Feedback-ID: i0c21471d:Fastmail
MIME-Version: 1.0
Date: Tue, 30 Jul 2024 18:02:51 -0300
From: "Euler Taveira" <euler@eulerto.com>
To: "Masahiko Sawada" <sawada.mshk@gmail.com>,
 "Bruce Momjian" <bruce@momjian.us>
Cc: PostgreSQL-documentation <pgsql-docs@postgresql.org>,
 "Peter Eisentraut" <peter@eisentraut.org>
Message-Id: <1802dbe8-d136-4ea3-bc31-bffc41b2cd27@app.fastmail.com>
In-Reply-To: 
 <CAD21AoCKqU8GPoAKA3V_-JD40tXR6VndekCV0Cx_9RMsGRVoYQ@mail.gmail.com>
References: <ZqX_4J-nFTQtmj6K@momjian.us>
 <CAD21AoCKqU8GPoAKA3V_-JD40tXR6VndekCV0Cx_9RMsGRVoYQ@mail.gmail.com>
Subject: Re: pg_createsubscriber clarification
Content-Type: multipart/mixed;
 boundary=756147858f7b40899293d40737b9fb91
Archived-At: 
 <https://www.postgresql.org/message-id/1802dbe8-d136-4ea3-bc31-bffc41b2cd27%40app.fastmail.com>
Precedence: bulk

--756147858f7b40899293d40737b9fb91
Content-Type: multipart/alternative;
 boundary=183911b403c245268f331bbd400ed497

--183911b403c245268f331bbd400ed497
Content-Type: text/plain
Content-Transfer-Encoding: 7bit

On Sun, Jul 28, 2024, at 1:17 PM, Masahiko Sawada wrote:
> I think the term "synchronization phase" comes from the description in
> "29.8.1. Initial Snapshot" section[1].

Yes.

> https://www.postgresql.org/docs/17/logical-replication-architecture.html#LOGICAL-REPLICATION-SNAPSHOT
> 
> How about the following?
> 
> The main difference between the logical replication setup and
> pg_createsubscriber is how they synchronize table data.
> pg_createsubscriber doesn't copy the initial table data because it
> uses the tables with their initial data on the target server. It only
> does the synchronization phase, which ensures each table is brought up
> to a synchronized state by applying changes using standard logical
> replication.

I slightly modified your proposal in the attached patch.

Thoughts?


--
Euler Taveira
EDB   https://www.enterprisedb.com/

--183911b403c245268f331bbd400ed497
Content-Type: text/html
Content-Transfer-Encoding: quoted-printable

<!DOCTYPE html><html><head><title></title><style type=3D"text/css">p.Mso=
Normal,p.MsoNoSpacing{margin:0}</style></head><body><div>On Sun, Jul 28,=
 2024, at 1:17 PM, Masahiko Sawada wrote:<br></div><blockquote type=3D"c=
ite" id=3D"qt" style=3D""><div>I think the term "synchronization phase" =
comes from the description in<br></div><div>"29.8.1. Initial Snapshot" s=
ection[1].<br></div></blockquote><div><br></div><div>Yes.</div><div><br>=
</div><blockquote type=3D"cite" id=3D"qt" style=3D""><div><a href=3D"htt=
ps://www.postgresql.org/docs/17/logical-replication-architecture.html#LO=
GICAL-REPLICATION-SNAPSHOT">https://www.postgresql.org/docs/17/logical-r=
eplication-architecture.html#LOGICAL-REPLICATION-SNAPSHOT</a><br></div><=
div><br></div><div>How about the following?<br></div><div><br></div><div=
>The main difference between the logical replication setup and<br></div>=
<div>pg_createsubscriber is how they synchronize table data.<br></div><d=
iv>pg_createsubscriber doesn't copy the initial table data because it<br=
></div><div>uses the tables with their initial data on the target server=
. It only<br></div><div>does the synchronization phase, which ensures ea=
ch table is brought up<br></div><div>to a synchronized state by applying=
 changes using standard logical<br></div><div>replication.<br></div></bl=
ockquote><div><br></div><div>I slightly modified your proposal in the at=
tached patch.</div><div><br></div><div>Thoughts?<br></div><div><br></div=
><div><br></div><div id=3D"sig112409080"><div class=3D"signature">--<br>=
</div><div class=3D"signature">Euler Taveira<br></div><div class=3D"sign=
ature">EDB&nbsp;&nbsp; <a href=3D"https://www.enterprisedb.com/">https:/=
/www.enterprisedb.com/</a><br></div></div><div><br></div></body></html>
--183911b403c245268f331bbd400ed497--

--756147858f7b40899293d40737b9fb91
Content-Disposition: attachment; filename="doc.patch"
Content-Type: text/x-patch; name="doc.patch"
Content-Transfer-Encoding: base64
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--756147858f7b40899293d40737b9fb91--