4039N/A SENDMAIL CONFIGURATION FILES
4039N/AThis document describes the sendmail configuration files. It
4039N/AIt also describes how to set options for sendmail which are explained
4039N/Ain the Sendmail Installation and Operation guide, which can be found
4039N/ARecall this URL throughout this document when references to
4039N/AUSING LDAP FOR ALIASES, MAPS, AND CLASSES
4039N/AANTI-SPAM CONFIGURATION CONTROL
4039N/AADDING NEW MAILERS OR RULESETS
4039N/ANON-SMTP BASED CONFIGURATIONS
4039N/AACCEPTING MAIL FOR MULTIPLE NAMES
4039N/AUSING USERDB TO MAP FULL NAMES
4039N/AMISCELLANEOUS SPECIAL FEATURES
4039N/ATWEAKING CONFIGURATION OPTIONS
4039N/A+--------------------------+
4039N/A| INTRODUCTION AND EXAMPLE |
4039N/A+--------------------------+
4039N/AConfiguration files are contained in the subdirectory "cf", with a
4039N/Asuffix ".mc". They must be run through "m4" to produce a ".cf" file.
4039N/AAlternatively, you can simply:
4039N/Aname of your configuration file. If you are running a version of M4
4039N/Athat understands the __file__ builtin (versions of GNU m4 >= 0.75 do
4039N/Athis, but the versions distributed with 4.4BSD and derivatives do not)
4039N/Aor the -I flag (ditto), then ${CFDIR} can be in an arbitrary directory.
4039N/AFor "traditional" versions, ${CFDIR} ***MUST*** be "..", or you MUST
4039N/ALet's examine a typical .mc file:
4039N/A # Copyright (c) 1998-2005 Sendmail, Inc. and its suppliers.
4039N/A # Copyright (c) 1983 Eric P. Allman. All rights reserved.
4039N/A # The Regents of the University of California. All rights reserved.
4039N/A # By using this file, you agree to the terms and conditions set
4039N/A # forth in the LICENSE file which can be found at the top level of
4039N/A # the sendmail distribution.
4039N/A # This is a Berkeley-specific configuration file for HP-UX
9.x. 4039N/A # It applies only to the Computer Science Division at Berkeley,
4039N/A # and should not be used elsewhere. It is provided on the sendmail
4039N/A # distribution as a sample only. To create your own configuration
4039N/A # file, create an appropriate domain file in ../domain, change the
4039N/A # `DOMAIN' macro below to reference that file, and copy the result
4039N/A # to a name of your own choosing.
4039N/AThe divert(-1) will delete the crud in the resulting output file.
4039N/AThe copyright notice can be replaced by whatever your lawyers require;
4039N/Aour lawyers require the one that is included in these files. A copyleft
4039N/Ais a copyright by another name. The divert(0) restores regular output.
4039N/A VERSIONID(`<SCCS or RCS version id>')
4039N/AVERSIONID is a macro that stuffs the version information into the
4039N/Aresulting file. You could use SCCS, RCS, CVS, something else, or
4039N/Aomit it completely. This is not the same as the version id included
4039N/AYou must specify an OSTYPE to properly configure things such as the
4039N/Apathname of the help and status files, the flags needed for the local
4039N/Amailer, and other important things. If you omit it, you will get an
4039N/Aerror when you try to build the configuration. Look at the ostype
4039N/Adirectory for the list of known operating system types.
4039N/AThis example is specific to the Computer Science Division at Berkeley.
4039N/AYou can use "DOMAIN(`generic')" to get a sufficiently bland definition
4039N/Athat may well work for you, or you can create a customized domain
4039N/Adefinition appropriate for your environment.
4039N/AThese describe the mailers used at the default CS site. The local
4039N/Amailer is always included automatically. Beware: MAILER declarations
4039N/Ashould only be followed by LOCAL_* sections. The general rules are
4039N/AThere are a few exceptions to this rule. Local macro definitions which
4039N/Ainfluence a FEATURE() should be done before that feature. For example,
4039N/Aa define(`PROCMAIL_MAILER_PATH', ...) should be done before
4039N/A+----------------------------+
4039N/A| A BRIEF INTRODUCTION TO M4 |
4039N/A+----------------------------+
4039N/ASendmail uses the M4 macro processor to ``compile'' the configuration
4039N/Afiles. The most important thing to know is that M4 is stream-based,
4039N/Athat is, it doesn't understand about lines. For this reason, in some
4039N/Aplaces you may see the word ``dnl'', which stands for ``delete
4039N/Athrough newline''; essentially, it deletes all characters starting
4039N/Aat the ``dnl'' up to and including the next newline character. In
4039N/Amost cases sendmail uses this only to avoid lots of unnecessary
4039N/AOther important directives are define(A, B) which defines the macro
4039N/A``A'' to have value ``B''. Macros are expanded as they are read, so
4039N/Aone normally quotes both values to prevent expansion. For example,
4039N/AOne word of warning: M4 macros are expanded even in lines that appear
4039N/Ato be comments. For example, if you have
4039N/Ait will not do what you expect, because the FEATURE(`foo') will be
4039N/Aexpanded. This also applies to
4039N/A # And then define the $X macro to be the return address
4039N/Abecause ``define'' is an M4 keyword. If you want to use them, surround
4039N/Athem with directed quotes, `like this'.
4039N/ASince m4 uses single quotes (opening "`" and closing "'") to quote
4039N/Aarguments, those quotes can't be used in arguments. For example,
4039N/Ait is not possible to define a rejection message containing a single
4039N/Aquote. Usually there are simple workarounds by changing those
4039N/Amessages; in the worst case it might be ok to change the value
4039N/Adirectly in the generated .cf file, which however is not advised.
4039N/Asendmail 8.9 has introduced a new configuration directory for sendmail
4039N/A8.10, all files will use this directory by default (some options may be
4039N/Aset by OSTYPE() files). This new directory should help to restore
4039N/Auniformity to sendmail's file locations.
4039N/ABelow is a table of some of the common changes:
4039N/ANote that all of these paths actually use a new m4 macro MAIL_SETTINGS_DIR
4039N/Ato create the pathnames. The default value of this variable is
4039N/A`
/etc/mail/'. If you set this macro to a different value, you MUST include
4039N/ANotice: all filenames used in a .mc (or .cf) file should be absolute
4039N/A(starting at the root,
i.e., with '/'). Relative filenames most
4039N/Alikely cause surprises during operations (unless otherwise noted).
4039N/AYou MUST define an operating system environment, or the configuration
4039N/Afile build will puke. There are several environments available; look
4039N/Aat the "ostype" directory for the current list. This macro changes
4039N/Athings like the location of the alias file and queue directory. Some
4039N/Aof these files are identical to one another.
4039N/AIt is IMPERATIVE that the OSTYPE occur before any MAILER definitions.
4039N/AIn general, the OSTYPE macro should go immediately after any version
4039N/Ainformation, and MAILER definitions should always go last.
4039N/AOperating system definitions are usually easy to write. They may define
4039N/Athe following variables (everything defaults, so an ostype file may be
4039N/Aempty). Unfortunately, the list of configuration-supported systems is
4039N/Anot as broad as the list of source-supported systems, since many of
4039N/Athe source contributors do not include corresponding ostype files.
4039N/A of the alias file(s). It can be a comma-separated
4039N/A list of names (but be sure you quote values with
4039N/A commas in them -- for example, use
4039N/A define(`ALIAS_FILE', `a,b')
4039N/A to get "a" and "b" both listed as alias files;
4039N/A otherwise the define() primitive only sees "a").
4039N/A containing information printed in response to
4039N/A queue files. To use multiple queues, supply
4039N/A a value ending with an asterisk. For
4039N/A directories or symbolic links to directories
4039N/A directories. The names 'qf', 'df', and 'xf' are
4039N/A reserved as specific subdirectories for the
4039N/A corresponding queue file types as explained in
4039N/A queue files for the MSP (Mail Submission Program).
