Date: Sun, 2 May 2021 17:52:22 -0700 From: Doug Hardie <bc979@lafn.org> To: "net@freebsd.org" <net@FreeBSD.org> Subject: Re: Problem with resolver/host Message-ID: <51F58892-1383-4698-8984-173CDE01AB67@sermon-archive.info> In-Reply-To: <041E6B07-0797-436F-909F-5B77ED77DAEC@sermon-archive.info> References: <041E6B07-0797-436F-909F-5B77ED77DAEC@sermon-archive.info>
next in thread | previous in thread | raw e-mail | index | archive | help
> On 20 April 2021, at 14:12, Doug Hardie <bc979@lafn.org> wrote:
>=20
> I suspect this is an issue with the resolver, but it could also be in =
host. I have the following in /etc/resolv.conf:
>=20
> # Generated by resolvconf
> nameserver fe80::213:72ff:fec3:180f%bge0
> nameserver fe80::120c:6bff:fee9:cdf7%bge0
>=20
> There is a DNS server running at the first address that supplies IPv6 =
addresses. When I ping the first address, it succeeds:
>=20
> test# ping fe80::213:72ff:fec3:180f%bge0
> PING6(56=3D40+8+8 bytes) fe80::6a5b:35ff:fed3:4eda%bge0 --> =
fe80::213:72ff:fec3:180f%bge0
> 16 bytes from fe80::213:72ff:fec3:180f%bge0, icmp_seq=3D0 hlim=3D64 =
time=3D0.321 ms
> 16 bytes from fe80::213:72ff:fec3:180f%bge0, icmp_seq=3D1 hlim=3D64 =
time=3D0.553 ms
> 16 bytes from fe80::213:72ff:fec3:180f%bge0, icmp_seq=3D2 hlim=3D64 =
time=3D0.544 ms
> 16 bytes from fe80::213:72ff:fec3:180f%bge0, icmp_seq=3D3 hlim=3D64 =
time=3D0.424 ms
>=20
> However, host client.global.example return an error:
>=20
> test# host client.global.example
> ;; connection timed out; no servers could be reached
>=20
> ktrace of host shows that either host or the resolver is trying to use =
the address without the %bge0 which is why that error occurs:
>=20
> 2027 host STRU struct sockaddr { AF_INET6, =
[fe80::213:72ff:fec3:180f]:53 }
> 2027 host RET sendto -1 errno 51 Network is unreachable
>=20
> Is this a bug or is there some configuration item I have missed:
>=20
I have filed a bug report: Bug 255316 on this. However, it is still =
sitting in limbo: =20
Assignee: freebsd-bugs (Nobody)
I am not sure that the patch I provided is not going to break something =
else. I am not sure this is the right place to ask these questions. =
The resolver is used for more than just finding the default route. I =
also have no idea how to build the parts needed to test it. I know the =
patch works fine in a simple test case where I extracted that portion of =
the code as a standalone test. I don't want to do a buildworld as that =
machine has no cooling and overheats quickly. =20
-- Doug=
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?51F58892-1383-4698-8984-173CDE01AB67>