Re: SUBSCRIPTIONS and pg_upgrade

Peter,

* Stephen Frost (sfrost(at)snowman(dot)net) wrote:
> * Peter Eisentraut (peter_e(at)gmx(dot)net) wrote:
> > Logical replication
> >
> > - Add PUBLICATION catalogs and DDL
> > - Add SUBSCRIPTION catalog and DDL
> > - Define logical replication protocol and output plugin
> > - Add logical replication workers
>
> I'm not entirely sure about the reasoning behind requiring a flag to
> include subscriptions in pg_dump output, as the documentation doesn't
> actually provide one, but if we are going to require that, shouldn't
> pg_upgrade use it, to make sure that the subscriptions are pulled
> forward to the upgraded cluster?
>
> Also, we should probably discuss that default in pg_dump to not include
> something in the database by default as that's not something we've ever
> done before. We made a very deliberate point to make sure that RLS
> didn't work by default with pg_dump to avoid the risk that we might not
> dump include everything in the database in the pg_dump output. I agree
> that it's not exactly the same, but even so, I was surprised to find
> out that subscriptions aren't included by default and I doubt I'd be
> alone.
>
> If this was all already discussed, I'm happy to go review the relevant
> thread(s). I'll admit that I didn't follow all of that discussion very
> closely, I'm just going through parts of pg_dump which are not being
> tested in our regression tests currently and discovered that dumping out
> subscriptions is not tested at all.

We should probably also throw an error if --include-subscriptions and
--data-only are used together, instead of just not dumping out the
subscriptions in that case, which is what happens now.

Surely, if the user asked for subscriptions to be included, we should
either throw an error or actually include them.

Thanks!

Stephen