4039N/ALOCAL_MAILER_FLAGS [Prmn9] The flags used by the local mailer. The
4039N/A flags lsDFMAw5:/|@q are always included.
4039N/ALOCAL_MAILER_ARGS [mail -d $u] The arguments passed to deliver local
4039N/ALOCAL_MAILER_MAX [undefined] If defined, the maximum size of local
4039N/A mail that you are willing to accept.
4039N/ALOCAL_MAILER_MAXMSGS [undefined] If defined, the maximum number of
4039N/A messages to deliver in a single connection. Only
4039N/A useful for LMTP local mailers.
4039N/ALOCAL_MAILER_CHARSET [undefined] If defined, messages containing 8-bit data
4039N/A that ARRIVE from an address that resolves to the
4039N/A local mailer and which are converted to MIME will be
4039N/A labeled with this character set.
4039N/ALOCAL_MAILER_EOL [undefined] If defined, the string to use as the
4039N/A end of line for the local mailer.
4039N/ALOCAL_MAILER_DSN_DIAGNOSTIC_CODE
4039N/A [X-Unix] The DSN Diagnostic-Code value for the
4039N/A local mailer. This should be changed with care.
4039N/ALOCAL_SHELL_PATH [
/bin/sh] The shell used to deliver piped email.
4039N/ALOCAL_SHELL_FLAGS [eu9] The flags used by the shell mailer. The
4039N/A flags lsDFM are always included.
4039N/ALOCAL_SHELL_ARGS [sh -c $u] The arguments passed to deliver "prog"
4039N/ALOCAL_SHELL_DIR [$z:/] The directory search path in which the
4039N/ALOCAL_MAILER_QGRP [undefined] The queue group for the local mailer.
4039N/ASMTP_MAILER_FLAGS [undefined] Flags added to SMTP mailer. Default
4039N/A flags are `mDFMuX' for all SMTP-based mailers; the
4039N/A "esmtp" mailer adds `a'; "smtp8" adds `8'; and
4039N/ARELAY_MAILER_FLAGS [undefined] Flags added to the relay mailer. Default
4039N/A flags are `mDFMuX' for all SMTP-based mailers; the
4039N/A relay mailer adds `a8'. If this is not defined,
4039N/A then SMTP_MAILER_FLAGS is used.
4039N/ASMTP_MAILER_MAX [undefined] The maximum size of messages that will
4039N/A be transported using the smtp, smtp8, esmtp, or dsmtp
4039N/ASMTP_MAILER_MAXMSGS [undefined] If defined, the maximum number of
4039N/A messages to deliver in a single connection for the
4039N/A smtp, smtp8, esmtp, or dsmtp mailers.
4039N/ASMTP_MAILER_MAXRCPTS [undefined] If defined, the maximum number of
4039N/A recipients to deliver in a single connection for the
4039N/A smtp, smtp8, esmtp, or dsmtp mailers.
4039N/ASMTP_MAILER_ARGS [TCP $h] The arguments passed to the smtp mailer.
4039N/A About the only reason you would want to change this
4039N/A would be to change the default port.
4039N/AESMTP_MAILER_ARGS [TCP $h] The arguments passed to the esmtp mailer.
4039N/ASMTP8_MAILER_ARGS [TCP $h] The arguments passed to the smtp8 mailer.
4039N/ADSMTP_MAILER_ARGS [TCP $h] The arguments passed to the dsmtp mailer.
4039N/ARELAY_MAILER_ARGS [TCP $h] The arguments passed to the relay mailer.
4039N/ASMTP_MAILER_QGRP [undefined] The queue group for the smtp mailer.
4039N/AESMTP_MAILER_QGRP [undefined] The queue group for the esmtp mailer.
4039N/ASMTP8_MAILER_QGRP [undefined] The queue group for the smtp8 mailer.
4039N/ADSMTP_MAILER_QGRP [undefined] The queue group for the dsmtp mailer.
4039N/ARELAY_MAILER_QGRP [undefined] The queue group for the relay mailer.
4039N/ARELAY_MAILER_MAXMSGS [undefined] If defined, the maximum number of
4039N/A messages to deliver in a single connection for the
4039N/ASMTP_MAILER_CHARSET [undefined] If defined, messages containing 8-bit data
4039N/A that ARRIVE from an address that resolves to one of
4039N/A the SMTP mailers and which are converted to MIME will
4039N/A be labeled with this character set.
4039N/ASMTP_MAILER_LL [990] The maximum line length for SMTP mailers
4039N/ARELAY_MAILER_LL [2040] The maximum line length for the relay mailer.
4039N/AUUCP_MAILER_FLAGS [undefined] Flags added to UUCP mailer. Default
4039N/A flags are `DFMhuU' (and `m' for uucp-new mailer,
4039N/A minus `U' for uucp-dom mailer).
4039N/AUUCP_MAILER_ARGS [uux - -r -z -a$g -gC $h!rmail ($u)] The arguments
4039N/AUUCP_MAILER_MAX [100000] The maximum size message accepted for
4039N/A transmission by the UUCP mailers.
4039N/AUUCP_MAILER_CHARSET [undefined] If defined, messages containing 8-bit data
4039N/A that ARRIVE from an address that resolves to one of
4039N/A the UUCP mailers and which are converted to MIME will
4039N/A be labeled with this character set.
4039N/AUUCP_MAILER_QGRP [undefined] The queue group for the UUCP mailers.
4039N/A program. This is also used by
4039N/APROCMAIL_MAILER_FLAGS [SPhnu9] Flags added to Procmail mailer. Flags
4039N/A DFM are always set. This is NOT used by
4039N/A FEATURE(`local_procmail'); tweak LOCAL_MAILER_FLAGS
4039N/APROCMAIL_MAILER_ARGS [procmail -Y -m $h $f $u] The arguments passed to
4039N/A the Procmail mailer. This is NOT used by
4039N/A FEATURE(`local_procmail'); tweak LOCAL_MAILER_ARGS
4039N/APROCMAIL_MAILER_MAX [undefined] If set, the maximum size message that
4039N/A will be accepted by the procmail mailer.
4039N/APROCMAIL_MAILER_QGRP [undefined] The queue group for the procmail mailer.
4039N/A Currently used for FEATURE(`local_lmtp') and
4039N/ALOCAL_PROG_QGRP [undefined] The queue group for the prog mailer.
4039N/ANote: to tweak Name_MAILER_FLAGS use the macro MODIFY_MAILER_FLAGS:
4039N/AMODIFY_MAILER_FLAGS(`Name', `change') where Name is the first part
4039N/Aof the macro Name_MAILER_FLAGS (note: that means Name is entirely in
4039N/Aupper case) and change can be: flags that should be used directly
4039N/A(thus overriding the default value), or if it starts with `+' (`-')
4039N/Athen those flags are added to (removed from) the default value.
4039N/A MODIFY_MAILER_FLAGS(`LOCAL', `+e')
4039N/Awill add the flag `e' to LOCAL_MAILER_FLAGS. Notice: there are
4039N/Aseveral smtp mailers all of which are manipulated individually.
4039N/ASee the section MAILERS for the available mailer names.
4039N/AWARNING: The FEATUREs local_lmtp and local_procmail set LOCAL_MAILER_FLAGS
4039N/Aunconditionally,
i.e., without respecting any definitions in an
4039N/AYou will probably want to collect domain-dependent defines into one
4039N/Afile, referenced by the DOMAIN macro. For example, the Berkeley
4039N/Adomain file includes definitions for several internal distinguished
4039N/AUUCP_RELAY The host that will accept UUCP-addressed email.
4039N/A If not defined, all UUCP sites must be directly
4039N/ABITNET_RELAY The host that will accept BITNET-addressed email.
4039N/A If not defined, the .BITNET pseudo-domain won't work.
4039N/ADECNET_RELAY The host that will accept DECNET-addressed email.
4039N/A If not defined, the .DECNET pseudo-domain and addresses
4039N/A of the form node::user will not work.
4039N/AFAX_RELAY The host that will accept mail to the .FAX pseudo-domain.
