Section: Maintenance Commands (8)
Index Return to Main Contents


dnsmasq - A lightweight DHCP and caching DNS server.  


dnsmasq [OPTION]...  


dnsmasq is a lightweight DNS, TFTP and DHCP server. It is intended to provide coupled DNS and DHCP service to a LAN.

Dnsmasq accepts DNS queries and either answers them from a small, local, cache or forwards them to a real, recursive, DNS server. It loads the contents of /etc/hosts so that local hostnames which do not appear in the global DNS can be resolved and also answers DNS queries for DHCP configured hosts.

The dnsmasq DHCP server supports static address assignments, multiple networks, DHCP-relay and RFC3011 subnet specifiers. It automatically sends a sensible default set of DHCP options, and can be configured to send any desired set of DHCP options, including vendor-encapsulated options. It includes a secure, read-only, TFTP server to allow net/PXE boot of DHCP hosts and also supports BOOTP.

Dnsmasq supports IPv6 for DNS, but not DHCP.  


Note that in general missing parameters are allowed and switch off functions, for instance "--pid-file" disables writing a PID file. On BSD, unless the GNU getopt library is linked, the long form of the options does not work on the command line; it is still recognised in the configuration file.
-h, --no-hosts
Don't read the hostnames in /etc/hosts.
-H, --addn-hosts=<file>
Additional hosts file. Read the specified file as well as /etc/hosts. If -h is given, read only the specified file. This option may be repeated for more than one additional hosts file.
-E, --expand-hosts
Add the domain to simple names (without a period) in /etc/hosts in the same way as for DHCP-derived names.
-T, --local-ttl=<time>
When replying with information from /etc/hosts or the DHCP leases file dnsmasq by default sets the time-to-live field to zero, meaning that the requestor should not itself cache the information. This is the correct thing to do in almost all situations. This option allows a time-to-live (in seconds) to be given for these replies. This will reduce the load on the server at the expense of clients using stale data under some circumstances.
-k, --keep-in-foreground
Do not go into the background at startup but otherwise run as normal. This is intended for use when dnsmasq is run under daemontools or launchd.
-d, --no-daemon
Debug mode: don't fork to the background, don't write a pid file, don't change user id, generate a complete cache dump on receipt on SIGUSR1, log to stderr as well as syslog, don't fork new processes to handle TCP queries.
-q, --log-queries
Log the results of DNS queries handled by dnsmasq. Enable a full cache dump on receipt of SIGUSR1.
-8, --log-facility=<facility>
Set the facility to which dnsmasq will send syslog entries, this defaults to DAEMON, and to LOCAL0 when debug mode is in operation. If the facilty given contains at least one '/' character, it is taken to be a filename, and dnsmasq logs to the given file, instead of syslog. (Errors whilst reading configuration will still go to syslog, but all output from a successful startup, and all output whilst running, will go exclusively to the file.)
Enable asynchronous logging and optionally set the limit on the number of lines which will be queued by dnsmasq when writing to the syslog is slow. Dnsmasq can log asynchronously: this allows it to continue functioning without being blocked by syslog, and allows syslog to use dnsmasq for DNS queries without risking deadlock. If the queue of log-lines becomes full, dnsmasq will log the overflow, and the number of messages lost. The default queue length is 5, a sane value would be 5-25, and a maximum limit of 100 is imposed.
-x, --pid-file=<path>
Specify an alternate path for dnsmasq to record its process-id in. Normally /var/run/dnsmasq.pid.
-u, --user=<username>
Specify the userid to which dnsmasq will change after startup. Dnsmasq must normally be started as root, but it will drop root privileges after startup by changing id to another user. Normally this user is "nobody" but that can be over-ridden with this switch.
-g, --group=<groupname>
Specify the group which dnsmasq will run as. The defaults to "dip", if available, to facilitate access to /etc/ppp/resolv.conf which is not normally world readable.
-v, --version
Print the version number.
-p, --port=<port>
Listen on <port> instead of the standard DNS port (53). Useful mainly for debugging.
-P, --edns-packet-max=<size>
Specify the largest EDNS.0 UDP packet which is supported by the DNS forwarder. Defaults to 1280, which is the RFC2671-recommended maximum for ethernet.
-Q, --query-port=<query_port>
Send outbound DNS queries from, and listen for their replies on, the specific UDP port <query_port> instead of using one chosen at runtime. Useful to simplify your firewall rules; without this, your firewall would have to allow connections from outside DNS servers to a range of UDP ports, or dynamically adapt to the port being used by the current dnsmasq instance.
-i, --interface=<interface name>
Listen only on the specified interface(s). Dnsmasq automatically adds the loopback (local) interface to the list of interfaces to use when the --interface option is used. If no --interface or --listen-address options are given dnsmasq listens on all available interfaces except any given in --except-interface options. IP alias interfaces (eg "eth1:0") cannot be used with --interface or --except-interface options, use --listen-address instead.
-I, --except-interface=<interface name>
Do not listen on the specified interface. Note that the order of --listen-address --interface and --except-interface options does not matter and that --except-interface options always override the others.
