Обсуждение: duplicate logging in pg_createsubscriber

Hi hackers,

While reviewing pg_createsubscriber logs (in another thread) I saw
some unexpected (almost) duplicate consecutive logs for the slot
creation in --dry-run mode.

e.g.
----------
pg_createsubscriber: creating the replication slot
"pg_createsubscriber_16386_e9a70df3" in database "db2"
pg_createsubscriber: create replication slot
"pg_createsubscriber_16386_e9a70df3" on publisher
----------

It didn't look right to me.

I found the code is doing:
- pg_log_info("creating the replication slot unconditional inside
create_logical_replication_slot
- pg_log_info("create replication slot...") immediately after call to
create_logical_replication_slot

Perhaps that 2nd log was once supposed to say "created" (past tense),
but even that seemed redundant.

Here is a small patch to remove the log duplication by keeping only
the log *within* the function.

Thoughts?

======
Kind Regards,
Peter Smith.
Fujitsu Asutralia

On Wed, Oct 08, 2025 at 05:29:20PM +1100, Peter Smith wrote:
> It didn't look right to me.
>
> I found the code is doing:
> - pg_log_info("creating the replication slot unconditional inside
> create_logical_replication_slot
> - pg_log_info("create replication slot...") immediately after call to
> create_logical_replication_slot
>
> Perhaps that 2nd log was once supposed to say "created" (past tense),
> but even that seemed redundant.
>
> Here is a small patch to remove the log duplication by keeping only
> the log *within* the function.
>
> Thoughts?

Agreed.  This was bugging me a bit today while looking at some logs
generated by the TAP tests.  We don't gain any additional information
here, so let's remove the duplicates.  I'll check that later, it's
late now.
--
Michael

On Thu, Oct 9, 2025 at 1:31 PM Chao Li <li.evan.chao@gmail.com> wrote:
>
>
>
> On Oct 8, 2025, at 14:29, Peter Smith <smithpb2250@gmail.com> wrote:
>
> Perhaps that 2nd log was once supposed to say "created" (past tense),
> but even that seemed redundant.
>
> Here is a small patch to remove the log duplication by keeping only
> the log *within* the function.
>
> Thoughts?
>
> ======
> Kind Regards,
> Peter Smith.
> Fujitsu Asutralia
> <v1-0001-remove-duplicate-logging-for-slot-creation.patch>
>
>
> The change is straightforward, especially I like the second change of adding “on publisher”, which makes the message
clearer.
>

Thanks.

> For dry-run mode, I see in setup_recovery(), it debug-logs a message with “# dry run mode recovery_target_lsn = xxx”,
Ithink it would be good to add similar debug logs in create_logical_replication_slot(), create_publication() and
drop_publication().
>

Yes, I am already attempting to address those other dry-run logging
issues in another thread [1].

======
[1]
https://www.postgresql.org/message-id/CAHut%2BPu84M8tgTjn%2B%2B4WCTETOX%3DNR4YmLn-%2B4B29UJeDTgoJ9Q%40mail.gmail.com

Kind Regards,
Peter Smith.
Fujitsu Australia

On Thu, Oct 09, 2025 at 01:36:36PM +1100, Peter Smith wrote:
> Thanks.

Done this one.
--
Michael

On Thu, Oct 9, 2025 at 4:03 PM Michael Paquier <michael@paquier.xyz> wrote:
>
> On Thu, Oct 09, 2025 at 01:36:36PM +1100, Peter Smith wrote:
> > Thanks.
>
> Done this one.
> --

Thanks for pushing!

======
Kind Regards,
Peter Smith.
Fujitsu Australia