4039N/A The "fax" mailer overrides this value.
4039N/ALOCAL_RELAY The site that will handle unqualified names -- that
4039N/A is, names without an @domain extension.
4039N/A Normally MAIL_HUB is preferred for this function.
4039N/A LOCAL_RELAY is mostly useful in conjunction with
4039N/A FEATURE(`stickyhost') -- see the discussion of
4039N/A stickyhost below. If not set, they are assumed to
4039N/A belong on this machine. This allows you to have a
4039N/A central site to store a company- or department-wide
4039N/A alias database. This only works at small sites,
4039N/A and only with some user agents.
4039N/ALUSER_RELAY The site that will handle lusers -- that is, apparently
4039N/A local names that aren't local accounts or aliases. To
4039N/A specify a local user instead of a site, set this to
4039N/AAny of these can be either ``mailer:hostname'' (in which case the
4039N/Amailer is the internal mailer name, such as ``uucp-new'' and the hostname
4039N/Ais the name of the host as appropriate for that mailer) or just a
4039N/A``hostname'', in which case a default mailer type (usually ``relay'',
4039N/Aa variant on SMTP) is used. WARNING: if you have a wildcard MX
4039N/Arecord matching your domain, you probably want to define these to
4039N/Ahave a trailing dot so that you won't get the mail diverted back
4039N/AThe domain file can also be used to define a domain name, if needed
4039N/A(using "DD<domain>") and set certain site-wide features. If all hosts
4039N/Aat your site masquerade behind one email name, you could also use
4039N/AYou do not have to define a domain -- in particular, if you are a
4039N/Asingle machine sitting off somewhere, it is probably more work than
4039N/Ait's worth. This is just a mechanism for combining "domain dependent
4039N/AThere are fewer mailers supported in this version than the previous
4039N/Aversion, owing mostly to a simpler world. As a general rule, put the
4039N/AMAILER definitions last in your .mc file.
4039N/Alocal The local and prog mailers. You will almost always
4039N/A need these; the only exception is if you relay ALL
4039N/A your mail to another site. This mailer is included
4039N/Asmtp The Simple Mail Transport Protocol mailer. This does
4039N/A not hide hosts behind a gateway or another other
4039N/A such hack; it assumes a world where everyone is
4039N/A running the name server. This file actually defines
4039N/A five mailers: "smtp" for regular (old-style) SMTP to
4039N/A other servers, "esmtp" for extended SMTP to other
4039N/A servers, "smtp8" to do SMTP to other servers without
4039N/A converting 8-bit data to MIME (essentially, this is
4039N/A your statement that you know the other end is 8-bit
4039N/A clean even if it doesn't say so), "dsmtp" to do on
4039N/A demand delivery, and "relay" for transmission to the
4039N/A RELAY_HOST, LUSER_RELAY, or MAIL_HUB.
4039N/Auucp The UNIX-to-UNIX Copy Program mailer. Actually, this
4039N/A know that the UUCP mailer at the other end can handle
4039N/A multiple recipients in one transfer. If the smtp mailer
4039N/A is included in your configuration, two other mailers
4039N/A ("uucp-dom" and "uucp-uudom") are also defined [warning: you
4039N/A MUST specify MAILER(`smtp') before MAILER(`uucp')]. When you
4039N/A include the uucp mailer, sendmail looks for all names in
4039N/A class {U} and sends them to the uucp-old mailer; all
4039N/A names in class {Y} are sent to uucp-new; and all
4039N/A names in class {Z} are sent to uucp-uudom. Note that
4039N/A this is a function of what version of rmail runs on
4039N/A the receiving end, and hence may be out of your control.
4039N/A See the section below describing UUCP mailers in more
4039N/Aprocmail An interface to procmail (does not come with sendmail).
4039N/A This is designed to be used in mailertables. For example,
4039N/A a common question is "how do I forward all mail for a given
4039N/A domain to a single person?". If you have this mailer
4039N/A defined, you could set up a mailertable reading:
4039N/A ! -oi -f $1 person@other.host
4039N/A to person@other.host. In a procmail script, $1 is the
4039N/A name of the sender and $2 is the name of the recipient.
4039N/A If you use this with FEATURE(`local_procmail'), the FEATURE
4039N/A Of course there are other ways to solve this particular
4039N/A problem,
e.g., a catch-all entry in a virtusertable.
4039N/AThe local mailer accepts addresses of the form "user+detail", where
4039N/Athe "+detail" is not used for mailbox matching but is available
4039N/Ato certain local mail programs (in particular, see
4039N/AFEATURE(`local_procmail')). For example, "eric", "eric+sendmail", and
4039N/A"eric+sww" all indicate the same user, but additional arguments <null>,
4039N/A"sendmail", and "sww" may be provided for use in sorting mail.
4039N/ASpecial features can be requested using the "FEATURE" macro. For
4039N/Afile to get values for class {w}. A FEATURE may contain up to 9
4039N/Aoptional parameters -- for example:
4039N/AThe default database map type for the table features can be set with
4039N/A define(`DATABASE_MAP_TYPE', `dbm')
4039N/Awhich would set it to use ndbm databases. The default is the Berkeley DB
4039N/Ahash database format. Note that you must still declare a database map type
4039N/Aif you specify an argument to a FEATURE. DATABASE_MAP_TYPE is only used
4039N/Aif no argument is given for the FEATURE. It must be specified before any
4039N/AAlso, features which can take a map definition as an argument can also take
4039N/Athe special keyword `LDAP'. If that keyword is used, the map will use the
4039N/ALDAP definition described in the ``USING LDAP FOR ALIASES, MAPS, AND
4039N/A alternate names for this host. This might be used if you
4039N/A were on a host that MXed for a dynamic set of other hosts.
4039N/A If the set is static, just including the line "Cw<name1>
4039N/A <name2> ..." (where the names are fully qualified domain
4039N/A names) is probably superior. The actual filename can be
4039N/A overridden by redefining confCW_FILE.
4039N/A names of users that will be ``trusted'', that is, able to
4039N/A set their envelope from address using -f without generating
4039N/A a warning message. The actual filename can be overridden
4039N/A a ``551 User has moved; please try <address>'' message.
4039N/A If this is set, you can alias people who have left
4039N/A to their new address with ".REDIRECT" appended.
4039N/Anouucp Don't route UUCP addresses. This feature takes one
4039N/A `reject': reject addresses which have "!" in the local
4039N/A part unless it originates from a system
4039N/A `nospecial': don't do anything special with "!".
4039N/A Warnings: 1. See the notice in the anti-spam section.
4039N/A 2. don't remove "!" from OperatorChars if `reject' is
4039N/Anocanonify Don't pass addresses to $[ ... $] for canonification
4039N/A except for unqualified names, which must not be used in this
4039N/A mode (violation of the standard). It can be changed by
4039N/A setting the DaemonPortOptions modifiers (M=). That is,
4039N/A FEATURE(`nocanonify') will be overridden by setting the
4039N/A 'c' flag. Conversely, if FEATURE(`nocanonify') is not used,
4039N/A it can be emulated by setting the 'C' flag
4039N/A (DaemonPortOptions=Modifiers=C). This would generally only
4039N/A be used by sites that only act as mail gateways or which have
4039N/A user agents that do full canonification themselves. You may
4039N/A "define(`confBIND_OPTS', `-DNSRCH -DEFNAMES')" to turn off
4039N/A the usual resolver options that do a similar thing.
4039N/A An exception list for FEATURE(`nocanonify') can be
4039N/A specified with CANONIFY_DOMAIN or CANONIFY_DOMAIN_FILE,
4039N/A i.e., a list of domains which are nevertheless passed to
4039N/A $[ ... $] for canonification. This is useful to turn on
4039N/A Another way to require canonification in the local
4039N/A domain is CANONIFY_DOMAIN(`$=m').