-2, --no-dhcp-interface=<interface name>
Do not provide DHCP or TFTP on the specified interface, but do provide DNS service.
-a, --listen-address=<ipaddr>
Listen on the given IP address(es). Both --interface and --listen-address options may be given, in which case the set of both interfaces and addresses is used. Note that if no --interface option is given, but --listen-address is, dnsmasq will not automatically listen on the loopback interface. To achieve this, its IP address,, must be explicitly given as a --listen-address option.
-z, --bind-interfaces
On systems which support it, dnsmasq binds the wildcard address, even when it is listening on only some interfaces. It then discards requests that it shouldn't reply to. This has the advantage of working even when interfaces come and go and change address. This option forces dnsmasq to really bind only the interfaces it is listening on. About the only time when this is useful is when running another nameserver (or another instance of dnsmasq) on the same machine. Setting this option also enables multiple instances of dnsmasq which provide DHCP service to run in the same machine.
-y, --localise-queries
Return answers to DNS queries from /etc/hosts which depend on the interface over which the query was received. If a name in /etc/hosts has more than one address associated with it, and at least one of those addresses is on the same subnet as the interface to which the query was sent, then return only the address(es) on that subnet. This allows for a server to have multiple addresses in /etc/hosts corresponding to each of its interfaces, and hosts will get the correct address based on which network they are attached to. Currently this facility is limited to IPv4.
-b, --bogus-priv
Bogus private reverse lookups. All reverse lookups for private IP ranges (ie 192.168.x.x, etc) which are not found in /etc/hosts or the DHCP leases file are answered with "no such domain" rather than being forwarded upstream.
-V, --alias=<old-ip>,<new-ip>[,<mask>]
Modify IPv4 addresses returned from upstream nameservers; old-ip is replaced by new-ip. If the optional mask is given then any address which matches the masked old-ip will be re-written. So, for instance --alias=,, will map to and to This is what Cisco PIX routers call "DNS doctoring".
-B, --bogus-nxdomain=<ipaddr>
Transform replies which contain the IP address given into "No such domain" replies. This is intended to counteract a devious move made by Verisign in September 2003 when they started returning the address of an advertising web page in response to queries for unregistered names, instead of the correct NXDOMAIN response. This option tells dnsmasq to fake the correct response when it sees this behaviour. As at Sept 2003 the IP address being returned by Verisign is
-f, --filterwin2k
Later versions of windows make periodic DNS requests which don't get sensible answers from the public DNS and can cause problems by triggering dial-on-demand links. This flag turns on an option to filter such requests. The requests blocked are for records of types SOA and SRV, and type ANY where the requested name has underscores, to catch LDAP requests.
-r, --resolv-file=<file>
Read the IP addresses of the upstream nameservers from <file>, instead of /etc/resolv.conf. For the format of this file see resolv.conf(5) the only lines relevant to dnsmasq are nameserver ones. Dnsmasq can be told to poll more than one resolv.conf file, the first file name specified overrides the default, subsequent ones add to the list. This is only allowed when polling; the file with the currently latest modification time is the one used.
-R, --no-resolv
Don't read /etc/resolv.conf. Get upstream servers only from the command line or the dnsmasq configuration file.
-1, --enable-dbus
Allow dnsmasq configuration to be updated via DBus method calls. The configuration which can be changed is upstream DNS servers (and corresponding domains) and cache clear. Requires that dnsmasq has been built with DBus support.
-o, --strict-order
By default, dnsmasq will send queries to any of the upstream servers it knows about and tries to favour servers to are known to be up. Setting this flag forces dnsmasq to try each query with each server strictly in the order they appear in /etc/resolv.conf
-n, --no-poll
Don't poll /etc/resolv.conf for changes.
Whenever /etc/resolv.conf is re-read, clear the DNS cache. This is useful when new nameservers may have different data than that held in cache.
-D, --domain-needed
Tells dnsmasq to never forward queries for plain names, without dots or domain parts, to upstream nameservers. If the name is not known from /etc/hosts or DHCP then a "not found" answer is returned.
-S, --local, --server=[/[<domain>]/[domain/]][<ipaddr>[#<port>][@<source>[#<port>]]]
Specify IP address of upstream severs directly. Setting this flag does not suppress reading of /etc/resolv.conf, use -R to do that. If one or more optional domains are given, that server is used only for those domains and they are queried only using the specified server. This is intended for private nameservers: if you have a nameserver on your network which deals with names of the form xxx.internal.thekelleys.org.uk at then giving the flag -S /internal.thekelleys.org.uk/ will send all queries for internal machines to that nameserver, everything else will go to the servers in /etc/resolv.conf. An empty domain specification, // has the special meaning of "unqualified names only" ie names without any dots in them. A non-standard port may be specified as part of the IP address using a # character. More than one -S flag is allowed, with repeated domain or ipaddr parts as required.

