Horton
Menu

Re: Draft for basic NUMA observability

Tomas Vondra <tomas@vondra.me>

View thread
From: Tomas Vondra <tomas@vondra.me>
To: Andres Freund <andres@anarazel.de>, Jakub Wartak <jakub.wartak@enterprisedb.com>
Cc: Bertrand Drouvot <bertranddrouvot.pg@gmail.com>, Alvaro Herrera <alvherre@alvh.no-ip.org>, Nazir Bilal Yavuz <byavuz81@gmail.com>, PostgreSQL Hackers <pgsql-hackers@postgresql.org>
Date: 2025-04-07T17:59:59Z
Lists: pgsql-hackers

Attachments

On 4/7/25 19:24, Andres Freund wrote:
> On 2025-04-04 19:07:12 +0200, Jakub Wartak wrote:
>> They actually look good to me. We've discussed earlier dropping
>> s/numa_//g for column names (after all views contain it already) so
>> they are fine in this regard.
>> There's also the question of consistency: (bufferid, page_num,
>> node_id) -- maybe should just drop "_" and that's it?
>> Well I would even possibly consider page_num -> ospagenumber, but that's ugly.
> 
> I'd go for os_page_num.

WFM. I've renamed "ospageid" to "os_page_num" in 0003, and I've also
renamed "node_id" to "numa_node" in 0002+0003, to make it clearer what
kind of node this is.

This reminds me whether it's fine to have "os_page_num" as int. Should
we make it bigint, perhaps?

Attached is v28, with the commit messages updated, added <warning> about
allocation of the memory, etc. I'll let the CI run the tests on it, and
then will push, unless someone has more comments.


regards

-- 
Tomas Vondra

Commits

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

  1. doc: Correct pg_shmem_allocations_numa.size data type

  2. Add pg_buffercache_numa view with NUMA node info

  3. Add support for basic NUMA awareness

  4. Introduce pg_shmem_allocations_numa view