Re: Weird XFS WAL problem

От: Matthew Wakeling
Тема: Re: Weird XFS WAL problem
Дата: ,
Msg-id: alpine.DEB.2.00.1006041009540.4083@aragorn.flymine.org
(см: обсуждение, исходный текст)
Ответ на: Re: Weird XFS WAL problem  (Greg Smith)
Список: pgsql-performance

Скрыть дерево обсуждения

Random Page Cost and Planner  (David Jarvis, )
 Re: Random Page Cost and Planner  (David Jarvis, )
 Re: Random Page Cost and Planner  ("Kevin Grittner", )
  Re: Random Page Cost and Planner  (David Jarvis, )
   Re: Random Page Cost and Planner  (Tom Lane, )
    Re: Random Page Cost and Planner  (David Jarvis, )
     Re: Random Page Cost and Planner  (Alexey Klyukin, )
      Re: Random Page Cost and Planner  (David Jarvis, )
   Re: Random Page Cost and Planner  (Rob Wultsch, )
    Re: Random Page Cost and Planner  (David Jarvis, )
     Re: Random Page Cost and Planner  (, )
   Re: Random Page Cost and Planner  ("Kevin Grittner", )
    Re: Random Page Cost and Planner  (David Jarvis, )
     Re: Random Page Cost and Planner  (David Jarvis, )
 Re: Random Page Cost and Planner  (, )
  Re: Random Page Cost and Planner  (David Jarvis, )
 Re: Random Page Cost and Planner  (David Jarvis, )
  Re: Random Page Cost and Planner  (Cédric Villemain, )
   Re: Random Page Cost and Planner  (David Jarvis, )
    Re: Random Page Cost and Planner  (Cédric Villemain, )
     Weird XFS WAL problem  (Craig James, )
      Re: Weird XFS WAL problem  (Mark Kirkwood, )
       Re: Weird XFS WAL problem  (Craig James, )
        Re: Weird XFS WAL problem  (Matthew Wakeling, )
         Re: Weird XFS WAL problem  ("Kevin Grittner", )
          Re: Weird XFS WAL problem  (Greg Smith, )
           Re: Weird XFS WAL problem  ("Kevin Grittner", )
            Re: Weird XFS WAL problem  (Scott Marlowe, )
             Re: Weird XFS WAL problem  ("Kevin Grittner", )
             Re: Weird XFS WAL problem  (Greg Smith, )
              Re: Weird XFS WAL problem  (Scott Marlowe, )
              Re: Weird XFS WAL problem  ("Kevin Grittner", )
              Re: Weird XFS WAL problem  (Matthew Wakeling, )
           Re: Weird XFS WAL problem  (Bruce Momjian, )
            Re: Weird XFS WAL problem  ("Kevin Grittner", )
             Re: Weird XFS WAL problem  (Bruce Momjian, )
              Re: Weird XFS WAL problem  ("Kevin Grittner", )
               Re: Weird XFS WAL problem  (Bruce Momjian, )
                Re: Weird XFS WAL problem  ("Kevin Grittner", )
                 Re: Weird XFS WAL problem  (Bruce Momjian, )
                 Re: Weird XFS WAL problem  (Greg Smith, )
                  Re: Weird XFS WAL problem  (Bruce Momjian, )
        Re: Weird XFS WAL problem  (Greg Smith, )
      Re: Weird XFS WAL problem  (Merlin Moncure, )
      Re: Weird XFS WAL problem  (Greg Smith, )
  Re: Random Page Cost and Planner  (Bryan Hinton, )
   Re: Random Page Cost and Planner  (David Jarvis, )

On Thu, 3 Jun 2010, Greg Smith wrote:
> And it's also quite reasonable for a RAID controller to respond to that
> "flush the whole cache" call by flushing its cache.

Remember that the RAID controller is presenting itself to the OS as a
large disc, and hiding the individual discs from the OS. Why should the OS
care what has actually happened to the individual discs' caches, as long
as that "flush the whole cache" command guarantees that the data is
persistent. Taking the RAID array as a whole, that happens when the data
hits the write-back cache.

The only circumstance where you actually need to flush the data to the
individual discs is when you need to take that disc away somewhere else
and read it on another system. That's quite a rare use case for a RAID
array (http://thedailywtf.com/Articles/RAIDing_Disks.aspx
notwithstanding).

> If the controller had some logic that said "it's OK to not flush the
> cache when that call comes in if my battery is working fine", that would
> make this whole problem go away.

The only place this can be properly sorted is the RAID controller.
Anywhere else would be crazy.

Matthew

--
"To err is human; to really louse things up requires root
 privileges."                 -- Alexander Pope, slightly paraphrased


В списке pgsql-performance по дате сообщения:

От: Andres Freund
Дата:
Сообщение: Re: performance regression with Linux 2.6.33 and glibc 2.12
От: Marc Cousin
Дата:
Сообщение: Re: performance regression with Linux 2.6.33 and glibc 2.12