From: | Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us> |
---|---|
To: | Arkhena <Arkhena(at)gmail(dot)com> |
Cc: | david(at)fetter(dot)org, andres(at)anarazel(dot)de, Pg Hackers <pgsql-hackers(at)postgresql(dot)org> |
Subject: | Re: [PATCH] Log CSV by default |
Date: | 2018-12-02 17:09:57 |
Message-ID: | 29693.1543770597@sss.pgh.pa.us |
Views: | Raw Message | Whole Thread | Download mbox | Resend email |
Thread: | |
Lists: | pgsql-hackers |
Arkhena <Arkhena(at)gmail(dot)com> writes:
> As a consultant with so many customers that simply install Postgres "out of
> the box", I think that changing log_destination to 'csvlog' and
> logging_collector to 'on' needs changing at least log_truncate_on_rotation
> (but certainly also log_filename and log_rotation_age) if we don't want my
> customers to run out of space sooner or later.
Yeah, that's another issue that we'd have to confront if we turn
logging_collector on by default: the default settings for log
rotation aren't exactly foolproof, and we'd need to make them so.
(Now having said that, the default setup *without* logging_collector
isn't exactly foolproof either, since wherever you send stderr to
will grow without bound. But maybe people are used to dealing with
that.)
Independently of changing logging_collector's default, perhaps we could
agree on changes to the rotation-related defaults so that it's less
risky to just turn logging_collector on without other changes?
regards, tom lane
From | Date | Subject | |
---|---|---|---|
Next Message | didier | 2018-12-02 17:39:56 | Re: [proposal] Add an option for returning SQLSTATE in psql error message |
Previous Message | Tom Lane | 2018-12-02 16:33:24 | Re: AllocSetContextCreate changes breake extensions |