Re: pgBufferUsage.blk_{read|write}_time are zero although there are pgBufferUsage.local_blks_{read|written} - Mailing list pgsql-hackers

From Nazir Bilal Yavuz
Subject Re: pgBufferUsage.blk_{read|write}_time are zero although there are pgBufferUsage.local_blks_{read|written}
Date
Msg-id CAN55FZ3tiZjYozcuN-8V+h2WiCduWo3geTJtiWv3LEnYnUQe3Q@mail.gmail.com
Whole thread Raw
In response to Re: pgBufferUsage.blk_{read|write}_time are zero although there are pgBufferUsage.local_blks_{read|written}  (Robert Haas <robertmhaas@gmail.com>)
Responses Re: pgBufferUsage.blk_{read|write}_time are zero although there are pgBufferUsage.local_blks_{read|written}
List pgsql-hackers
Hi,

On Tue, 3 Oct 2023 at 19:44, Robert Haas <robertmhaas@gmail.com> wrote:
>
> Given that, I'm inclined to agree that this is a bug. But we might
> need to go through and make sure all of the code that deals with these
> counters is on the same page about what the values represent. Maybe
> there is code lurking somewhere that thinks these counters are only
> supposed to include "shared" rather than, as the fragment above
> suggests, "shared/local".

Thank you for the guidance.

What do you think about the second patch, counting extend calls'
timings in blk_write_time? In my opinion if something increments
{shared|local}_blks_written, then it needs to be counted in
blk_write_time too. I am not sure why it is decided like that.

Regards,
Nazir Bilal Yavuz
Microsoft



pgsql-hackers by date:

Previous
From: "Hayato Kuroda (Fujitsu)"
Date:
Subject: RE: [PoC] pg_upgrade: allow to upgrade publisher node
Next
From: Amit Kapila
Date:
Subject: Re: [PoC] pg_upgrade: allow to upgrade publisher node