1RESOLVED.CONF(5) resolved.conf RESOLVED.CONF(5)
2
6 resolved.conf, resolved.conf.d - Network Name Resolution configuration
7 files
8
10 /etc/systemd/resolved.conf
11 /etc/systemd/resolved.conf.d/*.conf
13 /run/systemd/resolved.conf.d/*.conf
15 /usr/lib/systemd/resolved.conf.d/*.conf
17
19 These configuration files control local DNS and LLMNR name resolution.
20
22 The default configuration is set during compilation, so configuration
23 is only needed when it is necessary to deviate from those defaults.
24 Initially, the main configuration file in /etc/systemd/ contains
25 commented out entries showing the defaults as a guide to the
26 administrator. Local overrides can be created by editing this file or
27 by creating drop-ins, as described below. Using drop-ins for local
28 configuration is recommended over modifications to the main
29 configuration file.
30 In addition to the "main" configuration file, drop-in configuration
32 snippets are read from /usr/lib/systemd/*.conf.d/,
33 /usr/local/lib/systemd/*.conf.d/, and /etc/systemd/*.conf.d/. Those
34 drop-ins have higher precedence and override the main configuration
35 file. Files in the *.conf.d/ configuration subdirectories are sorted by
36 their filename in lexicographic order, regardless of in which of the
37 subdirectories they reside. When multiple files specify the same
38 option, for options which accept just a single value, the entry in the
39 file sorted last takes precedence, and for options which accept a list
40 of values, entries are collected as they occur in the sorted files.
41 When packages need to customize the configuration, they can install
43 drop-ins under /usr/. Files in /etc/ are reserved for the local
44 administrator, who may use this logic to override the configuration
45 files installed by vendor packages. Drop-ins have to be used to
46 override package drop-ins, since the main configuration file has lower
47 precedence. It is recommended to prefix all filenames in those
48 subdirectories with a two-digit number and a dash, to simplify the
49 ordering of the files.
50 To disable a configuration file supplied by the vendor, the recommended
52 way is to place a symlink to /dev/null in the configuration directory
53 in /etc/, with the same filename as the vendor configuration file.
54
56 The following options are available in the [Resolve] section:
57 DNS=
59 A space-separated list of IPv4 and IPv6 addresses to use as system
60 DNS servers. Each address can optionally take a port number
61 separated with ":", a network interface name or index separated
62 with "%", and a Server Name Indication (SNI) separated with "#".
63 When IPv6 address is specified with a port number, then the address
64 must be in the square brackets. That is, the acceptable full
65 formats are "111.222.333.444:9953%ifname#example.com" for IPv4 and
66 "[1111:2222::3333]:9953%ifname#example.com" for IPv6. DNS requests
67 are sent to one of the listed DNS servers in parallel to suitable
68 per-link DNS servers acquired from systemd-networkd.service(8) or
69 set at runtime by external applications. For compatibility reasons,
70 if this setting is not specified, the DNS servers listed in
71 /etc/resolv.conf are used instead, if that file exists and any
72 servers are configured in it. This setting defaults to the empty
73 list.
74 FallbackDNS=
76 A space-separated list of IPv4 and IPv6 addresses to use as the
77 fallback DNS servers. Please see DNS= for acceptable format of
78 addresses. Any per-link DNS servers obtained from systemd-
79 networkd.service(8) take precedence over this setting, as do any
80 servers set via DNS= above or /etc/resolv.conf. This setting is
81 hence only used if no other DNS server information is known. If
82 this option is not given, a compiled-in list of DNS servers is used
83 instead.
84 Domains=
86 A space-separated list of domains, optionally prefixed with "~",
87 used for two distinct purposes described below. Defaults to the
88 empty list.
89 Any domains not prefixed with "~" are used as search suffixes when
91 resolving single-label hostnames (domain names which contain no
92 dot), in order to qualify them into fully-qualified domain names
93 (FQDNs). These "search domains" are strictly processed in the order
94 they are specified in, until the name with the suffix appended is
95 found. For compatibility reasons, if this setting is not specified,
96 the search domains listed in /etc/resolv.conf with the search
97 keyword are used instead, if that file exists and any domains are
98 configured in it.
