Re: log_statement = 'mod' does not log all data modifying statements

Поиск
Список
Период
Сортировка
От Igor Neyman
Тема Re: log_statement = 'mod' does not log all data modifying statements
Дата
Msg-id A76B25F2823E954C9E45E32FA49D70ECCD515094@mail.corp.perceptron.com
обсуждение исходный текст
Ответ на Re: log_statement = 'mod' does not log all data modifying statements  (Stephen Frost <sfrost@snowman.net>)
Ответы Re: log_statement = 'mod' does not log all data modifying statements
Список pgsql-general

-----Original Message-----
From: pgsql-general-owner@postgresql.org [mailto:pgsql-general-owner@postgresql.org] On Behalf Of Stephen Frost
Sent: Wednesday, September 09, 2015 10:17 AM
To: Bruce Momjian <bruce@momjian.us>
Cc: Jack Christensen <jack@jackchristensen.com>; pgsql-general@postgresql.org
Subject: Re: [GENERAL] log_statement = 'mod' does not log all data modifying statements

* Bruce Momjian (bruce@momjian.us) wrote:
> On Fri, Jun 12, 2015 at 01:54:30PM -0500, Jack Christensen wrote:
> > I was recently surprised by changes that were not logged by
> > log_statement = 'mod'. After changing log_statement to 'all', I
> > found that the changes were occurring in a writable CTE.
> >
> > Is there a way to log all statements that update data?
>
> Not really.  Is this something we should document better?

pg_audit handles this (it is one of the specific goals of that project, in fact) and the case where modifications are
doneinside of DO blocks and PL code.  Further, it provides a great deal of additional flexibility. 

What we need here is better logging/auditing capabilities in core.
Improving the documentation for the back-branches is good but doesn't address this use-case.  It's unfortunate that we
don'thave a better answer at this time. 

Thanks!

Stephen

_______________________

Could you please provide reference to pg_audit?

TIA,
Igor Neyman



В списке pgsql-general по дате отправления:

Предыдущее
От: Florin Andrei
Дата:
Сообщение: avoid lock conflict between SELECT and TRUNCATE
Следующее
От: Pavel Suderevsky
Дата:
Сообщение: Re: Buffers: shared hit/read to shared_buffers dependence