Files
mesh-catalog/modules/dnsmasq/files/dnsmasq.conf
T
jschoubben c88be88ccf
mesh/delivery merged: source.main moved to novox/mesh-catalog@94805575379f
mesh/check checking c88be88ccf2f on 94805575379f
dnsmasq's public upstream resolvers are config values, defaulting to the two it always used
The two server lines were literals in the template, so changing an upstream needed a build. They are now the config keys upstream-resolver-first (1.1.1.1) and upstream-resolver-second (8.8.8.8), placed with ${config:...}; the service already restarts on its config file. no-resolv and every other line stay.

Claude-Session: https://claude.ai/code/session_01NcbgfFfQ4Z9UE5jK32PmE5
2026-11-08 23:48:59 +01:00

88 lines
5.1 KiB
Plaintext

# Managed by the mesh. dnsmasq's own defaults are replaced whole rather than
# patched, because this module owns the file and a patch would leave whatever
# was there before to be discovered later.
#
# This is one of the mesh's resolvers (novox/hq ADR 0194, 0223): it holds the
# `mesh-dns-resolver` seat, which more than one machine may hold, each answering
# the same names from the same roster. Every node and every container lists every
# holder and no public resolver, so whichever answers first gives the one answer.
# It holds the mesh's own names and nothing else (ADR 0191) — the roster is the
# mesh's, rendered into each holder; no other copy lives on any machine.
# What the mesh computed: one wildcard per machine — its name and everything
# under it — and the mesh's own suffix as a local domain, so a name under it is
# answered here or not at all and is never asked upstream. Rewritten whenever a
# machine joins or leaves, which is why the service below restarts on it: a
# reload makes dnsmasq re-read hosts files, not its configuration, and a
# wildcard is configuration.
conf-file=/etc/mesh-resolver/nodes.conf
# Every zone a module answers itself (ADR 0199): one forwarding line per zone, to
# the private address of the machine its module runs on and the port its
# answerer is published on. Nothing in a zone is answered here; what names a zone
# holds is the module's. Rewritten whenever a zone is declared or moves, and the
# service restarts on it for the same reason as above.
conf-file=/etc/mesh-resolver/zones.conf
# Where it answers: this machine's private address, so every node — and every
# container on every node, which copies its machine's resolvers — can ask; and
# loopback, for this machine's own use. Never a LAN address: a device that is
# not a member cannot reach what the mesh's names point at, and a LAN's resolver
# is its router's (ADR 0194).
#
# Named as an ADDRESS and not as the interface that carries it, on purpose. A
# container's query is addressed to this address but arrives on the runtime's
# bridge, and dnsmasq checks every query against what it was told to answer on:
# an `interface=` line admits queries by the interface they arrive on, a
# `listen-address=` line by the address they are sent to. With `interface=mesh0`
# alone, a query from a container was received and dropped without a word
# (novox/hq 04-ISSUES/110). The address admits it whatever bridge it comes in on.
#
# bind-interfaces, and never bind-dynamic (novox/hq issue 348). bind-dynamic
# re-reads the machine's addresses on every change of any link and closes each
# listener whose interface that reading did not find; a reading made while the
# container runtime deletes a bridge's virtual links fails ("index gone"), and
# then every listener is closed until a later change reads cleanly. On
# 2026-10-09 the private address was left unbound for twenty minutes, the daemon
# running and silent. bind-interfaces binds each address once, at start, and
# never closes it. It cannot bind an address that is not there yet, so the unit
# waits for it (the drop-in below); a new private address restarts the unit, as
# this file changing always did.
bind-interfaces
listen-address=${machine:address}
listen-address=127.0.0.1
# **It must never read resolv.conf to find out where to forward.** This
# machine's resolv.conf names this resolver — so a resolver that read it for
# upstreams would find itself, and every query it could not answer locally would
# loop until its receive queue filled. That is not theoretical: it filled with
# 15KB of queries and every lookup on the machine hung. no-resolv makes that loop
# impossible: the upstreams are the two lines below (the module's settings upstream-resolver-first and -second), and nothing on the machine
# can redirect them. The mesh's own names never reach them: the local= line in
# the file above stops them here, answered or refused.
no-resolv
server=${config:upstream-resolver-first}
server=${config:upstream-resolver-second}
# Both upstreams validate DNSSEC and say so with the Authenticated Data bit,
# and this passes that bit down rather than dropping it, which dnsmasq does
# unless told. It does not validate itself: an answer's trust is the upstream's
# and the path to it, which is what a forwarding resolver's trust always was.
# Without it a program that refuses to run behind a resolver that does not
# validate — a mail system's admin does exactly that check at start — cannot
# use this resolver (novox/hq 04-ISSUES/171).
proxy-dnssec
# A name without a dot is never forwarded, and reverse lookups of private ranges
# are answered here rather than asking the world who 10.x is.
domain-needed
bogus-priv
# **No hosts file, and no operator's files.** This machine's /etc/hosts is its
# own — the hosts module's block and the operator's lines (ADR 0199) — and the
# mesh's resolver answers every node, so a line written for one machine's own
# programs must not become an answer for all of them. Every per-node resolver
# went wrong exactly here: a hosts file read once at start, an operator's old
# line beside the mesh's, a drop-in read from a directory nobody owned.
no-hosts