Horton

Re: Support logical replication of DDLs

Zheng Li <zhengli10@gmail.com>

View thread
From: Zheng Li <zhengli10@gmail.com>
To: Amit Kapila <amit.kapila16@gmail.com>
Cc: Tom Lane <tgl@sss.pgh.pa.us>, vignesh C <vignesh21@gmail.com>, Ajin Cherian <itsajin@gmail.com>, "houzj.fnst@fujitsu.com" <houzj.fnst@fujitsu.com>, "wangw.fnst@fujitsu.com" <wangw.fnst@fujitsu.com>, Runqi Tian <runqidev@gmail.com>, Peter Smith <smithpb2250@gmail.com>, li jie <ggysxcq@gmail.com>, Dilip Kumar <dilipbalaut@gmail.com>, Alvaro Herrera <alvherre@alvh.no-ip.org>, Masahiko Sawada <sawada.mshk@gmail.com>, Japin Li <japinli@hotmail.com>, rajesh singarapu <rajesh.rs0541@gmail.com>, PostgreSQL Hackers <pgsql-hackers@lists.postgresql.org>
Date: 2023-03-29T16:53:31Z
Lists: pgsql-hackers

Commits

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

  1. Add a run_as_owner option to subscriptions.

  2. Refactor pgoutput_change().

  3. Print the correct aliases for DML target tables in ruleutils.

  4. Fix object identity string for transforms

  5. Add grantable MAINTAIN privilege and pg_maintain role.

  6. Get rid of recursion-marker values in enum AlterTableType

  7. Release cache tuple when no longer needed

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

  9. Refactor format_type APIs to be more modular

  10. Use wrappers of PG_DETOAST_DATUM_PACKED() more.

On Wed, Mar 29, 2023 at 5:13 AM Amit Kapila <amit.kapila16@gmail.com> wrote:
>
> On Wed, Mar 29, 2023 at 2:49 AM Zheng Li <zhengli10@gmail.com> wrote:
> >
> >
> > I agree that a full fledged DDL deparser and DDL replication is too
> > big of a task for one patch. I think we may consider approaching this
> > feature in the following ways:
> > 1. Phased development and testing as discussed in other emails.
> > Probably support table commands first (as they are the most common
> > DDLs), then the other commands in multiple phases.
> > 2. Provide a subscription option to receive the DDL change, raise a
> > notice and to skip applying the change. The users can listen to the
> > DDL notice and implement application logic to apply the change if
> > needed. The idea is we can start gathering user feedback by providing
> > a somewhat useful feature (compared to doing nothing about DDLs), but
> > also avoid heading straight into the potential footgun situation
> > caused by automatically applying any mal-formatted DDLs.
> >
>
> Doesn't this mean that we still need to support deparsing of such DDLs
> which is what I think we don't want?

I think we can send the plain DDL command string and the search_path
if we don't insist on applying it in the first version. Maybe the
deparser can be integrated later when we're confident that it's
ready/subset of it is ready.

> > 3. As cross-version DDL syntax differences are expected to be uncommon
> > (in real workload), maybe we can think about other options to handle
> > such edge cases instead of fully automating it? For example, what
> > about letting the user specify how a DDL should be replicated on the
> > subscriber by explicitly providing two versions of DDL commands in
> > some way?
> >
>
> As we are discussing in another related thread [1], if
> publisher_version > subscriber_version then it may not be possible to
> form a DDL at publisher which can be applied to the subscriber. OTOH,
> we need to think if there could be any problems with publisher_version
> < subscriber_version setup, and if so, what we want to do for it.
> Once, we have a clear answer to that then I think we will be in a
> better position to answer your question.
> [1] - https://www.postgresql.org/message-id/OS0PR01MB5716088E497BDCBCED7FC3DA94849%40OS0PR01MB5716.jpnprd01.prod.outlook.com

Regards,
Zane