Also permitted is a -S flag which gives a domain but no IP address; this tells dnsmasq that a domain is local and it may answer queries from /etc/hosts or DHCP but should never forward queries on that domain to any upstream servers. local is a synonym for server to make configuration files clearer in this case.

The optional second IP address after the @ character tells dnsmasq how to set the source address of the queries to this nameserver. It should be an address belonging to the machine on which dnsmasq is running otherwise this server line will be logged and then ignored. The query-port flag is ignored for any servers which have a source address specified but the port may be specified directly as part of the source address.

-A, --address=/<domain>/[domain/]<ipaddr>
Specify an IP address to return for any host in the given domains. Queries in the domains are never forwarded and always replied to with the specified IP address which may be IPv4 or IPv6. To give both IPv4 and IPv6 addresses for a domain, use repeated -A flags. Note that /etc/hosts and DHCP leases override this for individual names. A common use of this is to redirect the entire doubleclick.net domain to some friendly local web server to avoid banner ads. The domain specification works in the same was as for --server, with the additional facility that /#/ matches any domain. Thus --address=/#/ will always return for any query not answered from /etc/hosts or DHCP and not sent to an upstream nameserver by a more specific --server directive.
-m, --mx-host=<mx name>[[,<hostname>],<preference>]
Return an MX record named <mx name> pointing to the given hostname (if given), or the host specified in the --mx-target switch or, if that switch is not given, the host on which dnsmasq is running. The default is useful for directing mail from systems on a LAN to a central server. The preference value is optional, and defaults to 1 if not given. More than one MX record may be given for a host.
-t, --mx-target=<hostname>
Specify the default target for the MX record returned by dnsmasq. See --mx-host. If --mx-target is given, but not --mx-host, then dnsmasq returns a MX record containing the MX target for MX queries on the hostname of the machine on which dnsmasq is running.
-e, --selfmx
Return an MX record pointing to itself for each local machine. Local machines are those in /etc/hosts or with DHCP leases.
-L, --localmx
Return an MX record pointing to the host given by mx-target (or the machine on which dnsmasq is running) for each local machine. Local machines are those in /etc/hosts or with DHCP leases.
-W, --srv-host=<_service>.<_prot>.[<domain>],[<target>[,<port>[,<priority>[,<weight>]]]]
Return a SRV DNS record. See RFC2782 for details. If not supplied, the domain defaults to that given by --domain. The default for the target domain is empty, and the default for port is one and the defaults for weight and priority are zero. Be careful if transposing data from BIND zone files: the port, weight and priority numbers are in a different order. More than one SRV record for a given service/domain is allowed, all that match are returned.
-Y, --txt-record=<name>[[,<text>],<text>]
Return a TXT DNS record. The value of TXT record is a set of strings, so any number may be included, split by commas.
Return a PTR DNS record.
Return a DNS record associating the name with the primary address on the given interface. This flag specifies an A record for the given name in the same way as an /etc/hosts line, except that the address is not constant, but taken from the given interface. If the interface is down, not configured or non-existant, an empty record is returned. The matching PTR record is also created, mapping the interface address to the name. More than one name may be associated with an interface address by repeating the flag; in that case the first instance is used for the reverse address-to-name mapping.
-c, --cache-size=<cachesize>
Set the size of dnsmasq's cache. The default is 150 names. Setting the cache size to zero disables caching.
-N, --no-negcache
Disable negative caching. Negative caching allows dnsmasq to remember "no such domain" answers from upstream nameservers and answer identical queries without forwarding them again. This flag disables negative caching.
-0, --dns-forward-max=<queries>
Set the maximum number of concurrent DNS queries. The default value is 150, which should be fine for most setups. The only known situation where this needs to be increased is when using web-server log file resolvers, which can generate large numbers of concurrent queries.
-F, --dhcp-range=[[net:]network-id,]<start-addr>,<end-addr>[[,<netmask>],<broadcast>][,<default lease time>]