99 The domains prefixed with "~" are called "route-only domains". All
101 domains listed here (both search domains and route-only domains
102 after removing the "~" prefix) define a search path that preferably
103 directs DNS queries to this interface. This search path has an
104 effect only when suitable per-link DNS servers are known. Such
105 servers may be defined through the DNS= setting (see above) and
106 dynamically at run time, for example from DHCP leases. If no
107 per-link DNS servers are known, route-only domains have no effect.
108 Use the construct "~." (which is composed from "~" to indicate a
110 route-only domain and "." to indicate the DNS root domain that is
111 the implied suffix of all DNS domains) to use the DNS servers
112 defined for this link preferably for all domains.
113 See "Protocols and Routing" in systemd-resolved.service(8) for
115 details of how search and route-only domains are used.
116 LLMNR=
118 Takes a boolean argument or "resolve". Controls Link-Local
119 Multicast Name Resolution support (RFC 4795[1]) on the local host.
120 If true, enables full LLMNR responder and resolver support. If
121 false, disables both. If set to "resolve", only resolution support
122 is enabled, but responding is disabled. Note that systemd-
123 networkd.service(8) also maintains per-link LLMNR settings. LLMNR
124 will be enabled on a link only if the per-link and the global
125 setting is on.
126 MulticastDNS=
128 Takes a boolean argument or "resolve". Controls Multicast DNS
129 support (RFC 6762[2]) on the local host. If true, enables full
130 Multicast DNS responder and resolver support. If false, disables
131 both. If set to "resolve", only resolution support is enabled, but
132 responding is disabled. Note that systemd-networkd.service(8) also
133 maintains per-link Multicast DNS settings. Multicast DNS will be
134 enabled on a link only if the per-link and the global setting is
135 on.
136 DNSSEC=
138 Takes a boolean argument or "allow-downgrade".
139 If set to true, all DNS lookups are DNSSEC-validated locally
141 (excluding LLMNR and Multicast DNS). If the response to a lookup
142 request is detected to be invalid a lookup failure is returned to
143 applications. Note that this mode requires a DNS server that
144 supports DNSSEC. If the DNS server does not properly support DNSSEC
145 all validations will fail.
146 If set to "allow-downgrade", DNSSEC validation is attempted, but if
148 the server does not support DNSSEC properly, DNSSEC mode is
149 automatically disabled. Note that this mode makes DNSSEC validation
150 vulnerable to "downgrade" attacks, where an attacker might be able
151 to trigger a downgrade to non-DNSSEC mode by synthesizing a DNS
152 response that suggests DNSSEC was not supported.
153 If set to false, DNS lookups are not DNSSEC validated. In this
155 mode, or when set to "allow-downgrade" and the downgrade has
156 happened, the resolver becomes security-unaware and all forwarded
157 queries have DNSSEC OK (DO) bit unset.
158 Note that DNSSEC validation requires retrieval of additional DNS
160 data, and thus results in a small DNS lookup time penalty.
161 DNSSEC requires knowledge of "trust anchors" to prove data
163 integrity. The trust anchor for the Internet root domain is built
164 into the resolver, additional trust anchors may be defined with
165 dnssec-trust-anchors.d(5). Trust anchors may change at regular
166 intervals, and old trust anchors may be revoked. In such a case
167 DNSSEC validation is not possible until new trust anchors are
168 configured locally or the resolver software package is updated with
169 the new root trust anchor. In effect, when the built-in trust
170 anchor is revoked and DNSSEC= is true, all further lookups will
171 fail, as it cannot be proved anymore whether lookups are correctly
172 signed, or validly unsigned. If DNSSEC= is set to "allow-downgrade"
173 the resolver will automatically turn off DNSSEC validation in such
174 a case.
175 Client programs looking up DNS data will be informed whether
177 lookups could be verified using DNSSEC, or whether the returned
178 data could not be verified (either because the data was found
179 unsigned in the DNS, or the DNS server did not support DNSSEC or no
180 appropriate trust anchors were known). In the latter case it is
181 assumed that client programs employ a secondary scheme to validate
182 the returned DNS data, should this be required.
