Hello Masao-san and Melanie,
Thank you for your attention to this!
23.03.2025 19:26, Fujii Masao wrote:
With autovacuum = off, all of these fluctuations go away.
So you are suggesting disabling autovacuum during the postgres_fdw regression
test?
No, I'd prefer to have the test stable with working autovacuum. I just
wanted to point out that this failure is definitely caused by autovacuum
despite the lack of corresponding messages in the postmaster log.
Just my idea to stabilize the test with "RETURNING *" is to use WITH, like this:
WITH tmp AS (UPDATE ... RETURNING *) SELECT * FROM tmp ORDER BY ...
Thought?
Yeah, maybe such a trick will do.
23.03.2025 19:30, Melanie Plageman wrote:
On Sun, Mar 23, 2025 at 10:00 AM Alexander Lakhin <exclus...@gmail.com> wrote:
With autovacuum = off, all of these fluctuations go away.
If autovacuum is somehow corrupting the table, then isn't there a bug
that needs to be fixed? Also, are you only able to reproduce this on
master?
Yes, I reproduced this on master, but I don't think the table is
corrupted in this case, it just contains tuples in other order.
Best regards,
Alexander Lakhin
Neon (https://neon.tech)
