Re: pgsql: Recursively fsync() the data directory after a crash.

Поиск
Список
Период
Сортировка
Искать
От
Andrew Dunstan
Тема
Re: pgsql: Recursively fsync() the data directory after a crash.
Дата
в 03:37:32
Msg-id
55483AF1.50208@dunslane.net
Ответ на
Список
Дерево обсуждения
pgsql: Recursively fsync() the data directory after a crash. Robert Haas <rhaas@postgresql.org>
Re: pgsql: Recursively fsync() the data directory after a crash. Andrew Dunstan <andrew@dunslane.net>
Re: pgsql: Recursively fsync() the data directory after a crash. Robert Haas <robertmhaas@gmail.com>
Re: pgsql: Recursively fsync() the data directory after a crash. Peter Eisentraut <peter_e@gmx.net>
Re: pgsql: Recursively fsync() the data directory after a crash. Robert Haas <robertmhaas@gmail.com>
Re: pgsql: Recursively fsync() the data directory after a crash. Tom Lane <tgl@sss.pgh.pa.us>
Re: pgsql: Recursively fsync() the data directory after a crash. Robert Haas <robertmhaas@gmail.com>
Re: pgsql: Recursively fsync() the data directory after a crash. Tom Lane <tgl@sss.pgh.pa.us>
Re: pgsql: Recursively fsync() the data directory after a crash. Bruce Momjian <bruce@momjian.us>
Re: pgsql: Recursively fsync() the data directory after a crash. Tom Lane <tgl@sss.pgh.pa.us>
Re: pgsql: Recursively fsync() the data directory after a crash. Robert Haas <robertmhaas@gmail.com>
Re: pgsql: Recursively fsync() the data directory after a crash. Alvaro Herrera <alvherre@2ndquadrant.com>
Re: pgsql: Recursively fsync() the data directory after a crash. Robert Haas <robertmhaas@gmail.com>
Re: pgsql: Recursively fsync() the data directory after a crash. Tom Lane <tgl@sss.pgh.pa.us>

On 05/04/2015 02:23 PM, Robert Haas wrote:
> Recursively fsync() the data directory after a crash.
>
> Otherwise, if there's another crash, some writes from after the first
> crash might make it to disk while writes from before the crash fail
> to make it to disk.  This could lead to data corruption.
>
> Back-patch to all supported versions.
>

This appears to have broken Windows builds. At least it's broken by 
using the two argument form of open instead of the three argument form, 
for which we have a #define in the win32 case, and pg_win32_is_junction 
is a typo - the first _ should not be there.

cheers

andrew

В списке pgsql-committers по дате отправления
От: Peter Eisentraut
Дата:
От: Robert Haas
Дата:
FAQ