Add shared buffer hits to pg_stat_io

From: Melanie Plageman <melanieplageman(at)gmail(dot)com>
To: Pg Hackers <pgsql-hackers(at)postgresql(dot)org>
Cc: Maciek Sakrejda <m(dot)sakrejda(at)gmail(dot)com>, Justin Pryzby <pryzby(at)telsasoft(dot)com>, smilingsamay(at)gmail(dot)com, Kyotaro Horiguchi <horikyota(dot)ntt(at)gmail(dot)com>, Lukas Fittl <lukas(at)fittl(dot)com>, Magnus Hagander <magnus(at)hagander(dot)net>, Alvaro Herrera <alvherre(at)alvh(dot)no-ip(dot)org>
Subject: Add shared buffer hits to pg_stat_io
Date: 2023-02-25 20:16:40
Message-ID: CAAKRu_beMa9Hzih40=XPYqhDVz6tsgUGTrhZXRo=unp+szb=UA@mail.gmail.com
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-hackers

Hi,

As suggested in [1], the attached patch adds shared buffer hits to
pg_stat_io.

I remember at some point having this in the view and then removing it
but I can't quite remember what the issue was -- nor do I see a
rationale mentioned in the thread [2].

It might have had something to do with the interaction between the
now-removed "rejected" buffers column.

I am looking for input as to the order of this column in the view. I
think it should go after op_bytes since it is not relevant for
non-block-oriented IO. However, I'm not sure what the order of hits,
evictions, and reuses should be (all have to do with buffers).

While adding this, I noticed that I had made all of the IOOP columns
int8 in the view, and I was wondering if this is sufficient for hits (I
imagine you could end up with quite a lot of those).

- Melanie

[1] https://www.postgresql.org/message-id/20230209050319.chyyup4vtq4jzobq%40awork3.anarazel.de
[2] https://www.postgresql.org/message-id/flat/20230209050319.chyyup4vtq4jzobq%40awork3.anarazel.de#63ff7a97b7a5bb7b86c1a250065be7f9

Attachment Content-Type Size
v1-0001-Reorder-pgstatfuncs-local-enum.patch text/x-patch 1.0 KB
v1-0002-Add-IOOP_HIT-to-pg_stat_io.patch text/x-patch 11.5 KB

Responses

Browse pgsql-hackers by date

  From Date Subject
Next Message Regina Obe 2023-02-25 20:40:24 RE: Ability to reference other extensions by schema in extension scripts
Previous Message Andres Freund 2023-02-25 19:52:53 Re: stopgap fix for signal handling during restore_command