4039N/A A trailing dot is added to addresses with more than
4039N/A one component in it such that other features which
4039N/A expect a trailing dot (
e.g., virtusertable) will
4039N/A If `canonify_hosts' is specified as parameter,
i.e.,
4039N/A FEATURE(`nocanonify', `canonify_hosts'), then
4039N/A <user@host>, will be canonified (and hopefully fully
4039N/Astickyhost This feature is sometimes used with LOCAL_RELAY,
4039N/A although it can be used for a different effect with
4039N/A When used without MAIL_HUB, email sent to
4039N/A "user@local.host" are marked as "sticky" -- that
4039N/A is, the local addresses aren't matched against UDB,
4039N/A don't go through ruleset 5, and are not forwarded to
4039N/A the LOCAL_RELAY (if defined).
4039N/A With MAIL_HUB, mail addressed to "user@local.host"
4039N/A is forwarded to the mail hub, with the envelope
4039N/A address still remaining "user@local.host".
4039N/A Without stickyhost, the envelope would be changed
4039N/A to "user@mail_hub", in order to protect against
4039N/Amailertable Include a "mailer table" which can be used to override
4039N/A routing for particular domains (which are not in class {w},
4039N/A i.e. local host names). The argument of the FEATURE may be
4039N/A the key definition. If none is specified, the definition
4039N/A Keys in this database are fully qualified domain names
4039N/A or partial domains preceded by a dot -- for example,
4039N/A special case of the latter, "." matches any domain not
4039N/A covered by other keys. Values must be of the form:
4039N/A where "mailer" is the internal mailer name, and "domain"
4039N/A is where to send the message. These maps are not
4039N/A reflected into the message header. As a special case,
4039N/A will forward to the indicated user using the local mailer,
4039N/A will forward to the original user in the e-mail address
4039N/A using the local mailer, and
4039N/A will give an error message with the indicated SMTP reply
4039N/A code and message, where
D.S.N is an RFC 1893 compliant
4039N/Adomaintable Include a "domain table" which can be used to provide
4039N/A domain name mapping. Use of this should really be
4039N/A limited to your own domains. It may be useful if you
4039N/A change names (
e.g., your company changes names from
4039N/A FEATURE may be the key definition. If none is specified,
4039N/A The key in this table is the domain name; the value is
4039N/A the new (fully qualified) domain. Anything in the
4039N/A domaintable is reflected into headers; that is, this
4039N/Abitdomain Look up bitnet hosts in a table to try to turn them into
4039N/A internet addresses. The table can be built using the
4039N/A bitdomain program contributed by John Gardiner Myers.
4039N/A The argument of the FEATURE may be the key definition; if
4039N/A none is specified, the definition used is:
4039N/A Keys are the bitnet hostname; values are the corresponding
4039N/Auucpdomain Similar feature for UUCP hosts. The default map definition
4039N/A At the moment there is no automagic tool to build this
4039N/A Include the local host domain even on locally delivered
4039N/A mail. Normally it is not added on unqualified names.
4039N/A However, if you use a shared message store but do not use
4039N/A the same user name space everywhere, you may need the host
4039N/A name on local names. An optional argument specifies
4039N/A another domain to be added than the local.
4039N/Aallmasquerade If masquerading is enabled (using MASQUERADE_AS), this
4039N/A feature will cause recipient addresses to also masquerade
4039N/A as being from the masquerade host. Normally they get
4039N/A the local hostname. Although this may be right for
4039N/A ordinary users, it can break local aliases. For example,
4039N/A if you send to "localalias", the originating sendmail will
4039N/A find that alias and send to all members, but send the
4039N/A message with "To: localalias@masqueradehost". Since that
4039N/A alias likely does not exist, replies will fail. Use this
4039N/A feature ONLY if you can guarantee that the ENTIRE
4039N/A namespace on your masquerade host supersets all the
4039N/A Normally, any hosts listed in class {w} are masqueraded. If
4039N/A this feature is given, only the hosts listed in class {M} (see
4039N/A below: MASQUERADE_DOMAIN) are masqueraded. This is useful
4039N/A if you have several domains with disjoint namespaces hosted
4039N/A If masquerading is enabled (using MASQUERADE_AS) and
4039N/A MASQUERADE_DOMAIN (see below) is set, this feature will
4039N/A cause addresses to be rewritten such that the masquerading
4039N/A domains are actually entire domains to be hidden. All
4039N/A hosts within the masquerading domains will be rewritten
4039N/A to the masquerade name (used in MASQUERADE_AS). For example,
4039N/A NOTE: only domains within your jurisdiction and
4039N/A current hierarchy should be masqueraded using this.
4039N/A This feature prevents the local mailer from masquerading even
4039N/A if MASQUERADE_AS is used. MASQUERADE_AS will only have effect
4039N/A on addresses of mail going outside the local domain.
4039N/A If masquerading is enabled (using MASQUERADE_AS) or the
4039N/A genericstable is in use, this feature will cause envelope
4039N/A addresses to also masquerade as being from the masquerade
4039N/A host. Normally only the header addresses are masqueraded.
4039N/Agenericstable This feature will cause unqualified addresses (
i.e., without
4039N/A a domain) and addresses with a domain listed in class {G}
4039N/A to be looked up in a map and turned into another ("generic")
4039N/A form, which can change both the domain name and the user name.
4039N/A Notice: if you use an MSP (as it is default starting with
4039N/A 8.12), the MTA will only receive qualified addresses from the
4039N/A MSP (as required by the RFCs). Hence you need to add your
4039N/A domain to class {G}. This feature is similar to the userdb
4039N/A functionality. The same types of addresses as for
4039N/A masquerading are looked up,
i.e., only header sender
4039N/A addresses unless the allmasquerade
and/or masquerade_envelope
4039N/A features are given. Qualified addresses must have the domain
4039N/A part in class {G}; entries can be added to this class by the
4039N/A macros GENERICS_DOMAIN or GENERICS_DOMAIN_FILE (analogously
4039N/A to MASQUERADE_DOMAIN and MASQUERADE_DOMAIN_FILE, see below).
4039N/A The argument of FEATURE(`genericstable') may be the map
4039N/A definition; the default map definition is:
4039N/A The key for this table is either the full address, the domain
4039N/A (with a leading @; the localpart is passed as first argument)
4039N/A or the unqualified username (tried in the order mentioned);
4039N/A the value is the new user address. If the new user address
4039N/A does not include a domain, it will be qualified in the standard
4039N/A manner,
i.e., using $j or the masquerade name. Note that the
4039N/A address being looked up must be fully qualified. For local
4039N/A mail, it is necessary to use FEATURE(`always_add_domain')
4039N/A for the addresses to be qualified.
4039N/A The "+detail" of an address is passed as %1, so entries like
4039N/A and other forms are possible.
4039N/A If the genericstable is enabled and GENERICS_DOMAIN or
4039N/A GENERICS_DOMAIN_FILE is used, this feature will cause
4039N/A addresses to be searched in the map if their domain
4039N/A parts are subdomains of elements in class {G}.
4039N/Avirtusertable A domain-specific form of aliasing, allowing multiple
4039N/A virtual domains to be hosted on one machine. For example,
4039N/A if the virtuser table contains:
4039N/A joe@bar.com error:nouser 550 No such user here
4039N/A jax@bar.com error:5.7.0:550 Address invalid
4039N/A then mail addressed to info@foo.com will be sent to the
4039N/A address foo-info, mail addressed to info@bar.com will be
4039N/A delivered to bar-info, and mail addressed to anyone at
baz.org 4039N/A will be sent to jane@example.net, mail to joe@bar.com will
4039N/A be rejected with the specified error message, and mail to
4039N/A jax@bar.com will also have a RFC 1893 compliant error code
4039N/A The username from the original address is passed
4039N/A meaning someone@foo.org will be sent to someone@example.com.
4039N/A Additionally, if the local part consists of "user+detail"
4039N/A then "detail" is passed as %2 and "+detail" is passed as %3
4039N/A when a match against user+* is attempted, so entries like
4039N/A and other forms are possible. Note: to preserve "+detail"
4039N/A for a default case (@domain) %1%3 must be used as RHS.