Enable the DHCP server. Addresses will be given out from the range <start-addr> to <end-addr> and from statically defined addresses given in dhcp-host options. If the lease time is given, then leases will be given for that length of time. The lease time is in seconds, or minutes (eg 45m) or hours (eg 1h) or the literal "infinite". This option may be repeated, with different addresses, to enable DHCP service to more than one network. For directly connected networks (ie, networks on which the machine running dnsmasq has an interface) the netmask is optional. It is, however, required for networks which receive DHCP service via a relay agent. The broadcast address is always optional. On some broken systems, dnsmasq can listen on only one interface when using DHCP, and the name of that interface must be given using the interface option. This limitation currently affects OpenBSD before version 4.0. It is always allowed to have more than one dhcp-range in a single subnet. The optional network-id is a alphanumeric label which marks this network so that dhcp options may be specified on a per-network basis. When it is prefixed with 'net:' then its meaning changes from setting a tag to matching it. Only one tag may be set, but more than one tag may be matched. The end address may be replaced by the keyword static which tells dnsmasq to enable DHCP for the network specified, but not to dynamically allocate IP addresses. Only hosts which have static addresses given via dhcp-host or from /etc/ethers will be served.
-G, --dhcp-host=[<hwaddr>][,id:<client_id>|*][,net:<netid>][,<ipaddr>][,<hostname>][,<lease_time>][,ignore]
Specify per host parameters for the DHCP server. This allows a machine with a particular hardware address to be always allocated the same hostname, IP address and lease time. A hostname specified like this overrides any supplied by the DHCP client on the machine. It is also allowable to ommit the hardware address and include the hostname, in which case the IP address and lease times will apply to any machine claiming that name. For example --dhcp-host=00:20:e0:3b:13:af,wap,infinite tells dnsmasq to give the machine with hardware address 00:20:e0:3b:13:af the name wap, and an infinite DHCP lease. --dhcp-host=lap, tells dnsmasq to always allocate the machine lap the IP address Addresses allocated like this are not constrained to be in the range given by the --dhcp-range option, but they must be on the network being served by the DHCP server. It is allowed to use client identifiers rather than hardware addresses to identify hosts by prefixing with 'id:'. Thus: --dhcp-host=id:01:02:03:04,..... refers to the host with client identifier 01:02:03:04. It is also allowed to specify the client ID as text, like this: --dhcp-host=id:clientidastext,..... The special option id:* means "ignore any client-id and use MAC addresses only." This is useful when a client presents a client-id sometimes but not others. If a name appears in /etc/hosts, the associated address can be allocated to a DHCP lease, but only if a --dhcp-host option specifying the name also exists. The special keyword "ignore" tells dnsmasq to never offer a DHCP lease to a machine. The machine can be specified by hardware address, client ID or hostname, for instance --dhcp-host=00:20:e0:3b:13:af,ignore This is useful when there is another DHCP server on the network which should be used by some machines. The net:<network-id> sets the network-id tag whenever this dhcp-host directive is in use. This can be used to selectively send DHCP options just for this host. Ethernet addresses (but not client-ids) may have wildcard bytes, so for example --dhcp-host=00:20:e0:3b:13:*,ignore will cause dnsmasq to ignore a range of hardware addresses. Note that the "*" will need to be escaped or quoted on a command line, but not in the configuration file. Hardware addresses normally match any network (ARP) type, but it is possible to restrict them to a single ARP type by preceding them with the ARP-type (in HEX) and "-". so --dhcp-host=06-00:20:e0:3b:13:af, will only match a Token-Ring hardware address, since the ARP-address type for token ring is 6.
-Z, --read-ethers
Read /etc/ethers for information about hosts for the DHCP server. The format of /etc/ethers is a hardware address, followed by either a hostname or dotted-quad IP address. When read by dnsmasq these lines have exactly the same effect as --dhcp-host options containing the same information.
-O, --dhcp-option=[<network-id>,[<network-id>,]][vendor:[<vendor-class>],][<opt>|option:<opt-name>],[<value>[,<value>]]
Specify different or extra options to DHCP clients. By default, dnsmasq sends some standard options to DHCP clients, the netmask and broadcast address are set to the same as the host running dnsmasq, and the DNS server and default route are set to the address of the machine running dnsmasq. If the domain name option has been set, that is sent. This configuration allows these defaults to be overridden, or other options specified. The option, to be sent may be given as a decimal number or as "option:<option-name>" The option numbers are specified in RFC2132 and subsequent RFCs. The set of option-names known by dnsmasq can be discovered by running "dnsmasq --help dhcp". For example, to set the default route option to, do --dhcp-option=3, or --dhcp-option = option:router, and to set the time-server address to, do --dhcp-option = 42, or --dhcp-option = option:ntp-server, The special address is taken to mean "the address of the machine running dnsmasq". Data types allowed are comma separated dotted-quad IP addresses, a decimal number, colon-separated hex digits and a text string. If the optional network-ids are given then this option is only sent when all the network-ids are matched.