183 It is recommended to set DNSSEC= to true on systems where it is
185 known that the DNS server supports DNSSEC correctly, and where
186 software or trust anchor updates happen regularly. On other systems
187 it is recommended to set DNSSEC= to "allow-downgrade".
188 In addition to this global DNSSEC setting systemd-
190 networkd.service(8) also maintains per-link DNSSEC settings. For
191 system DNS servers (see above), only the global DNSSEC setting is
192 in effect. For per-link DNS servers the per-link setting is in
193 effect, unless it is unset in which case the global setting is used
194 instead.
195 Site-private DNS zones generally conflict with DNSSEC operation,
197 unless a negative (if the private zone is not signed) or positive
198 (if the private zone is signed) trust anchor is configured for
199 them. If "allow-downgrade" mode is selected, it is attempted to
200 detect site-private DNS zones using top-level domains (TLDs) that
201 are not known by the DNS root server. This logic does not work in
202 all private zone setups.
203 Defaults to "no".
205 DNSOverTLS=
207 Takes a boolean argument or "opportunistic". If true all
208 connections to the server will be encrypted. Note that this mode
209 requires a DNS server that supports DNS-over-TLS and has a valid
210 certificate. If the hostname was specified in DNS= by using the
211 format "address#server_name" it is used to validate its certificate
212 and also to enable Server Name Indication (SNI) when opening a TLS
213 connection. Otherwise the certificate is checked against the
214 server's IP. If the DNS server does not support DNS-over-TLS all
215 DNS requests will fail.
216 When set to "opportunistic" DNS request are attempted to send
218 encrypted with DNS-over-TLS. If the DNS server does not support
219 TLS, DNS-over-TLS is disabled. Note that this mode makes
220 DNS-over-TLS vulnerable to "downgrade" attacks, where an attacker
221 might be able to trigger a downgrade to non-encrypted mode by
222 synthesizing a response that suggests DNS-over-TLS was not
223 supported. If set to false, DNS lookups are send over UDP.
224 Note that DNS-over-TLS requires additional data to be send for
226 setting up an encrypted connection, and thus results in a small DNS
227 look-up time penalty.
228 Note that in "opportunistic" mode the resolver is not capable of
230 authenticating the server, so it is vulnerable to
231 "man-in-the-middle" attacks.
232 In addition to this global DNSOverTLS= setting systemd-
234 networkd.service(8) also maintains per-link DNSOverTLS= settings.
235 For system DNS servers (see above), only the global DNSOverTLS=
236 setting is in effect. For per-link DNS servers the per-link setting
237 is in effect, unless it is unset in which case the global setting
238 is used instead.
239 Defaults to "no".
241 Cache=
243 Takes a boolean or "no-negative" as argument. If "yes" (the
244 default), resolving a domain name which already got queried earlier
245 will return the previous result as long as it is still valid, and
246 thus does not result in a new network request. Be aware that
247 turning off caching comes at a performance penalty, which is
248 particularly high when DNSSEC is used. If "no-negative", only
249 positive answers are cached.
250 Note that caching is turned off by default for host-local DNS
252 servers. See CacheFromLocalhost= for details.
253 CacheFromLocalhost=
255 Takes a boolean as argument. If "no" (the default), and response
256 cames from host-local IP address (such as 127.0.0.1 or ::1), the
257 result wouldn't be cached in order to avoid potential duplicate
258 local caching.
259 DNSStubListener=
261 Takes a boolean argument or one of "udp" and "tcp". If "udp", a DNS
262 stub resolver will listen for UDP requests on addresses 127.0.0.53
263 and 127.0.0.54, port 53. If "tcp", the stub will listen for TCP
264 requests on the same addresses and port. If "yes" (the default),
265 the stub listens for both UDP and TCP requests. If "no", the stub
266 listener is disabled.
