On Mon, Sep 2, 2024 at 1:29 PM Melanie Plageman <melanieplage...@gmail.com> wrote: > I'll investigate more tomorrow, but based on my initial investigation, > there appears to be some interaction related to how much of the > relation is in shared buffers after creating the table and updating > it. If you set shared_buffers sufficiently high and prewarm the table > after the update, master has fewer WAL records reported by vacuum > verbose.
Fewer of what specific kind of WAL record? All of the details about useful work done by VACUUM were identical across versions. It was only the details related to WAL, buffers, and CPU time that changed. Perhaps I'm not thinking of something obvious. Maybe it's extra VISIBILITY records? But I'd expect the number of VISIBILITY records to match the number of pages frozen, given these particulars. VACUUM VERBOSE at least shows that that hasn't changed. -- Peter Geoghegan
