Horton

Re: Add support for specifying tables in pg_createsubscriber.

Shubham Khanna <khannashubham1197@gmail.com>

View thread
From: Shubham Khanna <khannashubham1197@gmail.com>
To: Peter Smith <smithpb2250@gmail.com>
Cc: "Zhijie Hou (Fujitsu)" <houzj.fnst@fujitsu.com>, vignesh C <vignesh21@gmail.com>, "Hayato Kuroda (Fujitsu)" <kuroda.hayato@fujitsu.com>, PostgreSQL Hackers <pgsql-hackers@lists.postgresql.org>
Date: 2025-09-15T05:24:19Z
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. Support existing publications in pg_createsubscriber.

Attachments

On Mon, Sep 15, 2025 at 6:01 AM Peter Smith <smithpb2250@gmail.com> wrote:
>
> Hi Shubham,
>
> IIUC the v6 will be rewritten to remove the new option, in favour of
> just redefining the --publication option behaviour.
>
> So, much of the current v6 will become obsolete. The below comment is
> just for one piece of code that I thought will survive the rewrite.
>
> ======
> src/bin/pg_basebackup/pg_createsubscriber.c
>
> setup_publisher:
>
> 1.
> + /*
> + * Check if publication already exists when
> + * --reuse-existing-publications is specified
> + */
> + if (opt->reuse_existing_pubs && check_publication_exists(conn,
> dbinfo[i].pubname, dbinfo[i].dbname))
> + {
> + pg_log_info("using existing publication \"%s\" in database \"%s\"",
> + dbinfo[i].pubname, dbinfo[i].dbname);
> + make_pub = false;
> + }
> +
>   /*
>   * Create publication on publisher. This step should be executed
>   * *before* promoting the subscriber to avoid any transactions between
>   * consistent LSN and the new publication rows (such transactions
>   * wouldn't see the new publication rows resulting in an error).
>   */
> - create_publication(conn, &dbinfo[i]);
> + if (make_pub)
> + {
> + create_publication(conn, &dbinfo[i]);
> + dbinfo[i].made_publication = true;
> + if (opt->reuse_existing_pubs)
> + pg_log_info("created publication \"%s\" in database \"%s\"",
> + dbinfo[i].pubname, dbinfo[i].dbname);
> + }
> + else
> + dbinfo[i].made_publication = false;
>
> I think there are still too many if/else here. This logic can be
> simplified like below:
>
> if (check_publication_exists(...))
> {
>   pg_log_info("using existing publication...");
>   dbinfo[i].made_publication = false;
> }
> else
> {
>   create_publication(conn, &dbinfo[i]);
>   pg_log_info("created publication ...");
>   dbinfo[i].made_publication = true;
> }
>

I have now removed the extra option and reworked the patch so that the
behavior is handled directly through the --publication option. This
way, the command will either reuse an existing publication (if it
already exists) or create a new one, making it simpler and more
intuitive for users.
The attached patch contains the suggested changes.

Thanks and regards,
Shubham Khanna.