Re: get_controlfile() can leak fds in the backend

Поиск
Список
Период
Сортировка
Искать
От
Andres Freund
Тема
Re: get_controlfile() can leak fds in the backend
Дата
Msg-id
20190227093231.ermj4gurayhxiyyo@alap3.anarazel.de
Ответ на
Список
Дерево обсуждения
get_controlfile() can leak fds in the backend Michael Paquier <michael@paquier.xyz>
Re: get_controlfile() can leak fds in the backend Joe Conway <mail@joeconway.com>
Re: get_controlfile() can leak fds in the backend Joe Conway <mail@joeconway.com>
Re: get_controlfile() can leak fds in the backend Michael Paquier <michael@paquier.xyz>
Re: get_controlfile() can leak fds in the backend Joe Conway <mail@joeconway.com>
Re: get_controlfile() can leak fds in the backend Michael Paquier <michael@paquier.xyz>
Re: get_controlfile() can leak fds in the backend Joe Conway <mail@joeconway.com>
Re: get_controlfile() can leak fds in the backend Michael Paquier <michael@paquier.xyz>
Re: get_controlfile() can leak fds in the backend Fabien COELHO <coelho@cri.ensmp.fr>
Re: get_controlfile() can leak fds in the backend Andres Freund <andres@anarazel.de>
Re: get_controlfile() can leak fds in the backend Fabien COELHO <coelho@cri.ensmp.fr>
Re: get_controlfile() can leak fds in the backend Michael Paquier <michael@paquier.xyz>
Re: get_controlfile() can leak fds in the backend Andres Freund <andres@anarazel.de>
Re: get_controlfile() can leak fds in the backend Fabien COELHO <coelho@cri.ensmp.fr>
Re: get_controlfile() can leak fds in the backend Andres Freund <andres@anarazel.de>
Re: get_controlfile() can leak fds in the backend Michael Paquier <michael@paquier.xyz>
Re: get_controlfile() can leak fds in the backend Andres Freund <andres@anarazel.de>
Re: get_controlfile() can leak fds in the backend Fabien COELHO <coelho@cri.ensmp.fr>
Hi,

On 2019-02-27 10:23:45 +0100, Fabien COELHO wrote:
> However, while at it, there is also the question of whether the control file
> should be locked when updated, eg with flock(2) to avoid race conditions
> between concurrent commands. ISTM that there is currently not such thing in
> the code, but that it would be desirable.

Shouldn't be necessary - the control file fits into a single page, and
writes of that size ought to always be atomic. And I also think
introducing flock usage for this would be quite disproportional.

Greetings,

Andres Freund

В списке pgsql-hackers по дате отправления
От: Fabien COELHO
Дата:
От: Dmitry Dolgov
Дата:
FAQ