Date: Tue, 11 Jan 2022 19:37:47 +0800
From: Julien Rouhaud <rjuju123@gmail.com>
To: RKN Sai Krishna <rknsaiforpostgres@gmail.com>
Cc: pgsql-hackers@lists.postgresql.org
Subject: Re: Query regarding replication slots
Message-ID: <20220111113747.7mk4stfbxo4zobqc@jrouhaud>
References: <CAMVpbFMt=jb+Krk2AWUh3+YQYOSEfTD0D+vhhNGK7f2Vz4_edA@mail.gmail.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <CAMVpbFMt=jb+Krk2AWUh3+YQYOSEfTD0D+vhhNGK7f2Vz4_edA@mail.gmail.com>
Archived-At: <https://www.postgresql.org/message-id/20220111113747.7mk4stfbxo4zobqc%40jrouhaud>
Precedence: bulk

On Tue, Jan 11, 2022 at 04:48:59PM +0530, RKN Sai Krishna wrote:
> Hi All,
> 
> I have a very basic question related to replication slots. Why should
> the master/primary server maintain the replication slot info like lsn
> corresponding to each standby server etc. Instead, why can't each
> standby server send the lsn that it needs, and master/primary server
> maintain the minimum lsn across all of the standby servers so that the
> information could be used for cleanup/removal of WAL segments?

Because the information is needed even if the standby servers are not
available, and if you only have a global xmin then you can't do much when
removing one of the slots.