On Sun, May 27, 2012 at 11:31 AM, Tom Lane <tgl@sss.pgh.pa.us> wrote:
> I don't recall exactly what problems drove us to make pg_upgrade do
> what it does with extensions, but we need a different fix for them.
Well, you need pg_upgrade to preserve the OIDs of objects that are
part of extensions just as you do for any other objects. If pg_dump
--binary-upgrade just emits "CREATE EXTENSION snarfle", for some
extension snarfle that provides an eponymous type, then the new
cluster is going to end up with a type with a different OID than than
whatever existed in the old cluster, and that's going to break all
sorts of things - e.g. arrays already on disk that contain the old
type OID.
I think that it would be an extremely fine thing if we could fix the
world so that we not preserve OIDs across a pg_upgrade; that would be
all kinds of wonderful. However, I think that's likely to be quite a
difficult project.
--
Robert Haas
EnterpriseDB: http://www.enterprisedb.com
The Enterprise PostgreSQL Company
