Re: Incorrect result of bitmap heap scan.

Peter Geoghegan <pg@bowt.ie>

From: Peter Geoghegan <pg@bowt.ie>

To: Andres Freund <andres@anarazel.de>

Cc: Matthias van de Meent <boekewurm+postgres@gmail.com>, Andrew Dunstan <andrew@dunslane.net>, Melanie Plageman <melanieplageman@gmail.com>, Thomas Munro <thomas.munro@gmail.com>, Konstantin Knizhnik <knizhnik@garret.ru>, PostgreSQL Hackers <pgsql-hackers@postgresql.org>, Alexander Kuzmenkov <akuzmenkov@timescale.com>, Tom Lane <tgl@sss.pgh.pa.us>, Heikki Linnakangas <hlinnaka@iki.fi>

Date: 2024-12-02T20:56:17Z

Lists: pgsql-hackers

Attachments

v1-0001-isolationtester-showing-broken-index-only-scans-w.patch (application/octet-stream) patch v1-0001

On Mon, Dec 2, 2024 at 2:43 PM Andres Freund <andres@anarazel.de> wrote:
> Attached is an isolationtest that reliably shows wrong query results.

Nice approach with the cursor!

I took what you wrote, and repurposed it to prove my old theory about
GiST index-only scans being broken due to the lack of an appropriate
interlock against concurrent TID recycling. See the attached patch.

-- 
Peter Geoghegan

Commits

Add test for HeapBitmapScan's broken skip_fetch optimization
- 24da5b239a4b 18.0 landed
Remove HeapBitmapScan's skip_fetch optimization
- 459e7bf8e2f8 18.0 landed
- b9ec8125d167 13.21 landed
- 4934d38759cd 14.18 landed
- 77d90d6d6334 15.13 landed
- 980727b84107 16.9 landed
- 78cb2466f752 17.5 landed
Allow bitmap scans to operate as index-only scans when possible.
- 7c70996ebf09 11.0 cited