6 PM
To:
Subject: Re: [GENERAL] pain of postgres upgrade with extensions
Greg Sabino Mullane wrote:
-BEGIN PGP SIGNED MESSAGE-
Hash: RIPEMD160
- dump version N database
- create empty version N+1 database
- install N+1's version of each needed contrib module into new databas
On Mar 12, 2008, at 3:19 PM, Tom Lane wrote:
- restore dump, ignoring "object already exists" errors
Couldn't one use the dump listing feature of pg_restore and comment
out the extensions when restoring? Not likely to be a big improvement
over "ignore" errors :-)
--
Sent via pgsql-gen
On Wed, Mar 12, 2008 at 07:34:03PM -, Greg Sabino Mullane wrote:
> > - dump version N database
> > - create empty version N+1 database
> > - install N+1's version of each needed contrib module into new database
> > - restore dump, ignoring "object already exists" errors
> >
> > There is a TODO
Greg Sabino Mullane wrote:
-BEGIN PGP SIGNED MESSAGE-
Hash: RIPEMD160
- dump version N database
- create empty version N+1 database
- install N+1's version of each needed contrib module into new database
- restore dump, ignoring "object already exists" errors
There is a TODO to fig
-BEGIN PGP SIGNED MESSAGE-
Hash: RIPEMD160
> - dump version N database
> - create empty version N+1 database
> - install N+1's version of each needed contrib module into new database
> - restore dump, ignoring "object already exists" errors
>
> There is a TODO to figure out some cleaner
paul rivers <[EMAIL PROTECTED]> writes:
> Is this something that wouldn't be fixed by:
> - dump 8.2 database
> - load dump into 8.3 database
> - for each extension, run the 8.2 drop extension script in 8.2's contrib
> - for each extension, run the 8.3 install extension script in 8.3's contrib
The
I noticed this immediately when using the PostgreSQL tool for examples of
dumps for creating export of database/table structure/data for the
MyJSQLView
application. I considered implementing a similar copy dump, but seems it
would
not be handled properly with SQL statements.
danap.
This is n
David Potts wrote:
This is not a flame about current or previous release of Postgres.
I have just gone through the awful experience of upgrading from Postgres
8.2 to 8.3 with a database that had one of the many Postgres extensions
included. The problem comes down to the way that Postgres extensi
