Date: Mon, 24 Feb 2020 12:01:19 -0800
From: Andres Freund <andres@anarazel.de>
To: Alvaro Herrera <alvherre@2ndquadrant.com>
Cc: Tom Lane <tgl@sss.pgh.pa.us>, Thomas Munro <thomas.munro@gmail.com>,
	=?utf-8?B?UsOpbWk=?= Zara <remi_zara@mac.com>,
	Michael Paquier <michael@paquier.xyz>,
	PostgreSQL Hackers <pgsql-hackers@lists.postgresql.org>
Subject: Re: pgsql: Add kqueue(2) support to the WaitEventSet API.
Message-ID: <20200224200119.x27tfdybwpzde3cy@alap3.anarazel.de>
References: <16353.1582555469@sss.pgh.pa.us>
 <20200224191453.GA711@alvherre.pgsql>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <20200224191453.GA711@alvherre.pgsql>
Precedence: bulk

Hi,

On 2020-02-24 16:14:53 -0300, Alvaro Herrera wrote:
> But the C runtime does:
> https://docs.microsoft.com/en-us/cpp/c-runtime-library/file-handling?view=vs-2019
> I suppose we do use the C runtime.  There's a reference to
> _setmaxstdio() being able to raise the limit from the default of 512 to
> up to 8192 open files.  We don't currently invoke that.
> https://docs.microsoft.com/en-us/cpp/c-runtime-library/reference/setmaxstdio?view=vs-2019

If we ever go for that, we should also consider raising the limit on
unix systems up to the hard limit when hitting the fd ceiling. I.e. get
the current limit with getrlimit(RLIMIT_NOFILE) and raise rlim_cur
[closer] to rlim_max with setrlimit.

Perhaps it'd even be worthwhile to just always raise the limit, if
possible, in set_max_safe_fds(), by max_safe_fds +
NUM_RESERVED_FDS. That way PLs, other shared libs, would have a more
usualy amount of FDs available. Rather than a fairly small number, but
only when the backend has been running for a while in the right
workload.

Greetings,

Andres Freund