Special processing is done on a text argument for option 119, to conform with RFC 3397. Text or dotted-quad IP addresses as arguments to option 120 are handled as per RFC 3361. Dotted-quad IP addresses which are followed by a slash and then a netmask size are encoded as described in RFC 3442.

Be careful: no checking is done that the correct type of data for the option number is sent, it is quite possible to persuade dnsmasq to generate illegal DHCP packets with injudicious use of this flag. When the value is a decimal number, dnsmasq must determine how large the data item is. It does this by examining the option number and/or the value, but can be overridden by appending a single letter flag as follows: b = one byte, s = two bytes, i = four bytes. This is mainly useful with encapsulated vendor class options (see below) where dnsmasq cannot determine data size from the option number. Option data which consists solely of periods and digits will be interpreted by dnsmasq as an IP address, and inserted into an option as such. To force a literal string, use quotes. For instance when using option 66 to send a literal IP address as TFTP server name, it is necessary to do --dhcp-option=66,

Encapsulated Vendor-class options may also be specified using --dhcp-option: for instance --dhcp-option=vendor:PXEClient,1, sends the encapsulated vendor class-specific option "mftp-address=" to any client whose vendor-class matches "PXEClient". The vendor-class matching is substring based (see --dhcp-vendorclass for details). If a vendor-class option (number 60) is sent by dnsmasq, then that is used for selecting encapsulated options in preference to any sent by the client. It is possible to omit the vendorclass completely; --dhcp-option=vendor:,1, in which case the encapsulated option is always sent. The address is not treated specially in encapsulated vendor class options.

