Date: Thu, 23 Jul 2020 17:49:09 -0400
From: Alvaro Herrera <alvherre@2ndquadrant.com>
To: Andres Freund <andres@anarazel.de>
Cc: pgsql-hackers@postgresql.org, Peter Geoghegan <pg@bowt.ie>
Subject: Re: heap_abort_speculative() sets xmin to Invalid* without
 HEAP_XMIN_INVALID
Message-ID: <20200723214909.GA19828@alvherre.pgsql>
MIME-Version: 1.0
Content-Type: text/plain; charset=iso-8859-1
Content-Disposition: inline
Content-Transfer-Encoding: 8bit
In-Reply-To: <20200723194042.bygzsr6m23pwwkhd@alap3.anarazel.de>
User-Agent: Mutt/1.10.1 (2018-07-13)
Precedence: bulk

On 2020-Jul-23, Andres Freund wrote:

> I think we should change heap_abort_speculative() to set
> HEAP_XMIN_INVALID in master.

+1

> But we can't really do anything about
> existing tuples without it - therefore we will have to forever take care
> about encountering that combination :(.
> 
> Perhaps we should instead or additionally make
> HeapTupleHeaderXminInvalid() explicitly check for InvalidTransactionId?

+1 for doing it as an additional fix, with a fat comment somewhere
explaining where such tuples would come from.

Additionally, but perhaps not very usefully, maybe we could have a
mechanism to inject such tuples so that code can be hardened against the
condition.

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