Date: Tue, 29 Jun 2021 16:14:46 +0200
From: Boris Kolpackov <boris@codesynthesis.com>
To: Alvaro Herrera <alvaro.herrera@2ndquadrant.com>
Cc: pgsql-hackers@lists.postgresql.org
Subject: Re: Pipeline mode and PQpipelineSync()
Message-ID: <boris.20210629160323@codesynthesis.com>
References: <boris.20210624103805@codesynthesis.com>
 <202106291354.wyn7ar6emdm6@alvherre.pgsql>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <202106291354.wyn7ar6emdm6@alvherre.pgsql>
Organization: Code Synthesis
User-Agent: Mutt/1.5.24 (2015-08-30)
Archived-At: 
 <https://www.postgresql.org/message-id/boris.20210629160323%40codesynthesis.com>
Precedence: bulk

Alvaro Herrera <alvaro.herrera@2ndquadrant.com> writes:

> No luck reproducing any problems with this sequence as yet.

Can you try to recreate the call flow as implemented here (it's
pretty much plain old C if you ignore error handling):

https://git.codesynthesis.com/cgit/odb/libodb-pgsql/tree/odb/pgsql/statement.cxx?h=bulk#n789

Except replacing `continue` on line 966 with `break` (that will
make the code read-biased which I find triggers the error more
readily, though I was able to trigger it both ways).

Then in an explicit transaction send 500 prepared insert statements
(see previous email for details) with 250'th having a duplicate
primary key.