Feedback-ID: ia2694551:Fastmail
Date: Thu, 26 Jan 2023 10:42:41 +0100
From: Alvaro Herrera <alvherre@alvh.no-ip.org>
To: "Drouvot, Bertrand" <bertranddrouvot.pg@gmail.com>
Cc: PostgreSQL Hackers <pgsql-hackers@lists.postgresql.org>
Subject: Re: Helper functions for wait_for_catchup() in Cluster.pm
Message-ID: <20230126094241.yjofprg2xcu43kop@alvherre.pgsql>
MIME-Version: 1.0
Content-Type: text/plain; charset=utf-8
Content-Disposition: inline
Content-Transfer-Encoding: 8bit
In-Reply-To: <606daf52-fcc8-bc5e-6e7a-75d9f7c29532@gmail.com>
Archived-At: <https://www.postgresql.org/message-id/20230126094241.yjofprg2xcu43kop%40alvherre.pgsql>
Precedence: bulk

On 2023-Jan-26, Drouvot, Bertrand wrote:

> On 1/24/23 7:27 PM, Alvaro Herrera wrote:

> > 1. I don't think wait_for_write_catchup is necessary, because
> > calling wait_for_catchup() and omitting the 'mode' and 'lsn' arguments
> > would already do the same thing.  So what we should do is patch places
> > that currently give those two arguments, so that they don't.
> 
> Agree but there is one place where the node passed as the second argument is not the "$self":
> 
> src/bin/pg_rewind/t/007_standby_source.pl:$node_b->wait_for_write_catchup('node_c', $node_a);
> 
> So it looks like there is still a need for wait_for_write_catchup().

Hmm, I think that one can use the more general wait_for_catchup.


> > 2. Because wait_for_replay_catchup is an instance method, passing the
> > second node as argument is needlessly noisy, because that's already
> > known as $self.  So we can just say
> > 
> >    $primary_node->wait_for_replay_catchup($standby_node);
> 
> Yeah, but same here, there is places where the node passed as the second argument is not the "$self":
> 
> src/bin/pg_rewind/t/007_standby_source.pl:$node_b->wait_for_replay_catchup('node_c', $node_a);
> src/test/recovery/t/001_stream_rep.pl:$node_standby_1->wait_for_replay_catchup($node_standby_2, $node_primary);
> src/test/recovery/t/001_stream_rep.pl:$node_standby_1->wait_for_replay_catchup($node_standby_2, $node_primary);
> src/test/recovery/t/001_stream_rep.pl:  $node_standby_1->wait_for_replay_catchup($node_standby_2, $node_primary);
> 
> So it looks like there is still a need for wait_for_replay_catchup() with 2 parameters.

Ah, cascading replication.  In that case, let's make the second
parameter optional.  If it's not given, $self is used.

-- 
Álvaro Herrera               48°01'N 7°57'E  —  https://www.EnterpriseDB.com/
"En las profundidades de nuestro inconsciente hay una obsesiva necesidad
de un universo lógico y coherente. Pero el universo real se halla siempre
un paso más allá de la lógica" (Irulan)