| From: | Peter Geoghegan <pg(at)bowt(dot)ie> |
|---|---|
| To: | Tomas Vondra <tomas(at)vondra(dot)me> |
| Cc: | Nikolay Samokhvalov <nik(at)postgres(dot)ai>, PostgreSQL Hackers <pgsql-hackers(at)lists(dot)postgresql(dot)org> |
| Subject: | Re: EXPLAIN: showing ReadStream / prefetch stats |
| Date: | 2026-09-22 15:35:29 |
| Message-ID: | CAH2-WzmLGVED5zgs5Sc-VfTFM-d=OdvYEbGyRsShDr5Q8ShW1w@mail.gmail.com |
| Views: | Whole Thread | Raw Message | Download mbox | Resend email |
| Thread: | |
| Lists: | pgsql-hackers |
On Tue, Sep 22, 2026 at 11:23 AM Tomas Vondra <tomas(at)vondra(dot)me> wrote:
> In other words, I/O and buffers may disagree - that's not a bug. Those
> counters are tracking different things.
>
> To handle this correctly, we'd need to find a way to not issue the I/O
> at all (since 862092932c9479b79732f3b441da05453ea5e06d we discard I/O
> that was not initiated yet).
Isn't this what Andres' experimental "Make read_stream_reset()/end()
not wait for IO" patch does? See:
This is one of the patches that I've been posting as part of the index
prefetching patch series for some time now.
--
Peter Geoghegan
| From | Date | Subject | |
|---|---|---|---|
| Next Message | Andres Freund | 2026-09-22 15:38:14 | Re: EXPLAIN: showing ReadStream / prefetch stats |
| Previous Message | Andres Freund | 2026-09-22 15:26:41 | Re: Teach pg_upgrade to deal with invalid databases |