| From: | David Geier <geidav(dot)pg(at)gmail(dot)com> |
|---|---|
| To: | Ayoub Kazar <kazarayoub2004(at)gmail(dot)com> |
| Cc: | KAZAR Ayoub <ma_kazar(at)esi(dot)dz>, Tomas Vondra <tomas(at)vondra(dot)me>, Jakub Wartak <jakub(dot)wartak(at)enterprisedb(dot)com>, Pg Hackers <pgsql-hackers(at)postgresql(dot)org> |
| Subject: | Re: Add pg_stat_vfdcache view for VFD cache statistics |
| Date: | 2026-09-09 14:59:54 |
| Message-ID: | 00379c7c-734a-48f9-b828-d79c4a657d56@gmail.com |
| Views: | Whole Thread | Raw Message | Download mbox | Resend email |
| Thread: | |
| Lists: | pgsql-hackers |
>> v7-0001: looks good to me.
>>
>> v7-0003: I'm wondering if we still want cache_bytes in pg_stat_vfdcache,
>> now where it's exposed via pg_backend_memory_contexts. It seems to me
>> that other stats functionality also doesn't expose memory info that is
>> accessible via pg_backend_memory_contexts. But I'm not completely sure
>> what's best here.
>>
> pg_backend_memory_contexts is local to current backend session, so AFAIK
> there's no other way of getting cluster-wide vfd cache memory usage.
> Therefore we need cache_bytes for this?
My bad. I hadn't realized cache_bytes is for the entire cluster.
--
David Geier
| From | Date | Subject | |
|---|---|---|---|
| Next Message | Андрей Казаринов | 2026-09-09 15:03:09 | Re:[PATCH] Allow subquery pull-up past inlineable CTEs |
| Previous Message | Aleksander Alekseev | 2026-09-09 14:54:24 | Re: {heap,minimal}_expand_tuple() are unused |