Re: pg_upgrade does not upgrade pg_stat_statements properly

On Thu, 29 Jul 2021 at 18:39, Bruce Momjian <bruce(at)momjian(dot)us> wrote:

> On Thu, Jul 29, 2021 at 06:29:11PM -0400, Tom Lane wrote:
> > Bruce Momjian <bruce(at)momjian(dot)us> writes:
> > > On Thu, Jul 29, 2021 at 06:19:56PM -0400, Tom Lane wrote:
> > >> I suggested awhile ago that pg_upgrade should look into
> > >> pg_available_extensions in the new cluster, and prepare
> > >> a script with ALTER EXTENSION UPDATE commands for
> > >> anything that's installed but is not the (new cluster's)
> > >> default version.
> >
> > > I can do that, but I would think a pg_dump/restore would also have this
> > > issue, so should this be more generic?
> >
> > No, because dump/restore does not have this issue. Regular pg_dump just
> > issues "CREATE EXTENSION" commands, so you automatically get the target
> > server's default version.
>
> Oh, so pg_upgrade does it differently so the oids are preserved?
>
>
I suspect this is part of --binary_upgrade mode

Dave

> --
> Bruce Momjian <bruce(at)momjian(dot)us> https://momjian.us
> EDB https://enterprisedb.com
>
> If only the physical world exists, free will is an illusion.
>
>