4039N/A There are two wildcards after "+": "+" matches only a non-empty
4039N/A to ensure that the parameters %2 and %3 are not empty.
4039N/A latter can be defined by the macros VIRTUSER_DOMAIN or
4039N/A VIRTUSER_DOMAIN_FILE (analogously to MASQUERADE_DOMAIN and
4039N/A MASQUERADE_DOMAIN_FILE, see below). If VIRTUSER_DOMAIN or
4039N/A VIRTUSER_DOMAIN_FILE is used, then the entries of class
4039N/A {VirtHost} are added to class {R},
i.e., relaying is allowed
4039N/A to (and from) those domains, which by default includes also
4039N/A all subdomains (see relay_hosts_only). The default map
4039N/A A new definition can be specified as the second argument of
4039N/A If the virtusertable is enabled and VIRTUSER_DOMAIN or
4039N/A VIRTUSER_DOMAIN_FILE is used, this feature will cause
4039N/A addresses to be searched in the map if their domain
4039N/A parts are subdomains of elements in class {VirtHost}.
4039N/Aldap_routing Implement LDAP-based e-mail recipient routing according to
4039N/A the Internet Draft draft-lachman-laser-ldap-mail-routing-01.
4039N/A This provides a method to re-route addresses with a
4039N/A domain portion in class {LDAPRoute} to either a
4039N/A different mail host or a different address. Hosts can
4039N/A be added to this class using LDAPROUTE_DOMAIN and
4039N/A LDAPROUTE_DOMAIN_FILE (analogously to MASQUERADE_DOMAIN and
4039N/A MASQUERADE_DOMAIN_FILE, see below).
4039N/A See the LDAP ROUTING section below for more information.
4039N/Anullclient This is a special case -- it creates a configuration file
4039N/A containing nothing but support for forwarding all mail to a
4039N/A central hub via a local SMTP-based network. The argument
4039N/A The only other feature that should be used in conjunction
4039N/A with this one is FEATURE(`nocanonify'). No mailers
4039N/A should be defined. No aliasing or forwarding is done.
4039N/Alocal_lmtp Use an LMTP capable local mailer. The argument to this
4039N/A feature is the pathname of an LMTP capable mailer. By
4039N/A confEBINDIR m4 variable -- making the default
4039N/A If a different LMTP capable mailer is used, its pathname
4039N/A can be specified as second parameter and the arguments
4039N/A WARNING: This feature sets LOCAL_MAILER_FLAGS unconditionally,
4039N/A i.e., without respecting any definitions in an OSTYPE setting.
4039N/Alocal_procmail Use procmail or another delivery agent as the local mailer.
4039N/A The argument to this feature is the pathname of the
4039N/A delivery agent, which defaults to PROCMAIL_MAILER_PATH.
4039N/A Note that this does NOT use PROCMAIL_MAILER_FLAGS or
4039N/A PROCMAIL_MAILER_ARGS for the local mailer; tweak
4039N/A LOCAL_MAILER_FLAGS and LOCAL_MAILER_ARGS instead, or
4039N/A specify the appropriate parameters. When procmail is used,
4039N/A the local mailer can make use of the
4039N/A "user+indicator@local.host" syntax; normally the +indicator
4039N/A is just tossed, but by default it is passed as the -a
4039N/A This feature can take up to three arguments:
4039N/A 1. Path to the mailer program
4039N/A 2. Argument vector including name of the program
4039N/A [default: procmail -Y -a $h -d $u]
4039N/A 3. Flags for the mailer [default: SPfhn9]
4039N/A Empty arguments cause the defaults to be taken.
4039N/A Note that if you are on a system with a broken
4039N/A setreuid() call, you may need to add -f $f to the procmail
4039N/A argument vector to pass the proper sender to procmail.
4039N/A For example, this allows it to use the maildrop
4039N/A WARNING: This feature sets LOCAL_MAILER_FLAGS unconditionally,
4039N/A i.e., without respecting any definitions in an OSTYPE setting.
4039N/Abestmx_is_local Accept mail as though locally addressed for any host that
4039N/A lists us as the best possible MX record. This generates
4039N/A additional DNS traffic, but should be OK for low to
4039N/A medium traffic hosts. The argument may be a set of
4039N/A domains, which will limit the feature to only apply to
4039N/A these domains -- this will reduce unnecessary DNS
4039N/A traffic. THIS FEATURE IS FUNDAMENTALLY INCOMPATIBLE WITH
4039N/A WILDCARD MX RECORDS!!! If you have a wildcard MX record
4039N/A that matches your domain, you cannot use this feature.
4039N/Asmrsh Use the SendMail Restricted SHell (smrsh) provided
4039N/A to programs. This improves the ability of the local
4039N/A system administrator to control what gets run via
4039N/A e-mail. If an argument is provided it is used as the
4039N/A pathname to smrsh; otherwise, the path defined by
4039N/A confEBINDIR is used for the smrsh binary -- by default,
4039N/A By default, the sendmail configuration files do not permit
4039N/A mail relaying (that is, accepting mail from outside your
4039N/A local host (class {w}) and sending it to another host than
4039N/A your local host). This option sets your site to allow
4039N/A mail relaying from any site to any site. In almost all
4039N/A cases, it is better to control relaying more carefully
4039N/A with the access map, class {R}, or authentication. Domains
4039N/A can be added to class {R} by the macros RELAY_DOMAIN or
4039N/A RELAY_DOMAIN_FILE (analogously to MASQUERADE_DOMAIN and
4039N/A MASQUERADE_DOMAIN_FILE, see below).
4039N/A This option allows any host in your domain as defined by
4039N/A class {m} to use your server for relaying. Notice: make
4039N/A sure that your domain is not just a top level domain,
4039N/A e.g., com. This can happen if you give your host a name
4039N/A By default, names that are listed as RELAY in the access
4039N/A db and class {R} are treated as domain names, not host names.
4039N/A will all be accepted for relaying. This feature changes
4039N/A the behaviour to lookup individual host names only.
4039N/A Turns on the ability to allow relaying based on the MX
4039N/A records of the host portion of an incoming recipient; that
4039N/A description below for more information before using this
4039N/A feature. Also, see the KNOWNBUGS entry regarding bestmx
4039N/A FEATURE(`relay_based_on_MX') does not necessarily allow
4039N/A routing of these messages which you expect to be allowed,
4039N/A if route address syntax (or %-hack syntax) is used. If
4039N/A this is a problem, add entries to the access-table or use
4039N/A FEATURE(`loose_relay_check').
4039N/A Allows relaying if the mail sender is listed as RELAY in
4039N/A the access map. If an optional argument `domain' (this
4039N/A is the literal word `domain', not a placeholder) is given,
4039N/A relaying can be allowed just based on the domain portion
4039N/A of the sender address. This feature should only be used if
4039N/A absolutely necessary as the sender address can be easily
4039N/A forged. Use of this feature requires the "From:" tag to
4039N/A be used for the key in the access map; see the discussion
4039N/A of tags and FEATURE(`relay_mail_from') in the section on
4039N/A anti-spam configuration control.
4039N/A Allows relaying if the domain portion of the mail sender
4039N/A is a local host. This should only be used if absolutely
4039N/A necessary as it opens a window for spammers. Specifically,
4039N/A they can send mail to your mail server that claims to be
4039N/A from your domain (either directly or via a routed address),
4039N/A and you will go ahead and relay it out to arbitrary hosts
4039N/A Normally, MAIL FROM: commands in the SMTP session will be
4039N/A refused if the connection is a network connection and the
4039N/A sender address does not include a domain name. If your
4039N/A setup sends local mail unqualified (
i.e., MAIL FROM:<joe>),
4039N/A you will need to use this feature to accept unqualified
4039N/A sender addresses. Setting the DaemonPortOptions modifier
4039N/A 'u' overrides the default behavior,
i.e., unqualified
4039N/A addresses are accepted even without this FEATURE.
