From: Andrew Gierth <andrew@tao11.riddles.org.uk>
To: Peter Geoghegan <pg@bowt.ie>
Cc: PostgreSQL mailing lists <pgsql-bugs@lists.postgresql.org>,
  Thomas Munro <thomas.munro@gmail.com>,  iijima.yun@koumakan.jp
Subject: Re: BUG #17245: Index corruption involving deduplicated entries
In-Reply-To: 
 <CAH2-Wzn4mNEPbQ2Jrs=0tfF=iDHNpi7dSgPy2k1iHMdswR11MA@mail.gmail.com>
	(Peter Geoghegan's message of "Sun, 24 Oct 2021 14:52:17 -0700")
Message-ID: <87pmru6i9u.fsf@news-spur.riddles.org.uk>
References: <17245-ddf06aaf85735f36@postgresql.org>
	<CAH2-Wzn4mNEPbQ2Jrs=0tfF=iDHNpi7dSgPy2k1iHMdswR11MA@mail.gmail.com>
User-Agent: Gnus/5.13 (Gnus v5.13) Emacs/26.3 (berkeley-unix)
Date: Mon, 25 Oct 2021 01:56:36 +0100
MIME-Version: 1.0
Content-Type: text/plain
Archived-At: 
 <https://www.postgresql.org/message-id/87pmru6i9u.fsf%40news-spur.riddles.org.uk>
Precedence: bulk

>>>>> "Peter" == Peter Geoghegan <pg@bowt.ie> writes:

 Peter> PostgreSQL 14 is the first version where the default
 Peter> wal_sync_method is fdatasync() on FreeBSD -- though only with
 Peter> FreeBSD 13. See commit f900a79e. Perhaps that has something to
 Peter> do with the problem seen here.

Nope. The database in question was restored from a pg_dumpall backup
three weeks ago and I'm told there have been no crashes or even unclean
restarts since then.

 Peter> I suspect that there was TID recycling by VACUUM involved here.

That also doesn't seem likely from the heap page images I saw, which
suggested a fairly low update rate, HOT updates, and page cleanup rather
than much vacuum activity (in particular there were several LP_REDIRECT
pointers and few free pointers).

-- 
Andrew (irc:RhodiumToad)