Re: Recommended optimisations slows down PostgreSQL 8.4

От: Stephen Frost
Тема: Re: Recommended optimisations slows down PostgreSQL 8.4
Дата: ,
Msg-id: 20110812210742.GB12765@tamriel.snowman.net
(см: обсуждение, исходный текст)
Ответ на: Re: Recommended optimisations slows down PostgreSQL 8.4  (Waldo Nell)
Список: pgsql-performance

Скрыть дерево обсуждения

Recommended optimisations slows down PostgreSQL 8.4  (Waldo Nell, )
 Re: Recommended optimisations slows down PostgreSQL 8.4  ("", )
  Re: Recommended optimisations slows down PostgreSQL 8.4  (Waldo Nell, )
   Re: Recommended optimisations slows down PostgreSQL 8.4  (Merlin Moncure, )
    Re: Recommended optimisations slows down PostgreSQL 8.4  (Waldo Nell, )
     Re: Recommended optimisations slows down PostgreSQL 8.4  ("Kevin Grittner", )
      Re: Recommended optimisations slows down PostgreSQL 8.4  (Waldo Nell, )
     Re: Recommended optimisations slows down PostgreSQL 8.4  (Stephen Frost, )
 Re: Recommended optimisations slows down PostgreSQL 8.4  (Greg Smith, )
  Re: Recommended optimisations slows down PostgreSQL 8.4  (Waldo Nell, )
   Re: Recommended optimisations slows down PostgreSQL 8.4  (Greg Smith, )
 Re: Recommended optimisations slows down PostgreSQL 8.4  (Tom Lane, )

* Waldo Nell () wrote:
> The fsync = off was because the production system runs on a uber expensive SAN system with multipathing over Fibre
Channel,it is on UPS and backup generators in a secure datacenter, and we have daily backups we can fall back to. 

So, two points: #1- the uber-expensive SAN should make twiddling fsync
have much less of an effect on performance than in a non-SAN/non-BBWC
environment, so you might validate that you really need it off.  #2- the
SAN, FC, UPS, etc will be of no help if the OS or PG crash.  Seems
pretty harsh to resort back to a daily backup in the event the OS reboots
due to some wacky NMI, or the ASR going haywire..

    Thanks,

        Stephen

Вложения

В списке pgsql-performance по дате сообщения:

От: Craig Ringer
Дата:
Сообщение: Re: strange pgbench results (as if blocked at the end)
От: Greg Smith
Дата:
Сообщение: Re: strange pgbench results (as if blocked at the end)