Re: Visibility bug with prepared transaction with subtransactions on standby

Alexander Lakhin <exclusion@gmail.com>

From: Alexander Lakhin <exclusion@gmail.com>

To: Heikki Linnakangas <hlinnaka@iki.fi>, pgsql-hackers <pgsql-hackers@postgresql.org>

Cc: Konstantin Knizhnik <knizhnik@neon.tech>

Date: 2024-12-16T08:00:00Z

Lists: pgsql-hackers

Hello Heikki,

27.06.2024 21:35, Heikki Linnakangas wrote:
> All I heard is crickets, so committed and backported to all supported versions.
>

Recently hornet made some noise too: [1], by failing on the test
modification added with e9c8747ee (in REL_13_STABLE):
# issuing query via background psql: SELECT count(*) FROM t_009_tbl_standby_mvcc
# pump_until: process terminated unexpectedly when searching for "(?^:background_psql: QUERY_SEPARATOR)" with stream: ""
query failed: psql:<stdin>:6: ERROR:  relation "t_009_tbl_standby_mvcc" does not exist

It looks like t_009_tbl_standby_mvcc had not arrived to standby yet when
it was queried.

I can reproduce the same with TEMP_CONFIG containing:
recovery_min_apply_delay = '500ms'

All the other tests (I ran check-world) pass with this setting. So maybe
this test lacks waiting for standby synchronization.

[1] https://buildfarm.postgresql.org/cgi-bin/show_log.pl?nm=hornet&dt=2024-12-13%2022%3A14%3A10

Best regards,
Alexander

Commits

Same data as JSON: GET /api/v1/messages/:b64id/commits the thread's linked commits as JSON, with link sources. API reference →

Make 009_twophase.pl test pass with recovery_min_apply_delay set
- e438b0f4241e 13.19 landed
- f5f5e7b47351 14.16 landed
- 2af2457fa6d8 15.11 landed
- 3231cb57d886 16.7 landed
- d1253c67fa29 17.3 landed
- 1dfeb6af7f87 18.0 landed
Fix MVCC bug with prepared xact with subxacts on standby
- 9dbf8ab48c21 14.13 landed
- 0e2f3d78bf5a 15.8 landed
- b5b418b689ec 16.4 landed
- cbfbda78413a 17.0 landed
- 5dea6628b32d 12.20 landed
- e9c8747ee934 13.16 landed
Rework handling of subtransactions in 2PC recovery
- 49e928154978 10.0 cited