Re: WIP patch - INSERT-able log statements - Mailing list pgsql-patches

From Greg Smith
Subject Re: WIP patch - INSERT-able log statements
Date
Msg-id Pine.GSO.4.64.0702212335370.29101@westnet.com
Whole thread Raw
In response to Re: WIP patch - INSERT-able log statements  ("FAST PostgreSQL" <fastpgs@fast.fujitsu.com.au>)
List pgsql-patches
On Thu, 22 Feb 2007, FAST PostgreSQL wrote:

> As we are triggering the sql output in log_destination, if the user
> gives 'syslog,sql' as options he is going to get two different looking
> logs (in terms of contents) depending upon his settings.

Yes, exactly; it's a good thing.  People add and remove things from the
text logs to make them easier to read.  It's one of the reasons they're
harder to process.  Since readability isn't a requirement for the SQL
formatted ones, you can pack a lot more into there and make it available
easily anyway.

I keep having every part of this conversation twice, so here's take two on
this one.  The things that people want out of the text logs are not
necessarily the same things they want from the SQL ones.  For example, I
have a situation where the semantics of the syslog output is being driven
by Sarbanes-Oxley related mechanics.  But the SQL logs are be based on my
requirements, which is to include close enough to everything that it might
as well be the whole set, in case I forgot something I find I need later.
The SQL logs are *completely* different from the syslog setup.

--
* Greg Smith gsmith@gregsmith.com http://www.gregsmith.com Baltimore, MD

pgsql-patches by date:

Previous
From: Alvaro Herrera
Date:
Subject: Re: [previously on HACKERS] "Compacting" a relation
Next
From: Peter Eisentraut
Date:
Subject: Re: [HACKERS] BLCKSZ fun facts