This works in exactly the same way as --dhcp-option except that the option will always be sent, even if the client does not ask for it in the parameter request list. This is sometimes needed, for example when sending options to PXELinux.
-U, --dhcp-vendorclass=<network-id>,<vendor-class>
Map from a vendor-class string to a network id tag. Most DHCP clients provide a "vendor class" which represents, in some sense, the type of host. This option maps vendor classes to tags, so that DHCP options may be selectively delivered to different classes of hosts. For example dhcp-vendorclass=printers,Hewlett-Packard JetDirect will allow options to be set only for HP printers like so: --dhcp-option=printers,3, The vendor-class string is substring matched against the vendor-class supplied by the client, to allow fuzzy matching.
-j, --dhcp-userclass=<network-id>,<user-class>
Map from a user-class string to a network id tag (with substring matching, like vendor classes). Most DHCP clients provide a "user class" which is configurable. This option maps user classes to tags, so that DHCP options may be selectively delivered to different classes of hosts. It is possible, for instance to use this to set a different printer server for hosts in the class "accounts" than for hosts in the class "engineering".
-4, --dhcp-mac=<network-id>,<MAC address>
Map from a MAC address to a network-id tag. The MAC address may include wildcards. For example --dhcp-mac=3com,01:34:23:*:*:* will set the tag "3com" for any host whose MAC address matches the pattern.
--dhcp-circuitid=<network-id>,<circuit-id>, --dhcp-remoteid=<network-id>,<remote-id>
Map from RFC3046 relay agent options to network-id tags. This data may be provided by DHCP relay agents. The circuit-id or remote-id is normally given as colon-separated hex, but is also allowed to be a simple string. If an exact match is achieved between the circuit or agent ID and one provided by a relay agent, the network-id tag is set.
Map from RFC3993 subscriber-d relay agent options to network-id tags.
-J, --dhcp-ignore=<network-id>[,<network-id>]
When all the given network-ids match the set of network-ids derived from the net, host, vendor and user classes, ignore the host and do not allocate it a DHCP lease.
When all the given network-ids match the set of network-ids derived from the net, host, vendor and user classes, ignore any hostname provided by the host. Note that, unlike dhcp-ignore, it is permissable to supply no netid tags, in which case DHCP-client supplied hostnames are always ignored, and DHCP hosts are added to the DNS using only dhcp-host configuration in dnsmasq and the contents of /etc/hosts and /etc/ethers.
-M, --dhcp-boot=[net:<network-id>,]<filename>,[<servername>[,<server address>]]
Set BOOTP options to be returned by the DHCP server. Server name and address are optional: if not provided, the name is left empty, and the address set to the address of the machine running dnsmasq. If dnsmasq is providing a TFTP service (see --enable-tftp ) then only the filename is required here to enable network booting. If the optional network-id(s) are given, they must match for this configuration to be sent. Note that network-ids are prefixed by "net:" to distinguish them.
-X, --dhcp-lease-max=<number>
Limits dnsmasq to the specified maximum number of DHCP leases. The default is 150. This limit is to prevent DoS attacks from hosts which create thousands of leases and use lots of memory in the dnsmasq process.
-K, --dhcp-authoritative
Should be set when dnsmasq is definitely the only DHCP server on a network. It changes the behaviour from strict RFC compliance so that DHCP requests on unknown leases from unknown hosts are not ignored. This allows new hosts to get a lease without a tedious timeout under all circumstances. It also allows dnsmasq to rebuild its lease database without each client needing to reaquire a lease, if the database is lost.
-3, --bootp-dynamic
Enable dynamic allocation of IP addresses to BOOTP clients. Use this with care, since each address allocated to a BOOTP client is leased forever, and therefore becomes permanently unavailable for re-use by other hosts.
-5, --no-ping
By default, the DHCP server will attempt to ensure that an address in not in use before allocating it to a host. It does this by sending an ICMP echo request (aka "ping") to the address in question. If it gets a reply, then the address must already be in use, and another is tried. This flag disables this check. Use with caution.
Extra logging for DHCP: log all the options sent to DHCP clients and the netid tags used to determine them.
-l, --dhcp-leasefile=<path>
Use the specified file to store DHCP lease information. If this option is given but no dhcp-range option is given then dnsmasq version 1 behaviour is activated. The file given is assumed to be an ISC dhcpd lease file and parsed for leases which are then added to the DNS system if they have a hostname. This functionality may have been excluded from dnsmasq at compile time, in which case an error will occur. In any case note that ISC leasefile integration is a deprecated feature. It should not be used in new installations, and will be removed in a future release.
-6 --dhcp-script=<path>