4039N/A If this FEATURE is not used, the DaemonPortOptions modifier
4039N/A 'f' can be used to enforce fully qualified addresses.
4039N/A Normally, MAIL FROM: commands in the SMTP session will be
4039N/A refused if the host part of the argument to MAIL FROM:
4039N/A cannot be located in the host name service (
e.g., an A or
4039N/A MX record in DNS). If you are inside a firewall that has
4039N/A only a limited view of the Internet host name space, this
4039N/A could cause problems. In this case you probably want to
4039N/A use this feature to accept all domains on input, even if
4039N/Aaccess_db Turns on the access database feature. The access db gives
4039N/A you the ability to allow or refuse to accept mail from
4039N/A specified domains for administrative reasons. Moreover,
4039N/A it can control the behavior of sendmail in various situations.
4039N/A By default, the access database specification is:
4039N/A See the anti-spam configuration control section for further
4039N/A important information about this feature. Notice:
4039N/A "-T<TMPF>" is meant literal, do not replace it by anything.
4039N/A Turns on the ability to block incoming mail for certain
4039N/A recipient usernames, hostnames, or addresses. For
4039N/A example, you can block incoming mail to user nobody,
4039N/A These specifications are put in the access db as
4039N/A described in the anti-spam configuration control section
4039N/Adelay_checks The rulesets check_mail and check_relay will not be called
4039N/A when a client connects or issues a MAIL command, respectively.
4039N/A Instead, those rulesets will be called by the check_rcpt
4039N/A ruleset; they will be skipped under certain circumstances.
4039N/A See "Delay all checks" in the anti-spam configuration control
4039N/A section. Note: this feature is incompatible to the versions
4039N/Ause_client_ptr If this feature is enabled then check_relay will override
4039N/A its first argument with $&{client_ptr}. This is useful for
4039N/A rejections based on the unverified hostname of client,
4039N/A which turns on the same behavior as in earlier sendmail
4039N/A about check_relay, {client_name}, and {client_ptr}.
4039N/Adnsbl Turns on rejection, discarding, or quarantining of hosts
4039N/A found in a DNS based list. The first argument is used as
4039N/A the domain in which blocked hosts are listed. A second
4039N/A argument can be used to change the default error message,
4039N/A or select one of the operations `discard' and `quarantine'.
4039N/A Without that second argument, the error message will be
4039N/A Rejected: IP-ADDRESS listed at SERVER
4039N/A where IP-ADDRESS and SERVER are replaced by the appropriate
4039N/A information. By default, temporary lookup failures are
4039N/A ignored. This behavior can be changed by specifying a
4039N/A third argument, which must be either `t' or a full error
4039N/A message. See the anti-spam configuration control section for
4039N/A an example. The dnsbl feature can be included several times
4039N/A to query different DNS based rejection lists. See also
4039N/A enhdnsbl for an enhanced version.
4039N/A Set the DNSBL_MAP mc option to change the default map
4039N/A definition from `host'. Set the DNSBL_MAP_OPT mc option
4039N/A to add additional options to the map specification used.
4039N/A Some DNS based rejection lists cause failures if asked
4039N/A for AAAA records. If your sendmail version is compiled
4039N/A with IPv6 support (NETINET6) and you experience this
4039N/A define(`DNSBL_MAP', `dns -R A')
4039N/A before the first use of this feature. Alternatively you
4039N/A can use enhdnsbl instead (see below). Moreover, this
4039N/A statement can be used to reduce the number of DNS retries,
4039N/A define(`DNSBL_MAP', `dns -R A -r2')
4039N/A See below (EDNSBL_TO) for an explanation.
4039N/Aenhdnsbl Enhanced version of dnsbl (see above). Further arguments
4039N/A (up to 5) can be used to specify specific return values
4039N/A from lookups. Temporary lookup failures are ignored unless
4039N/A a third argument is given, which must be either `t' or a full
4039N/A error message. By default, any successful lookup will
4039N/A generate an error. Otherwise the result of the lookup is
4039N/A compared with the supplied argument(s), and only if a match
4039N/A occurs an error is generated. For example,
4039N/A will reject the e-mail if the lookup returns the value
4039N/A ``127.0.0.2.'', or generate a 451 response if the lookup
4039N/A temporarily failed. The arguments can contain metasymbols
4039N/A as they are allowed in the LHS of rules. As the example
4039N/A shows, the default values are also used if an empty argument,
4039N/A i.e., `', is specified. This feature requires that sendmail
4039N/A Set the EDNSBL_TO mc option to change the DNS retry count
4039N/A from the default value of 5, this can be very useful when
4039N/A a DNS server is not responding, which in turn may cause
4039N/A clients to time out (an entry stating
4039N/Aratecontrol Enable simple ruleset to do connection rate control
4039N/A checking. This requires entries in access_db of the form
4039N/A The RHS specifies the maximum number of connections
4039N/A (an integer number) over the time interval defined
4039N/A by ConnectionRateWindowSize, where 0 means unlimited.
4039N/A Take the following example:
4039N/A 10.1.2.3 can only make up to 4 connections, the
4039N/A general limit it 10, and 127.0.0.1 can make an unlimited
4039N/A number of connections per ConnectionRateWindowSize.
4039N/A See also CONNECTION CONTROL.
4039N/Aconncontrol Enable a simple check of the number of incoming SMTP
4039N/A connections. This requires entries in access_db of the
4039N/A The RHS specifies the maximum number of open connections
4039N/A Take the following example:
4039N/A 10.1.2.3 can only have up to 4 open connections, the
4039N/A general limit it 10, and 127.0.0.1 does not have any
4039N/A See also CONNECTION CONTROL.
4039N/Amtamark Experimental support for "Marking Mail Transfer Agents in
4039N/A Reverse DNS with TXT RRs" (MTAMark), see
4039N/A draft-stumpf-dns-mtamark-01. Optional arguments are:
4039N/A 550 Rejected: $&{client_addr} not listed as MTA
4039N/A 2. Temporary lookup failures are ignored unless a second
4039N/A argument is given, which must be either `t' or a full
4039N/A 3. Lookup prefix, default: _perm._smtp._srv. This should
4039N/A not be changed unless the draft changes it.
4039N/A FEATURE(`mtamark', `', `t')
4039N/Alookupdotdomain Look up also .domain in the access map. This allows to
4039N/A match only subdomains. It does not work well with
4039N/A FEATURE(`relay_hosts_only'), because most lookups for
4039N/A subdomains are suppressed by the latter feature.
4039N/A Normally, if % addressing is used for a recipient,
e.g. 4039N/A user%site@othersite, and othersite is in class {R}, the
4039N/A check_rcpt ruleset will strip @othersite and recheck
4039N/A user@site for relaying. This feature changes that
4039N/A behavior. It should not be needed for most installations.
4039N/A Preserve the name of the recipient host if LUSER_RELAY is
4039N/A used. Without this option, the domain part of the
4039N/A recipient address will be replaced by the host specified as
4039N/A LUSER_RELAY. This feature only works if the hostname is
4039N/A passed to the mailer (see mailer triple in
op.me). Note
4039N/A that in the default configuration the local mailer does not
4039N/A receive the hostname,
i.e., the mailer triple has an empty
4039N/A Preserve the +detail portion of the address when passing
4039N/A address to local delivery agent. Disables alias and
4039N/A .forward +detail stripping (
e.g., given user+detail, only
4039N/A that address will be looked up in the alias file; user+* and
4039N/A user will not be looked up). Only use if the local
4039N/A delivery agent in use supports +detail addressing.
4039N/Acompat_check Enable ruleset check_compat to look up pairs of addresses
4039N/A with the Compat: tag -- Compat:sender<@>recipient -- in the
4039N/A access map. Valid values for the RHS include
4039N/A DISCARD silently discard recipient
4039N/A TEMP: return a temporary error
4039N/A ERROR: return a permanent error
4039N/A In the last two cases, a 4xy/5xy SMTP reply code should
4039N/Ano_default_msa Don't generate the default MSA daemon,
i.e.,
4039N/A DAEMON_OPTIONS(`Port=587,Name=MSA,M=E')
4039N/A To define a MSA daemon with other parameters, use this
4039N/A FEATURE and introduce new settings via DAEMON_OPTIONS().
