Re: Missing data_sync_elevel() for some calls of pg_fsync()?

Поиск
Список
Период
Сортировка
От Craig Ringer
Тема Re: Missing data_sync_elevel() for some calls of pg_fsync()?
Дата
Msg-id CAMsr+YGkAmdjhTwW1a4AbJsu-oLfMHcNeRW=jXqkQ6d=H3Ly3A@mail.gmail.com
обсуждение исходный текст
Ответ на Re: Missing data_sync_elevel() for some calls of pg_fsync()?  (Thomas Munro <thomas.munro@gmail.com>)
Список pgsql-hackers
Дерево обсуждения
On Mon, 2 Dec 2019 at 13:43, Thomas Munro <thomas.munro@gmail.com> wrote:

1.  Some code opens a file, writes some stuff to it, closes, and then
fsyncs it, and if that fails and and it ever retries it'll redo all of
those steps again.  We know that some kernels might have thrown away
the data, but we don't care about the copy in the kernel's cache
because we'll write it out again next time around.

Can we trust the kernel to be reporting the EIO or ENOSPC only from writeback buffers for the actual file we're fsync()ing though? Not from buffers it flushed while performing our fsync() request, failed to flush, and complained about?

I'm not confident I want to assume that.
--
 Craig Ringer                   http://www.2ndQuadrant.com/
 2ndQuadrant - PostgreSQL Solutions for the Enterprise

В списке pgsql-hackers по дате отправления:

Предыдущее
От: Thomas Munro
Дата:
Сообщение: Re: Missing data_sync_elevel() for some calls of pg_fsync()?
Следующее
От: Craig Ringer
Дата:
Сообщение: Re: Add a GUC variable that control logical replication