Horton

Re: pgBufferUsage.blk_{read|write}_time are zero although there are pgBufferUsage.local_blks_{read|written}

Michael Paquier <michael@paquier.xyz>

View thread
From: Michael Paquier <michael@paquier.xyz>
To: Robert Haas <robertmhaas@gmail.com>
Cc: Melanie Plageman <melanieplageman@gmail.com>, Nazir Bilal Yavuz <byavuz81@gmail.com>, pgsql-hackers <pgsql-hackers@postgresql.org>
Date: 2023-10-19T02:25:13Z
Lists: pgsql-hackers

Commits

Same data as JSON: GET /api/v1/messages/:b64id/commits the thread's linked commits as JSON, with link sources. API reference →

  1. Fix description of I/O timing info for shared buffers in EXPLAIN (BUFFERS)

  2. pg_stat_statements: Add local_blk_{read|write}_time

  3. Add local_blk_{read|write}_time I/O timing statistics for local blocks

  4. Rename I/O timing statistics columns to shared_blk_{read|write}_time

  5. Count write times when extending relation files for shared buffers

  6. Add JIT deform_counter

On Tue, Oct 03, 2023 at 12:44:36PM -0400, Robert Haas wrote:
> My first thought was to wonder whether this was even a bug. I
> remembered that EXPLAIN treats shared, local, and temp buffers as
> three separate categories of things. But it seems that someone decided
> to conflate two of them for I/O timing purposes:
> 
>             if (has_timing)
>             {
>                 appendStringInfoString(es->str, "
> shared/local");
> 
> ^^^^ Notice this bit in particular.

I was reviewing the whole, and this is an oversight specific to
efb0ef909f60, because we've never incremented the write/read counters
for local buffers, even with this commit applied, for both the EXPLAIN
reports and anything stored in pg_stat_statement.  It seems to me that
the origin of the confusion comes down to pg_stat_database where
blk_{read|write}_time increments on both local and shared blocks, but
on EXPLAIN this stuff only reflects data about shared buffers.  So the
"shared" part of the string is right, but the "local" part is not in
v15 and v16.
--
Michael