From: Chris Browne <cbbrowne@acm.org>
Subject: Re: "COPY TO stdout" statements occurrence in log files
Date: Fri, 14 Jan 2011 16:19:42 -0500
Organization: cbbrowne Computing Inc
Lines: 24
Message-ID: <8739ovdvyp.fsf@cbbrowne.afilias-int.info>
References: <AANLkTinUziwN+9fMRin0n3uKhfqB1B3WfX5gLyO9zq6f@mail.gmail.com>
	<AANLkTikF_tSo0EyDCWf5qkmPbuKdHe1Mi-+0O3Fvrq9u@mail.gmail.com>
Mime-Version: 1.0
Content-Type: text/plain
User-Agent: Gnus/5.13 (Gnus v5.13) Emacs/24.0.50 (gnu/linux)
Cancel-Lock: sha1:pTvD3AAQmbVqvo/SmCbAhIJPvJc=
To: pgsql-performance@postgresql.org

msakrejda@truviso.com (Maciek Sakrejda) writes:
>> Is this normal? I'm afraid because my application doesn't run this kind of
>> statement, so how can I know what is doing these commands? Maybe pg_dump?
>
> I think pg_dump is likely, yes, if you have that scheduled. I don't
> think anything in the log file will identify it as pg_dump explicitly
> (I believe as far as the server is concerned, pg_dump is just another
> client), but if you're concerned about this, you can add the client
> pid (%p) to log_line_prefix in postgresql.conf, log the pg_dump pid
> through whatever mechanism manages that, and compare.

That's an option...  More are possible...

1.  Our DBAs have been known to create users specifically for doing
backups ("dumpy").  It doesn't seem like a *huge* proliferation of users
to have some 'utility' user names for common processes.

2.  In 9.1, there will be a new answer, as there's a GUC to indicate the
"application_name".
-- 
"Programming today  is a race  between software engineers  striving to
build bigger and better  idiot-proof programs, and the Universe trying
to  produce  bigger  and  better  idiots.  So  far,  the  Universe  is
winning."  -- Rich Cook