Whenever a new DHCP lease is created, or an old one destroyed, the binary specified by this option is run. The arguments to the process are "add", "old" or "del", the MAC address of the host (or "<null>"), the IP address, and the hostname, if known. "add" means a lease has been created, "del" means it has been destroyed, "old" is a notification of an existing lease when dnsmasq starts or a change to MAC address or hostname of an existing lease (also, lease length or expiry and client-id, if leasefile-ro is set). The process is run as root (assuming that dnsmasq was originally run as root) even if dnsmasq is configured to change UID to an unprivileged user. The environment is inherited from the invoker of dnsmasq, and if the host provided a client-id, this is stored in the environment variable DNSMASQ_CLIENT_ID. If the client provides vendor-class or user-class information, these are provided in DNSMASQ_VENDOR_CLASS and DNSMASQ_USER_CLASS0..DNSMASQ_USER_CLASSn variables, but only for "add" actions or "old" actions when a host resumes an existing lease, since these data are not held in dnsmasq's lease database. If dnsmasq was compiled with HAVE_BROKEN_RTC, then the length of the lease (in seconds) is stored in DNSMASQ_LEASE_LENGTH, otherwise the time of lease expiry is stored in DNSMASQ_LEASE_EXPIRES. If a lease used to have a hostname, which is removed, an "old" event is generated with the new state of the lease, ie no name, and the former name is provided in the environment variable DNSMASQ_OLD_HOSTNAME. All file decriptors are closed except stdin, stdout and stderr which are open to /dev/null (except in debug mode). The script is not invoked concurrently: if subsequent lease changes occur, the script is not invoked again until any existing invocation exits. At dnsmasq startup, the script will be invoked for all existing leases as they are read from the lease file. Expired leases will be called with "del" and others with "old". <path> must be an absolute pathname, no PATH search occurs.
-9, --leasefile-ro
Completely suppress use of the lease database file. The file will not be created, read, or written. Change the way the lease-change script (if one is provided) is called, so that the lease database may be maintained in external storage by the script. In addition to the invocations given in --dhcp-script the lease-change script is called once, at dnsmasq startup, with the single argument "init". When called like this the script should write the saved state of the lease database, in dnsmasq leasefile format, to stdout and exit with zero exit code. Setting this option also forces the leasechange script to be called on changes to the client-id and lease length and expiry time.
Treat DHCP request packets arriving at any of the <alias> interfaces as if they had arrived at <interface>. This option is only available on FreeBSD and Dragonfly BSD, and is necessary when using "old style" bridging, since packets arrive at tap interfaces which don't have an IP address.
-s, --domain=<domain>
Specifies the domain for the DHCP server. This has two effects; firstly it causes the DHCP server to return the domain to any hosts which request it, and secondly it sets the domain which it is legal for DHCP-configured hosts to claim. The intention is to constrain hostnames so that an untrusted host on the LAN cannot advertise its name via dhcp as e.g. "microsoft.com" and capture traffic not meant for it. If no domain suffix is specified, then any DHCP hostname with a domain part (ie with a period) will be disallowed and logged. If suffix is specified, then hostnames with a domain part are allowed, provided the domain part matches the suffix. In addition, when a suffix is set then hostnames without a domain part have the suffix added as an optional domain part. Eg on my network I can set --domain=thekelleys.org.uk and have a machine whose DHCP hostname is "laptop". The IP address for that machine is available from dnsmasq both as "laptop" and "laptop.thekelleys.org.uk". If the domain is given as "#" then the domain is read from the first "search" directive in /etc/resolv.conf (or equivalent).
Enable the TFTP server function. This is deliberately limited to that needed to net-boot a client: Only reading is allowed, and only in binary/octet mode. The tsize and blksize extensions are supported.
Look for files to transfer using TFTP relative to the given directory. When this is set, TFTP paths which include ".." are rejected, to stop clients getting outside the specified root. Absolute paths (starting with /) are allowed, but they must be within the tftp-root.
Enable TFTP secure mode: without this, any file which is readble by the dnsmasq process under normal unix access-control rules is available via TFTP. When the --tftp-secure flag is given, only files owned by the user running the dnsmasq process are accessible. If dnsmasq is being run as root, different rules apply: --tftp-secure has no effect, but only files which have the world-readable bit set are accessible. It is not recommended to run dnsmasq as root with TFTP enabled, and certainly not without specifying --tftp-root. Doing so can expose any world-readable file on the server to any host on the net.
Set the maximum number of concurrent TFTP connections allowed. This defaults to 50. When serving a large number of TFTP connections, per-process file descriptor limits may be encountered. Dnsmasq needs one file descriptor for each concurrent TFTP connection and one file descriptor per unique file (plus a few others). So serving the same file simultaneously to n clients will use require about n + 10 file descriptors, serving different files simultaneously to n clients will require about (2*n) + 10 descriptors.
Stop the TFTP server from negotiating the "blocksize" option with a client. Some buggy clients request this option but then behave badly when it is granted.
-C, --conf-file=<file>
Specify a different configuration file. The conf-file option is also allowed in configuration files, to include multiple configuration files.
-7, --conf-dir=<directory>
Read all the files in the given directory as configuration files. Files whose names end in ~ or start with . or start and end with # are skipped. This flag may be given on the command line or in a configuration file.


