Horton

Re: Reducing power consumption on idle servers

Thomas Munro <thomas.munro@gmail.com>

View thread View on postgr.es
From: Thomas Munro <thomas.munro@gmail.com>
To: Simon Riggs <simon.riggs@enterprisedb.com>
Cc: Robert Haas <robertmhaas@gmail.com>, Kyotaro Horiguchi <horikyota.ntt@gmail.com>, Andres Freund <andres@anarazel.de>, Zheng Li <zhengli10@gmail.com>, Jim Nasby <nasbyj@amazon.com>, Tom Lane <tgl@sss.pgh.pa.us>, PostgreSQL Hackers <pgsql-hackers@postgresql.org>
Date: 2022-11-13T21:28:14Z
Lists: pgsql-hackers 
On Mon, Nov 14, 2022 at 5:52 AM Simon Riggs
<simon.riggs@enterprisedb.com> wrote:
> The attached patch is a reduced version of the original. It covers only:
> * deprecation of the promote_trigger_file - there are no tests that
> use that, hence why there is no test coverage for the patch
> * changing the sleep time of the startup process to 60s
> * docs and comments

LPGTM.  If we also fix the bogus SIGALRM wakeups[1], then finally a
completely idle recovery process looks like:

kevent(8,0x0,0,{ },1,{ 60.000000000 })           = 0 (0x0)
kevent(8,0x0,0,{ },1,{ 60.000000000 })           = 0 (0x0)
kevent(8,0x0,0,{ },1,{ 60.000000000 })           = 0 (0x0)

Presumably it would have no timeout at all in the next release.

[1] https://www.postgresql.org/message-id/CALj2ACUiYn+ZmPGUVmGeoY1u7ino2qsvqrnufk8sWPvK3A8yJA@mail.gmail.com

Commits

  1. Remove promote_trigger_file.

  2. Add pg_promote function

  3. pg_ctl promote