Horton

Re: BUG: Postgres 14 + vacuum_defer_cleanup_age + FOR UPDATE + UPDATE

Andres Freund <andres@anarazel.de>

View thread View on postgr.es
From: Andres Freund <andres@anarazel.de>
To: Mark Dilger <mark.dilger@enterprisedb.com>
Cc: Heikki Linnakangas <hlinnaka@iki.fi>, Andrey Borodin <amborodin86@gmail.com>, Michail Nikolaev <michail.nikolaev@gmail.com>, Thomas Munro <thomas.munro@gmail.com>, Matthias van de Meent <boekewurm+postgres@gmail.com>, PostgreSQL Hackers <pgsql-hackers@postgresql.org>, Mark Dilger <mark.dilger@enterprisedb.com>
Date: 2023-03-09T00:15:58Z
Lists: pgsql-hackers

Attachments

Hi,

On 2023-02-06 13:02:05 -0800, Andres Freund wrote:
> I didn't quite feel confident pushing a fix for this just before a minor
> release, so I'll push once the minor releases are tagged. A quite minimal fix
> to GetFullRecentGlobalXmin() in 12-13 (returning FirstNormalTransactionId if
> epoch == 0 and RecentGlobalXmin > nextxid_xid), and the slightly larger fix in
> 14+.

Pushed that.


Mark:

I worked some more on the fixes for amcheck, and fixes for amcheck.

The second amcheck fix ends up correcting some inaccurate output in the tests
- previously xids from before xid 0 were reported to be in the future.

Previously there was no test case exercising exceeding nextxid, without
wrapping around into the past. I added that at the end of
004_verify_heapam.pl, because renumbering seemed too annoying.

What do you think?


Somewhat random note:

Is it intentional that we VACUUM FREEZE test ROWCOUNT times? That's
effectively O(ROWCOUNT^2), albeit with small enough constants to not really
matter. I don't think we need to insert the rows one-by-one either. Changing
that to a single INSERT and FREEZE shaves 10-12% off the tests.  I didn't
change that, but we also fire off a psql for each tuple for heap_page_items(),
with offset $N no less. That seems to be another 500ms.

Greetings,

Andres Freund

Commits

  1. Fix incorrect TAP test ordering

  2. pg_amcheck: Minor test speedups

  3. amcheck: Fix FullTransactionIdFromXidAndCtx() for xids before epoch 0

  4. amcheck: Fix ordering bug in update_cached_xid_range()

  5. Fix corruption due to vacuum_defer_cleanup_age underflowing 64bit xids

  6. Add hardening to catch invalid TIDs in indexes.

  7. Use full 64-bit XID for checking if a deleted GiST page is old enough.

  8. Avoid early reuse of btree pages, causing incorrect query results.