At startup, dnsmasq reads /etc/dnsmasq.conf, if it exists. (On FreeBSD, the file is /usr/local/etc/dnsmasq.conf ) (but see the -C and -7 options.) The format of this file consists of one option per line, exactly as the long options detailed in the OPTIONS section but without the leading "--". Lines starting with # are comments and ignored. For options which may only be specified once, the configuration file overrides the command line. Quoting is allowed in a config file: between " quotes the special meanings of ,:. and # are removed and the following escapes are allowed: \\ \" \t \a \b \r and \n. The later corresponding to tab, bell, backspace, return and newline.  


When it receives a SIGHUP, dnsmasq clears its cache and then re-loads /etc/hosts and /etc/ethers. If --no-poll is set SIGHUP also re-reads /etc/resolv.conf. SIGHUP does NOT re-read the configuration file.

When it receives a SIGUSR1, dnsmasq writes cache statistics to the system log. It writes the cache size, the number of names which have had to removed from the cache before they expired in order to make room for new names and the total number of names that have been inserted into the cache. In --no-daemon mode or when full logging is enabled (-q), a complete dump of the contents of the cache is made.

Dnsmasq is a DNS query forwarder: it it not capable of recursively answering arbitrary queries starting from the root servers but forwards such queries to a fully recursive upstream DNS server which is typically provided by an ISP. By default, dnsmasq reads /etc/resolv.conf to discover the IP addresses of the upstream nameservers it should use, since the information is typically stored there. Unless --no-poll is used, dnsmasq checks the modification time of /etc/resolv.conf (or equivalent if --resolv-file is used) and re-reads it if it changes. This allows the DNS servers to be set dynamically by PPP or DHCP since both protocols provide the information. Absence of /etc/resolv.conf is not an error since it may not have been created before a PPP connection exists. Dnsmasq simply keeps checking in case /etc/resolv.conf is created at any time. Dnsmasq can be told to parse more than one resolv.conf file. This is useful on a laptop, where both PPP and DHCP may be used: dnsmasq can be set to poll both /etc/ppp/resolv.conf and /etc/dhcpc/resolv.conf and will use the contents of whichever changed last, giving automatic switching between DNS servers.

Upstream servers may also be specified on the command line or in the configuration file. These server specifications optionally take a domain name which tells dnsmasq to use that server only to find names in that particular domain.

In order to configure dnsmasq to act as cache for the host on which it is running, put "nameserver" in /etc/resolv.conf to force local processes to send queries to dnsmasq. Then either specify the upstream servers directly to dnsmasq using --server options or put their addresses real in another file, say /etc/resolv.dnsmasq and run dnsmasq with the -r /etc/resolv.dnsmasq option. This second technique allows for dynamic update of the server addresses by PPP or DHCP.

Addresses in /etc/hosts will "shadow" different addresses for the same names in the upstream DNS, so "mycompany.com" in /etc/hosts will ensure that queries for "mycompany.com" always return even if queries in the upstream DNS would otherwise return a different address. There is one exception to this: if the upstream DNS contains a CNAME which points to a shadowed name, then looking up the CNAME through dnsmasq will result in the unshadowed address associated with the target of the CNAME. To work around this, add the CNAME to /etc/hosts so that the CNAME is shadowed too.

The network-id system works as follows: For each DHCP request, dnsmasq collects a set of valid network-id tags, one from the dhcp-range used to allocate the address, one from any matching dhcp-host and possibly many from matching vendor classes and user classes sent by the DHCP client. Any dhcp-option which has network-id tags will be used in preference to an untagged dhcp-option, provided that _all_ the tags match somewhere in the set collected as described above. The prefix '#' on a tag means 'not' so --dhcp=option=#purple,3, sends the option when the network-id tag purple is not in the set of valid tags.

If the network-id in a dhcp-range is prefixed with 'net:' then its meaning changes from setting a tag to matching it. Thus if there is more than dhcp-range on a subnet, and one is tagged with a network-id which is set (for instance from a vendorclass option) then hosts which set the netid tag will be allocated addresses in the tagged range.

The DHCP server in dnsmasq will function as a BOOTP server also, provided that the MAC address and IP address for clients are given, either using dhcp-host configurations or in /etc/ethers , and a dhcp-range configuration option is present to activate the DHCP server on a particular network. (Setting --bootp-dynamic removes the need for static address mappings.) The filename parameter in a BOOTP request is matched against netids in dhcp-option configurations, as is the tag "bootp", allowing some control over the options returned to different classes of hosts.



The default values for resource limits in dnsmasq are generally conservative, and appropriate for embedded router type devices with slow processors and limited memory. On more capable hardware, it is possible to increase the limits, and handle many more clients. The following applies to dnsmasq-2.37: earlier versions did not scale as well.

Dnsmasq is capable of handling DNS and DHCP for at least a thousand clients. Clearly to do this the value of --dhcp-lease-max must be increased, and lease times should not be very short (less than one hour). The value of --dns-forward-max can be increased: start with it equal to the number of clients and increase if DNS seems slow. Note that DNS performance depends too on the performance of the upstream nameservers. The size of the DNS cache may be increased: the hard limit is 10000 names and the default (150) is very low. Sending SIGUSR1 to dnsmasq makes it log information which is useful for tuning the cache size. See the NOTES section for details.

The built-in TFTP server is capable of many simultaneous file transfers: the absolute limit is related to the number of file-handles allowed to a process and the ability of the select() system call to cope with large numbers of file handles. If the limit is set too high using --tftp-max it will be scaled down and the actual limit logged at start-up. Note that more transfers are possible when the same file is being sent than when each transfer sends a different file.

It is possible to use dnsmasq to block Web advertising by using a list of known banner-ad servers, all resolving to or, in /etc/hosts or an additional hosts file. The list can be very long, dnsmasq has been tested successfully with one million names. That size file needs a 1GHz processor and about 60Mb of RAM.












hosts(5), resolver(5)  


This manual page was written by Simon Kelley <simon@thekelleys.org.uk>.




linux.jgfs.net manual pages