Date: Tue, 25 Aug 2020 11:22:00 +0900
From: Michael Paquier <michael@paquier.xyz>
To: rams nalabolu <ramsveeru441@gmail.com>
Cc: pgsql-docs@lists.postgresql.org
Subject: Re: PostgreSQL switchover process
Message-ID: <20200825022200.GB2017@paquier.xyz>
References: 
 <CAHq7gtAW9CPeKsjxdHfJRvs=9==C8od1M4-fbxw+jgvRc9VE=w@mail.gmail.com>
MIME-Version: 1.0
Content-Type: multipart/signed; micalg=pgp-sha512;
	protocol="application/pgp-signature"; boundary="Yylu36WmvOXNoKYn"
Content-Disposition: inline
In-Reply-To: 
 <CAHq7gtAW9CPeKsjxdHfJRvs=9==C8od1M4-fbxw+jgvRc9VE=w@mail.gmail.com>
Precedence: bulk


--Yylu36WmvOXNoKYn
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline

On Mon, Aug 24, 2020 at 03:56:43PM -0400, rams nalabolu wrote:
> I did the following steps to perform switchover
> 1) ran a checkpoint on primary.
> 2) verified secondary status and it is up to date with primary.
> 3) stopped primary using pg_ctl -mf i.e force.

This is not a force mode, but the fast mode, where all existing
connections are forcibly stopped, and that the shutdown is clean, with
a shutdown checkpoint generated before finishing the shutdown
sequence.  During this shutdown, the primary makes sure that all
standbys have flushed WAL up to the point of the shutdown checkpoint.
Note that this makes rather unnecessary the checkpoint you ran on the
primary in step 1.

> 4) promoted primary using trigger file and verified the status of the new
> primary; it is not in recovery mode and running on new timeline ID 2.

I think that you mean promotion of the standby here.

> 5) I did create the recovery.conf file on old primary i.e new standby with
>       primaray_coninfo: new primary(B)
>       recovery_target_timeline='latest'
>       primary_slot_name: "slot I created in new_standby"

Typo here.  You mean primary_conninfo.

> 6) I started the new standby. It is up and running but it is not in
> recovery mode.

This switchover flow is a good base, so it should be possible to reuse
your previous primary as a standby.

> But the wal sender process on primary and receiver process on secondary are
> not running. i.e it looks like my replication is broken.
> what steps am I missing here without rebuilding a new standby i.e old
> primary I want the replication happen b/w my new primary and
> secondary.

It may be many things without more information.  Could you check
pg_stat_replication on the primary and pg_stat_wal_receiver on the
standby?  Most likely something is wrong with primary_conninfo, but
the logs of the standby should have enough information to let you know
what happened.  Another thing you are not telling is the version of
PostgreSQL you are using here.  In 12 and newer versions,
support for recovery.conf has been removed, requiring roughly the
creation of standby.signal with all recovery parameters set in
postgresql.conf if you want to set up a standby.
--
Michael

--Yylu36WmvOXNoKYn
Content-Type: application/pgp-signature; name="signature.asc"

-----BEGIN PGP SIGNATURE-----
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=tbEu
-----END PGP SIGNATURE-----

--Yylu36WmvOXNoKYn--