4039N/Amsp Defines config file for Message Submission Program.
4039N/A to use it. An optional argument can be used to override
4039N/A the default of `[localhost]' to use as host to send all
4039N/A e-mails to. Note that MX records will be used if the
4039N/A specified hostname is not in square brackets (
e.g.,
4039N/A [hostname]). If `MSA' is specified as second argument then
4039N/A port 587 is used to contact the server. Example:
4039N/A Some more hints about possible changes can be found below
4039N/A in the section MESSAGE SUBMISSION PROGRAM.
4039N/A FEATURE(`msp', `[127.0.0.1]')
4039N/A by default. If you have a machine with IPv6 only,
4039N/A FEATURE(`msp', `[IPv6:::1]')
4039N/A If you want to continue using '[localhost]', (the behavior
4039N/Aqueuegroup A simple example how to select a queue group based
4039N/A on the full e-mail address or the domain of the
4039N/A recipient. Selection is done via entries in the
4039N/A access map using the tag QGRP:, for example:
4039N/A QGRP:friend@some.org others
4039N/A where "main", "others", and "local" are names of
4039N/A queue groups. If an argument is specified, it is used
4039N/A queue groups and possible queue manipulations.
4039N/Agreet_pause Adds the greet_pause ruleset which enables open proxy
4039N/A and SMTP slamming protection. The feature can take an
4039N/A argument specifying the milliseconds to wait:
4039N/A FEATURE(`greet_pause', `5000') dnl 5 seconds
4039N/A If FEATURE(`access_db') is enabled, an access database
4039N/A lookup with the GreetPause tag is done using client
4039N/A hostname, domain, IP address, or subnet to determine the
4039N/A When using FEATURE(`access_db'), the optional
FEATURE(`greet_pause') argument becomes the default if
nothing is found in the access database. A ruleset called
Local_greet_pause can be used for local modifications,
e.g.,
block_bad_helo Reject messages from SMTP clients which provide a
HELO/EHLO argument which is either unqualified, or is one of our own
names (
i.e., the server name instead of the client name).
This check is performed at RCPT stage and disabled for the
- authenticated sessions,
- connections from IP addresses in class $={R}.
Currently access_db lookups can not be used to
(selectively) disable this test, moreover,
require_rdns Reject mail from connecting SMTP clients without proper
rDNS (reverse DNS), functional gethostbyaddr() resolution.
Note: this feature will cause false positives,
i.e., there
are legitimate MTAs that do not have proper DNS entries.
Rejecting mails from those MTAs is a local policy decision.
The basic policy is to reject message with a 5xx error if
the IP address fails to resolve. However, if this is a
temporary failure, a 4xx temporary failure is returned.
If the look-up succeeds, but returns an apparently forged
value, this is treated as a temporary failure with a 4xx
Exceptions based on access entries are discussed below.
Any IP address matched using $=R (the "relay-domains" file)
is excepted from the rules. Since we have explicitly
allowed relaying for this host, based on IP address, we
The philosophical assumption here is that most users do
not control their rDNS. They should be able to send mail
through their ISP, whether or not they have valid rDNS.
The class $=R, roughly speaking, contains those IP addresses
and address ranges for which we are the ISP, or are acting
If `delay_checks' is in effect (recommended), then any
sender who has authenticated is also excepted from the
restrictions. This happens because the rules produced by
this FEATURE() will not be applied to authenticated senders
(assuming `delay_checks').
will whitelist IP address 1.2.3.4, so that the rDNS
blocking does apply to that IP address
will have the effect of forcing a temporary failure for
that address to be treated as a permanent failure.
badmx Reject envelope sender addresses (MAIL) whose domain part
resolves to a "bad" MX record. By default these are
MX records which resolve to A records that match the
This default regular expression can be overridden by
specifying an argument,
e.g.,
FEATURE(`badmx', `^127\.0\.0\.1')
Note: this feature requires that the sendmail binary
has been compiled with the options MAP_REGEX and
It's hard to get UUCP mailers right because of the extremely ad hoc
nature of UUCP addressing. These config files are really designed
for domain-based addressing, even for UUCP sites.
There are four UUCP mailers available. The choice of which one to
use is partly a matter of local preferences and what is running at
the other end of your UUCP connection. Unlike good protocols that
define what will go over the wire, UUCP uses the policy that you
should do what is right for the other end; if they change, you have
to change. This makes it hard to do the right thing, and discourages
people from updating their software. In general, if you can avoid
The major choice is whether to go for a domainized scheme or a
non-domainized scheme. This depends entirely on what the other
end will recognize. If at all possible, you should encourage the
other end to go to a domain-based system -- non-domainized addresses
don't work entirely properly.
uucp-old (obsolete name: "uucp")
This is the oldest, the worst (but the closest to UUCP) way of
sending messages across UUCP connections. It does bangify
everything and prepends $U (your UUCP name) to the sender's
address (which can already be a bang path itself). It can
only send to one address at a time, so it spends a lot of
time copying duplicates of messages. Avoid this if at all
uucp-new (obsolete name: "suucp")
The same as above, except that it assumes that in one rmail
command you can specify several recipients. It still has a
This UUCP mailer keeps everything as domain addresses.
Basically, it uses the SMTP mailer rewriting rules. This mailer
is only included if MAILER(`smtp') is specified before
Unfortunately, a lot of UUCP mailer transport agents require
bangified addresses in the envelope, although you can use
domain-based addresses in the message header. (The envelope
shows up as the From_ line on UNIX mail.) So....
This is a cross between uucp-new (for the envelope addresses)
and uucp-dom (for the header addresses). It bangifies the
envelope sender (From_ line in messages) without adding the
local hostname, unless there is no host name on the address
at all (
e.g., "wolf") or the host component is a UUCP host name
instead of a domain name ("somehost!wolf" instead of
"
some.dom.ain!wolf"). This is also included only if MAILER(`smtp')
is also specified earlier.
summarizes the sender rewriting for various mailers.
Mailer sender rewriting in the envelope
------ ------ -------------------------
uucp-{old,new} wolf grasp!wolf
uucp-dom wolf wolf@grasp.insa-lyon.fr
uucp-{old,new} wolf@fr.net grasp!
fr.net!wolf
uucp-dom wolf@fr.net wolf@fr.net
uucp-uudom wolf@fr.net
fr.net!wolf
uucp-{old,new} somehost!wolf grasp!somehost!wolf
uucp-dom somehost!wolf somehost!wolf@grasp.insa-lyon.fr
If you are using one of the domainized UUCP mailers, you really want
to convert all UUCP addresses to domain format -- otherwise, it will
do it for you (and probably not the way you expected). For example,
if you have the address foo!bar!baz (and you are not sending to foo),
will not add the local hostname. You can do this using the uucpdomain
For more complex configurations, you can define special rules.
The macro LOCAL_RULE_3 introduces rules that are used in canonicalizing
the names. Any modifications made here are reflected in the header.
A common use is to convert old UUCP addresses to SMTP addresses using
the UUCPSMTP macro. For example:
will cause addresses of the form "decvax!user" and "research!user"
to be converted to "user@decvax.dec.com" and "user@research.att.com"
This could also be used to look up hosts in a database map:
R$* < @ $+ > $* $: $1 < @ $(hostmap $2 $) > $3
This map would be defined in the LOCAL_CONFIG portion, as shown below.
Similarly, LOCAL_RULE_0 can be used to introduce new parsing rules.
For example, new rules are needed to parse hostnames that you accept
via MX records. For example, you might have:
pointing at this host; this rule catches the message and forwards it on
You can also tweak rulesets 1 and 2 using LOCAL_RULE_1 and LOCAL_RULE_2.
These rulesets are normally empty.