267 The DNS stub resolver on 127.0.0.53 provides the full feature set
269 of the local resolver, which includes offering LLMNR/MulticastDNS
270 resolution. The DNS stub resolver on 127.0.0.54 provides a more
271 limited resolver, that operates in "proxy" mode only, i.e. it will
272 pass most DNS messages relatively unmodified to the current
273 upstream DNS servers and back, but not try to process the messages
274 locally, and hence does not validate DNSSEC, or offer up
275 LLMNR/MulticastDNS. (It will translate to DNS-over-TLS
276 communication if needed however.)
277 Note that the DNS stub listener is turned off implicitly when its
279 listening address and port are already in use.
280 DNSStubListenerExtra=
282 Takes an IPv4 or IPv6 address to listen on. The address may be
283 optionally prefixed with a protocol name ("udp" or "tcp") separated
284 with ":". If the protocol is not specified, the service will listen
285 on both UDP and TCP. It may be also optionally suffixed by a
286 numeric port number with separator ":". When an IPv6 address is
287 specified with a port number, then the address must be in the
288 square brackets. If the port is not specified, then the service
289 uses port 53. Note that this is independent of the primary DNS stub
290 configured with DNSStubListener=, and only configures additional
291 sockets to listen on. This option can be specified multiple times.
292 If an empty string is assigned, then the all previous assignments
293 are cleared. Defaults to unset.
294 Examples:
296 DNSStubListenerExtra=192.168.10.10
298 DNSStubListenerExtra=2001:db8:0:f102::10
299 DNSStubListenerExtra=192.168.10.11:9953
300 DNSStubListenerExtra=[2001:db8:0:f102::11]:9953
301 DNSStubListenerExtra=tcp:192.168.10.12
302 DNSStubListenerExtra=udp:2001:db8:0:f102::12
303 DNSStubListenerExtra=tcp:192.168.10.13:9953
304 DNSStubListenerExtra=udp:[2001:db8:0:f102::13]:9953
305 ReadEtcHosts=
308 Takes a boolean argument. If "yes" (the default), systemd-resolved
309 will read /etc/hosts, and try to resolve hosts or address by using
310 the entries in the file before sending query to DNS servers.
311 ResolveUnicastSingleLabel=
313 Takes a boolean argument. When false (the default),
314 systemd-resolved will not resolve A and AAAA queries for
315 single-label names over classic DNS. Note that such names may still
316 be resolved if search domains are specified (see Domains= above),
317 or using other mechanisms, in particular via LLMNR or from
318 /etc/hosts. When true, queries for single-label names will be
319 forwarded to global DNS servers even if no search domains are
320 defined.
321 This option is provided for compatibility with configurations where
323 public DNS servers are not used. Forwarding single-label names to
324 servers not under your control is not standard-conformant, see IAB
325 Statement[3], and may create a privacy and security risk.
326 StaleRetentionSec=SECONDS
328 Takes a duration value, which determines the length of time DNS
329 resource records can be retained in the cache beyond their Time To
330 Live (TTL). This allows these records to be returned as stale
331 records. By default, this value is set to zero, meaning that DNS
332 resource records are not stored in the cache after their TTL
333 expires.
334 This is useful when a DNS server failure occurs or becomes
336 unreachable. In such cases, systemd-resolved continues to use the
337 stale records to answer DNS queries, particularly when no valid
338 response can be obtained from the upstream DNS servers. However,
339 this doesn't apply to NXDOMAIN responses, as those are still
340 perfectly valid responses. This feature enhances resilience against
341 DNS infrastructure failures and outages.
342 systemd-resolved always attempts to reach the upstream DNS servers
344 first, before providing the client application with any stale data.
345 If this feature is enabled, cache will not be flushed when changing
346 servers.
347
349 systemd(1), systemd-resolved.service(8), systemd-networkd.service(8),
350 dnssec-trust-anchors.d(5), resolv.conf(5)
351
353 1. RFC 4795
354 https://tools.ietf.org/html/rfc4795
355 2. RFC 6762
357 https://tools.ietf.org/html/rfc6762
358 3. IAB Statement
360 https://www.iab.org/documents/correspondence-reports-documents/2013-2/iab-statement-dotless-domains-considered-harmful/
361systemd 254 RESOLVED.CONF(5)