On Tue, 2023-10-31 at 08:23 +0100, Luca Ferrari wrote:
> What puzzled me was that such xmax was non-zero for a
> while (days). Inspecting pg_class.reloptions I found that accidentally
> (?) disabled autovacuum on such table, my fault!
> I should have checked pg_stat_user_tables at first!
VACUUM won't modify "xmax", nor will it remove the row.
Yours,
Laurenz Albe