Re: Vacuum statistics

Hi Karina,
Thank you for the review!  v44 is attached.

On Wed, Sep 23, 2026 at 8:57 AM Karina Litskevich
<litskevichkarina(at)gmail(dot)com> wrote:
> CI is complaining about more than one definition of
> <table id="extvacuumstatistics-pg-stats-vacuum-indexes-columns">

Fixed, the extra table is removed from 0007.

> If you add a new external function in
> visibilitymap.c, i.e., visibilitymap_clear_rel, you should also
> add it in the INTERFACE ROUTINES list in the beginning of the file.

Done.

> I'd also suggest that the variation of the visibilitymap_clear
> function taking a Relation should be the main one (and be called
> visibilitymap_clear), and the variant taking a RelFileLocator
> should be the additional one for those who don't have a Relation
> (and be called visibilitymap_clear_no_stats or something). I am
> not insisting, though.

I left it as is for now.  Master has just changed visibilitymap_clear()
to take a RelFileLocator, and I didn't want to change it back.  I can
rename it if others prefer.

I also changed one more thing in 0004, after a question from Vlada
Pogozheskaya <v(dot)pogozheskaya(at)postgrespro(dot)ru>. vacuum_interrupt_count is
now added to the database stats in pgstat_update_dbstats(), the same way
as xact_rollback, instead of in AtEOXact_PgStat_Database().
So nothing is done during transaction abort anymore.

--
Regards,
Alena Rybakina
Yandex