| From: | Shinya Kato <shinya11(dot)kato(at)gmail(dot)com> |
|---|---|
| To: | Hüseyin Demir <huseyin(dot)d3r(at)gmail(dot)com> |
| Cc: | pgsql-hackers(at)lists(dot)postgresql(dot)org |
| Subject: | Re: Add autovacuum_warning to surface concurrent vacuum collisions |
| Date: | 2026-07-14 06:04:55 |
| Message-ID: | CAOzEurSGGFkSM40qX4O91U3o5pbbwi-WS7qruNdXZ1WkU4V84g@mail.gmail.com |
| Views: | Whole Thread | Raw Message | Download mbox | Resend email |
| Thread: | |
| Lists: | pgsql-hackers |
On Fri, Jun 26, 2026 at 11:29 PM Hüseyin Demir <huseyin(dot)d3r(at)gmail(dot)com> wrote:
> On any busy OLTP system with autovacuum_max_workers > 1, workers will skip tables held by other workers in every vacuum cycle. That is not a transient edge case and it is the steady state of a loaded database.
You are right. The skip is the mechanism working as intended, and on a
loaded system a per-table log line every cycle is noise rather than a
warning. The checkpoint_warning analogy does not hold either, since
that fires on a deviation from healthy behavior and rate-limits
itself, and a bare boolean does neither. I agree a per-table LOG is
the wrong surface for this.
--
Best regards,
Shinya Kato
NTT OSS Center
| From | Date | Subject | |
|---|---|---|---|
| Next Message | Shinya Kato | 2026-07-14 06:06:01 | Re: Add autovacuum_warning to surface concurrent vacuum collisions |
| Previous Message | Arne Roland | 2026-07-14 04:55:26 | Re: Key joins |