Date: Thu, 11 Nov 2021 17:22:34 -0800
From: Noah Misch <noah@leadboat.com>
To: Michael Paquier <michael@paquier.xyz>, tgl@sss.pgh.pa.us
Cc: pgsql-hackers@postgresql.org
Subject: Re: XLogReadRecord() error in XlogReadTwoPhaseData()
Message-ID: <20211112012234.GA1020314@rfd.leadboat.com>
References: <20211107013157.GB790288@rfd.leadboat.com>
 <YYiqxtEqhgb5gIWL@paquier.xyz>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <YYiqxtEqhgb5gIWL@paquier.xyz>
User-Agent: Mutt/1.5.24 (2015-08-30)
Archived-At: 
 <https://www.postgresql.org/message-id/20211112012234.GA1020314%40rfd.leadboat.com>
Precedence: bulk

On Mon, Nov 08, 2021 at 01:42:46PM +0900, Michael Paquier wrote:
> On Sat, Nov 06, 2021 at 06:31:57PM -0700, Noah Misch wrote:
> > On Sun, Oct 24, 2021 at 04:35:02PM -0700, Noah Misch wrote:
> > > https://buildfarm.postgresql.org/cgi-bin/show_log.pl?nm=kittiwake&dt=2021-10-24%2012%3A01%3A10
> > > got an interesting v9.6 failure [...]:
> > > 
> > > 2021-10-24 14:25:29.263 CEST [34569:175] pgbench ERROR:  could not read two-phase state from xlog at 0/158F4E0
> > > 2021-10-24 14:25:29.263 CEST [34569:176] pgbench STATEMENT:  COMMIT PREPARED 'c1';

Tom Lane reported another instance today:
https://buildfarm.postgresql.org/cgi-bin/show_log.pl?nm=tadarida&dt=2021-11-11%2013%3A29%3A58

Each of the three failures happened on a sparc64 Debian+gcc machine.  I had
tried ~8000 iterations on thorntail, another sparc64 Debian+gcc animal,
without reproducing this.

> > As a first step, let's report the actual XLogReadRecord() error message.
> > Attached.
> 
> Good catch!  This looks good.

Pushed.