Re: Versioning policy and pg_upgrade - Mailing list pgsql-www

From Bruce Momjian
Subject Re: Versioning policy and pg_upgrade
Date
Msg-id 201008170046.o7H0kGG20551@momjian.us
Whole thread Raw
In response to Re: Versioning policy and pg_upgrade  (Tom Lane <tgl@sss.pgh.pa.us>)
Responses Re: Versioning policy and pg_upgrade  (Josh Berkus <josh@agliodbs.com>)
List pgsql-www
Tom Lane wrote:
> Thom Brown <thom@linux.com> writes:
> > But I propose expanding it to read:
> 
> > "Major releases usually change the internal format of system tables
> > and data files. These changes are often complex, so we don't maintain
> > backward compatibility for data files. A dump/reload of the database
> > cluster is required for major upgrades for upgrading from 8.2 or
> > earlier.  For upgrading from version 8.3 and later, there is the
> > option to use pg_upgrade (also known as pg_migrator prior to version
> > 9.0), capable of in-place upgrades.  This means, for example, you can
> > upgrade any minor version of 8.3 to 8.4 without having to perform a
> > full backup and restore, resulting in a significantly faster upgrade."
> 
> > Too wordy?
> 
> I don't really think that pg_upgrade should be recommended for 8.3 to
> 8.4 upgrades.  There are too many limitations and it's not been well
> tested.  I don't even trust it very much yet for 8.4 to 9.0 updates...

Hey, I heard that!  ;-)

--  Bruce Momjian  <bruce@momjian.us>        http://momjian.us EnterpriseDB
http://enterprisedb.com
 + It's impossible for everything to be true. +


pgsql-www by date:

Previous
From: Thom Brown
Date:
Subject: Re: (news.postgresql.org) pgsql.hackers and pgsql.performance not updating
Next
From: Noah Misch
Date:
Subject: Re: (news.postgresql.org) pgsql.hackers and pgsql.performance not updating