Horton
Menu

Re: running logical replication as the subscription owner

Robert Haas <robertmhaas@gmail.com>

View thread
From: Robert Haas <robertmhaas@gmail.com>
To: Jelte Fennema <postgres@jeltef.nl>
Cc: Jeff Davis <pgsql@j-davis.com>, PostgreSQL-development <pgsql-hackers@postgresql.org>, Andres Freund <andres@anarazel.de>, Noah Misch <noah@leadboat.com>
Date: 2023-03-24T18:36:50Z
Lists: pgsql-hackers 
On Fri, Mar 24, 2023 at 2:14 PM Jelte Fennema <postgres@jeltef.nl> wrote:
> Yes, I totally agree. I now realise that wasn't clear at all from the wording in my previous email. I'm fine with both behaviours. I mainly meant that if we actually think the new behaviour is better (which honestly I'm not convinced of yet), then some follow up patch would probably be good. I definitely don't want to block this patch on any of that though. Both behaviors would be vastly better than the current one in my opinion. So if others wanted the behaviour in your patch, I'm completely fine with that.

Makes sense. I hope a few more people will comment on what they think
we should do here, especially Andres and Noah.

> > Yeah, if we stick with the current approach we should probably add
> > tests for that stuff.
>
> Even if we don't, we should still have tests showing that the security restrictions that we intend to put in place actually do their job.

Yeah, I just don't want to write the tests and then decide to change
the behavior and then have to write them over again. It's not so much
fun that I'm yearning to do it twice.

-- 
Robert Haas
EDB: http://www.enterprisedb.com

Commits

Same data as JSON: GET /api/v1/messages/:b64id/commits the thread's linked commits as JSON, with link sources. API reference →

  1. Honor run_as_owner option in tablesync worker.

  2. Document new pg_subscription columns.

  3. Add a run_as_owner option to subscriptions.

  4. Perform logical replication actions as the table owner.

  5. Add new predefined role pg_create_subscription.

  6. Respect permissions within logical replication.

  7. Empty search_path in logical replication apply worker and walsender.

  8. Empty search_path in Autovacuum and non-psql/pgbench clients.