postgresql.conf evaluation of duplicate keys - Mailing list pgsql-general

From Martin Gerdes
Subject postgresql.conf evaluation of duplicate keys
Date
Msg-id 4F69DF67.5080005@dser.de
Whole thread Raw
Responses Re: postgresql.conf evaluation of duplicate keys  (David Kerr <dmk@mr-paradox.net>)
List pgsql-general
I've got a question relating to how the postgres configuration is parsed:

If I write into the following into postgresql.conf:
shared_buffers = 24MB
shared_buffers = 32MB

and start up postgres, the command 'show shared_buffers;' answers
'32MB'. That means the later value in the configuration file took
precedence. Which is great and what I would expect.
Now the question: Is this just a happy accident and could change at any
time, or is this documented and guaranteed behaviour (and if so, can you
point me to the relevant documentation)? I tried to search for it, but
my google foo is just not strong enough :-)

Answer to the (probably) inevitable question: "What are you trying to
achieve?" :
I am trying to define a set of defaults, which will get deployed to
multiple servers, and will be overwritten automatically. Yet I also want
to have the option to change individual parameters on one server. So I
want to introduce a line "#==do not modify above this line==".
Everything above the line will be replaced, everything below it left
untouched.
So if postgres is actually also parsing a file "postgresql.local" whose
values are guaranteed to take precedence over "postgresql.conf", that
would solve my problem as well.

Many thanks in advance!

Martin


pgsql-general by date:

Previous
From: Merlin Moncure
Date:
Subject: Re: POSTGRESQL Newbie
Next
From: Andrew Sullivan
Date:
Subject: software in domains (was: POSTGRESQL Newbie)