresql architecture ?