> This is *not* an oversight. It is by design. See commit b07642dbcd8d. The
> documentation says
I don't see in the commit message why inserts in an aborted transaction
must count towards n_ins_since_vacuum.
> Estimated number of rows inserted since this table was last vacuumed
>
> Those rows were actually inserted. They are physically in the data files. And
> that's what matters for the autovacuum algorithm.
Correct,but they are dead tuples that are physically in the files, and
are accounted for through
n_dead_tup and are then cleaned up based on the
autovacuum_vacuum_scale_factor|threshold
calculation.
The purpose of b07642dbcd8d is to trigger autovacuum for append only/mainly
workloads that don't generate dead tuples.
--
Sami Imseih
Amazon Web Services (AWS)
