Re: promotion related handling in pg_sync_replication_slots()

From: Bertrand Drouvot <bertranddrouvot(dot)pg(at)gmail(dot)com>
To: shveta malik <shveta(dot)malik(at)gmail(dot)com>
Cc: "Zhijie Hou (Fujitsu)" <houzj(dot)fnst(at)fujitsu(dot)com>, Amit Kapila <amit(dot)kapila16(at)gmail(dot)com>, Bharath Rupireddy <bharath(dot)rupireddyforpostgres(at)gmail(dot)com>, pgsql-hackers <pgsql-hackers(at)postgresql(dot)org>
Subject: Re: promotion related handling in pg_sync_replication_slots()
Date: 2024-04-16 09:22:06
Message-ID: Zh5DPsa38T0NnJEe@ip-10-97-1-34.eu-west-3.compute.internal
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-hackers

Hi,

On Tue, Apr 16, 2024 at 02:06:45PM +0530, shveta malik wrote:
> On Tue, Apr 16, 2024 at 1:55 PM Zhijie Hou (Fujitsu)
> <houzj(dot)fnst(at)fujitsu(dot)com> wrote:
> > I personally feel adding the additional check for sync_replication_slots may
> > not improve the situation here. Because the GUC sync_replication_slots can
> > change at any point, the GUC could be false when performing this addition check
> > and is set to true immediately after the check, so It could not simplify the logic
> > anyway.
>
> +1.
> I feel doc and "cannot synchronize replication slots concurrently"
> check should suffice.
>
> In the scenario which Hou-San pointed out, if after performing the
> GUC check in SQL function, this GUC is enabled immediately and say
> worker is started sooner than the function could get chance to sync,
> in that case as well, SQL function will ultimately get error "cannot
> synchronize replication slots concurrently", even though GUC is
> enabled. Thus, I feel we should stick with samer error in all
> scenarios.

Okay, fine by me, let's forget about checking sync_replication_slots then.

Regards,

--
Bertrand Drouvot
PostgreSQL Contributors Team
RDS Open Source Databases
Amazon Web Services: https://aws.amazon.com

In response to

Responses

Browse pgsql-hackers by date

  From Date Subject
Next Message Aleksander Alekseev 2024-04-16 09:48:23 [PATCH] Replace magic constant 3 with NUM_MERGE_MATCH_KINDS
Previous Message Stefan Fercot 2024-04-16 08:55:08 Re: Add recovery to pg_control and remove backup_label