Date: Tue, 9 Jul 2019 17:46:30 -0400
From: Alvaro Herrera <alvherre@2ndquadrant.com>
To: Andres Freund <andres@anarazel.de>
Cc: Michael Paquier <michael@paquier.xyz>, marko@joh.to,
	pgsql-bugs@lists.postgresql.org,
	Magnus Hagander <magnus@hagander.net>
Subject: Re: BUG #15888: Bogus "idle in transaction" state for logical
 decoding client after creating a slot
Message-ID: <20190709214630.GA15632@alvherre.pgsql>
MIME-Version: 1.0
Content-Type: text/plain; charset=iso-8859-1
Content-Disposition: inline
Content-Transfer-Encoding: 8bit
In-Reply-To: <20190709201557.y5yt57z2za2hu6l7@alap3.anarazel.de>
User-Agent: Mutt/1.9.4 (2018-02-28)
Precedence: bulk

On 2019-Jul-09, Andres Freund wrote:

> The problem doesn't exist in later versions, because there
> exec_replication_command() does an explicit
> 	/* Report to pgstat that this process is now idle */
> 	pgstat_report_activity(STATE_IDLE, NULL);
> That's due to
> 
> commit d02974e32e028fc078d8f5eca1d6a4516efb0aa6
> Author: Magnus Hagander <magnus@hagander.net>
> Date:   2017-12-29 16:19:51 +0100
> 
>     Properly set base backup backends to active in pg_stat_activity
> 
> which is post 9.6.  Not sure I really feel a huge appetite for whacking
> this around in the back branches.

Hmm ... so what that commit did is precisely to fix this bug.
Magnus thought at the time he was fixing a pg10 bug,
https://postgr.es/m/CABUevEwX4g8y=gmgfPzxFKS7gqjSYNR949Xc96OQm=YXJmh_Og@mail.gmail.com
but apparently now we see that the bug was older than that.  Maybe it's
okay to backpatch further?

(Booh to missing "Discussion:" tags in commit messages)

-- 
Álvaro Herrera                https://www.2ndQuadrant.com/
PostgreSQL Development, 24x7 Support, Remote DBA, Training & Services