References: <871rr3ohbo.fsf@ars-thinkpad>
 <CA+fd4k70BXLTm-N6q18LrL=GbKtwY3-2++UVFw05SvFTkZgTyQ@mail.gmail.com>
User-agent: mu4e 1.1.0; emacs 26.0.50
From: Arseny Sher <a.sher@postgrespro.ru>
To: Masahiko Sawada <masahiko.sawada@2ndquadrant.com>
Cc: PostgreSQL Hackers <pgsql-hackers@postgresql.org>,
 Alvaro Herrera <alvherre@alvh.no-ip.org>
Subject: Re: logical copy_replication_slot issues
In-reply-to: 
 <CA+fd4k70BXLTm-N6q18LrL=GbKtwY3-2++UVFw05SvFTkZgTyQ@mail.gmail.com>
Date: Mon, 10 Feb 2020 17:01:25 +0300
Message-ID: <87v9oelex6.fsf@ars-thinkpad>
MIME-Version: 1.0
Content-Type: text/plain
Precedence: bulk


Masahiko Sawada <masahiko.sawada@2ndquadrant.com> writes:

> I've attached the draft patch fixing this issue but I'll continue
> investigating it more deeply.

There also should be a check that source slot itself has consistent
snapshot (valid confirmed_flush) -- otherwise it might be possible to
create not initialized slot which is probably not an error, but weird
and somewhat meaningless. Paranoically, this ought to be checked in both
src slot lookups.

With this patch it seems like the only thing
create_logical_replication_slot does is ReplicationSlotCreate, which
questions the usefulness of this function. On the second look,
CreateInitDecodingContext checks plugin sanity (ensures it exists), so
probably it's fine.


-- cheers, arseny