Date: Tue, 07 Oct 2025 16:00:48 -0700 From: Cy Schubert <Cy.Schubert@cschubert.com> To: Cy Schubert <Cy.Schubert@cschubert.com> Cc: jlduran+freebsd@freebsd.org, Cy Schubert <cy@freebsd.org>, src-committers@freebsd.org, dev-commits-src-all@freebsd.org, dev-commits-src-main@freebsd.org Subject: Re: git: b2efd602aea8 - main - unbound: Vendor import 1.24.0 Message-ID: <20251007230048.64C8F39B@slippy.cwsent.com> In-Reply-To: <20251007221221.22BCC29E@slippy.cwsent.com> References: <202510071524.597FOTr3063628@gitrepo.freebsd.org> <CAPwQLcep1-Pt4P1QHiJpnZrMXXicG5Bc_tpxa6MtjVZT%2BvpJmg@mail.gmail.com> <20251007221221.22BCC29E@slippy.cwsent.com>
next in thread | previous in thread | raw e-mail | index | archive | help
In message <20251007221221.22BCC29E@slippy.cwsent.com>, Cy Schubert writes:
> In message <CAPwQLcep1-Pt4P1QHiJpnZrMXXicG5Bc_tpxa6MtjVZT+vpJmg@mail.gmail.c
> om>
> , Jose Luis Duran writes:
> > Hello Cy,
> >
> > According to the changelog:
> >
> > "The default value increase for so-sndbuf is to mitigate a cross-layer
> > issue where the UDP socket send buffers are exhausted waiting for
> > ARP/NDP resolution. Thanks to Reflyable for the report.
> >
> > To help the server start more easily, the setsockopt for sndbuf buffer
> > size prints a warning instead of a failure to start the server if it
> > can not set the buffer size."
> >
> > This means so-sndbuf is now 4m. Which triggers a warning, detailed in
> > upstream commit 713b5db5 ("- Fix to print warning for when so-sndbuf
> > setsockopt is not granted.").
> >
> > I wonder if we should revert back to using "so-sndbuf: 0" as the
> > default for FreeBSD? Or is there a better solution/workaround?
>
> There are in fact two commits.
>
> 1ef7b4a24 adjusted so-sndbuf to default to 4m.
>
> 03772d10f changed the default from to 1m.
>
> I assume you're suggesting reverting both upstream commits for FreeBSD?
Looking at the commit that started unbound down this path we see this in
their commit log before bumping the default to 4m:
- Change default for so-sndbuf to 1m, to mitigate a cross-layer
issue where the UDP socket send buffers are exhausted waiting
for ARP/NDP resolution. Thanks to Reflyable for the report.
I use local_unbound on one of my machines here. It has 8 GB RAM. I don't
see any warning messages WRT the socket option not accepted. The first
question that comes to mind, to try to understand your environment, how
much RAM does this machine have and of that how much is allocated to the
kernel?
--
Cheers,
Cy Schubert <Cy.Schubert@cschubert.com>
FreeBSD UNIX: <cy@FreeBSD.org> Web: https://FreeBSD.org
NTP: <cy@nwtime.org> Web: https://nwtime.org
e**(i*pi)+1=0
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20251007230048.64C8F39B>