Horton

Re: HOT chain validation in verify_heapam()

Andres Freund <andres@anarazel.de>

View thread View on postgr.es
From: Andres Freund <andres@anarazel.de>
To: Peter Geoghegan <pg@bowt.ie>
Cc: pgsql-hackers@lists.postgresql.org, Himanshu Upadhyaya <upadhyaya.himanshu@gmail.com>, Robert Haas <robertmhaas@gmail.com>, Aleksander Alekseev <aleksander@timescale.com>, PostgreSQL Hackers <pgsql-hackers@postgresql.org>
Date: 2022-11-16T06:55:33Z
Lists: pgsql-hackers 
Hi,

On 2022-11-14 15:07:05 -0800, Peter Geoghegan wrote:
> I'd really like to know if the scary HOT chain freezing scenario is
> possible, for the very obvious reason. Have you tried to write a test
> case for that?

I tried. Unfortunately, even if the bug exists, we currently don't have the
infrastructure to write isolationtester tests for it. There's just too many
points where we'd need to wait where I don't know of ways to wait with
isolationtester.

I'm quite certain that it's possible to end up freezing an earlier row
versions in a hot chain in < 14, I got there with careful gdb
orchestration. Of course possible I screwed something up, given I did it once,
interactively. Not sure if trying to fix it is worth the risk of backpatching
all the necessary changes to switch to the retry approach.

Greetings,

Andres Freund

Commits

  1. amcheck: Generalize one of the recently-added update chain checks.

  2. amcheck: Tighten up validation of redirect line pointers.

  3. amcheck: Fix verify_heapam for tuples where xmin or xmax is 0.

  4. amcheck: Fix a few bugs in new update chain validation.

  5. Fix new test case to work on (some?) big-endian architectures.

  6. Don't test HEAP_XMAX_INVALID when freezing xmax.