From: Chris Browne <cbbrowne@acm.org>
Subject: Re: 2 PGSQL instances in the same server
Date: Thu, 04 Nov 2010 12:38:03 -0400
Organization: cbbrowne Computing Inc
Lines: 35
Message-ID: <87lj59rqms.fsf@cbbrowne.afilias-int.info>
References: <AANLkTik5huKGe7MK0UeEM4R=9dhF+84wDF8x=CWb-CU8@mail.gmail.com>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
User-Agent: Gnus/5.13 (Gnus v5.13) Emacs/23.2 (gnu/linux)
Cancel-Lock: sha1:T4sh0moaIffhj+9Dj2mnJL6ZHyE=
To: pgsql-admin@postgresql.org

rbarreca@gmail.com (Ramiro Barreca) writes:
> We need to have, for migration, either an instance of our actual PG 8.4.4 and a
> new one of PG 9.0 for testing.
> Where can we found a paper for helping us?
> Our platform is Centos 5.4 x86 64 bits

The "complication" to this is that it is quite likely not simple to do
this using pre-packaged (e.g. - RPM files) Postgres builds.

Anyone that's involved with building Postgres tends to have tooling
handy to help have multiple versions of Postgres around.

I have a script I use for this sort of thing:
    <https://github.com/cbbrowne/pginit>

Greg Smith has something analagous, though more sophisticated:
    <https://github.com/gregs1104/peg>

It amounts to:

  1.  Installing both versions of software;
  2.  Possibly using scripts to help manage where the results get
      installed.

If you're accustomed to entrusting installation totally to RPM-based
installations of Postgres, you're headed into new territory to have to
choose and configure where the installations are done.  But it's really
no big deal - it's not unusual for me to install several different
instances of Postgres in the course of an hour, refreshing some
instances I use to test Slony-I.
-- 
output = ("cbbrowne" "@" "linuxfinances.info")
"In the  free software world, a  rising tide DOES lift  all boats, and
once the  user has tasted  Unix it's easy  for them to  switch between
Unices." -- david parsons