A similar macro is LOCAL_CONFIG. This introduces lines added after the
boilerplate option setting but before rulesets. Do not declare rulesets in
the LOCAL_CONFIG section. It can be used to declare local database maps or
+---------------------------+
| MASQUERADING AND RELAYING |
+---------------------------+
You can have your host masquerade as another using
This causes mail being sent to be labeled as coming from the
indicated
host.domain, rather than $j. One normally masquerades as
one of one's own subdomains (for example, it's unlikely that
Berkeley would choose to masquerade as an MIT site). This
behaviour is modified by a plethora of FEATUREs; in particular, see
masquerade_envelope, allmasquerade, limited_masquerade, and
masquerade_entire_domain.
The masquerade name is not normally canonified, so it is important
that it be your One True Name, that is, fully qualified and not a
CNAME. However, if you use a CNAME, the receiving side may canonify
it for you, so don't think you can cheat CNAME mapping this way.
Normally the only addresses that are masqueraded are those that come
from this host (that is, are either unqualified or in class {w}, the list
of local domain names). You can augment this list, which is realized
The effect of this is that although mail to user@otherhost.domain
will not be delivered locally, any mail including any user@otherhost.domain
will, when relayed, be rewritten to have the MASQUERADE_AS address.
This can be a space-separated list of names.
If these names are in a file, you can use
MASQUERADE_DOMAIN_FILE(`filename')
to read the list of names from the indicated file (
i.e., to add
To exempt hosts or subdomains from being masqueraded, you can use
This can come handy if you want to masquerade a whole domain
except for one (or a few) host(s). If these names are in a file,
MASQUERADE_EXCEPTION_FILE(`filename')
Normally only header addresses are masqueraded. If you want to
masquerade the envelope as well, use
FEATURE(`masquerade_envelope')
There are always users that need to be "exposed" -- that is, their
internal site name should be displayed instead of the masquerade name.
Root is an example (which has been "exposed" by default prior to 8.10).
You can add users to this list using
EXPOSED_USER(`usernames')
This adds users to class {E}; you could also use
EXPOSED_USER_FILE(`filename')
You can also arrange to relay all unqualified names (that is, names
without @host) to a relay host. For example, if you have a central
email server, you might relay to that host so that users don't have
to have .forward files or aliases. You can do this using
define(`LOCAL_RELAY', `mailer:hostname')
The ``mailer:'' can be omitted, in which case the mailer defaults to
"relay". There are some user names that you don't want relayed, perhaps
because of local aliases. A common example is root, which may be
locally aliased. You can add entries to this list using
This adds users to class {L}; you could also use
LOCAL_USER_FILE(`filename')
If you want all incoming mail sent to a centralized hub, as for a
define(`MAIL_HUB', `mailer:hostname')
Again, ``mailer:'' defaults to "relay". If you define both LOCAL_RELAY
and MAIL_HUB _AND_ you have FEATURE(`stickyhost'), unqualified names will
be sent to the LOCAL_RELAY and other local names will be sent to MAIL_HUB.
Note: there is a (long standing) bug which keeps this combination from
working for addresses of the form user+detail.
Names in class {L} will be delivered locally, so you MUST have aliases or
FEATURE(`stickyhost'), the following combinations of settings will have the
email sent to.... eric eric@mastodon.CS.Berkeley.EDU
MAIL_HUB set as above (no local aliasing) (aliasing done)
If you do not have FEATURE(`stickyhost') set, then LOCAL_RELAY and
MAIL_HUB act identically, with MAIL_HUB taking precedence.
If you want all outgoing mail to go to a central relay site, define
SMART_HOST as well. Briefly:
LOCAL_RELAY applies to unqualified names (
e.g., "eric").
MAIL_HUB applies to names qualified with the name of the
local host (
e.g., "eric@mastodon.CS.Berkeley.EDU").
SMART_HOST applies to names qualified with other hosts or
bracketed addresses (
e.g., "eric@mastodon.CS.Berkeley.EDU"
However, beware that other relays (
e.g., UUCP_RELAY, BITNET_RELAY,
DECNET_RELAY, and FAX_RELAY) take precedence over SMART_HOST, so if you
really want absolutely everything to go to a single central site you will
need to unset all the other relays -- or better yet, find or build a
minimal config file that does this.
For duplicate suppression to work properly, the host name is best
specified with a terminal dot:
note the trailing dot ---^
+-------------------------------------------+
| USING LDAP FOR ALIASES, MAPS, AND CLASSES |
+-------------------------------------------+
LDAP can be used for aliases, maps, and classes by either specifying your
own LDAP map specification or using the built-in default LDAP map
specification. The built-in default specifications all provide lookups
which match against either the machine's fully qualified hostname (${j}) or
a "cluster". The cluster allows you to share LDAP entries among a large
number of machines without having to enter each of the machine names into
each LDAP entry. To set the LDAP cluster name to use for a particular
machine or set of machines, set the confLDAP_CLUSTER m4 variable to a
unique name. For example:
define(`confLDAP_CLUSTER', `Servers')
Here, the word `Servers' will be the cluster name. As an example, assume
Some of the LDAP LDIF examples below show use of the Servers cluster.
Every entry must have either a sendmailMTAHost or sendmailMTACluster
attribute or it will be ignored. Be careful as mixing clusters and
individual host records can have surprising results (see the CAUTION
this schema (and therefore the lookups and examples below) is experimental
at this point as it has had little public review. Therefore, it may change
in future versions. Feedback via sendmail-YYYY@support.sendmail.org is
encouraged (replace YYYY with the current year,
e.g., 2005).
The ALIAS_FILE (O AliasFile) option can be set to use LDAP for alias
lookups. To use the default schema, simply use:
define(`ALIAS_FILE', `ldap:')
By doing so, you will use the default schema which expands to a map
ldap -k (&(objectClass=sendmailMTAAliasObject)
(sendmailMTAAliasGrouping=aliases)
(|(sendmailMTACluster=${sendmailMTACluster})
-v sendmailMTAAliasValue,sendmailMTAAliasSearch:FILTER:sendmailMTAAliasObject,sendmailMTAAliasURL:URL:sendmailMTAAliasObject
NOTE: The macros shown above ${sendmailMTACluster} and $j are not actually
used when the binary expands the `ldap:' token as the AliasFile option is
not actually macro-expanded when read from the
sendmail.cf file.
Example LDAP LDIF entries might be:
dn: sendmailMTAKey=sendmail-list, dc=sendmail, dc=org
objectClass: sendmailMTAAlias
objectClass: sendmailMTAAliasObject
sendmailMTAAliasGrouping: aliases
sendmailMTAKey: sendmail-list
sendmailMTAAliasValue: ca@example.org
sendmailMTAAliasValue: eric
sendmailMTAAliasValue: gshapiro@example.com
dn: sendmailMTAKey=owner-sendmail-list, dc=sendmail, dc=org
objectClass: sendmailMTAAlias
objectClass: sendmailMTAAliasObject
sendmailMTAAliasGrouping: aliases
sendmailMTAKey: owner-sendmail-list
sendmailMTAAliasValue: eric
dn: sendmailMTAKey=postmaster, dc=sendmail, dc=org
objectClass: sendmailMTAAlias
objectClass: sendmailMTAAliasObject
sendmailMTAAliasGrouping: aliases
sendmailMTACluster: Servers
sendmailMTAKey: postmaster
sendmailMTAAliasValue: eric
Here, the aliases sendmail-list and owner-sendmail-list will be available
CAUTION: aliases are additive so that entries like these:
dn: sendmailMTAKey=bob, dc=sendmail, dc=org
objectClass: sendmailMTAAlias
objectClass: sendmailMTAAliasObject
sendmailMTAAliasGrouping: aliases
sendmailMTACluster: Servers
sendmailMTAAliasValue: eric
dn: sendmailMTAKey=bobetrn, dc=sendmail, dc=org
objectClass: sendmailMTAAlias
objectClass: sendmailMTAAliasObject
sendmailMTAAliasGrouping: aliases
sendmailMTAAliasValue: gshapiro
would mean that on all of the hosts in the cluster, mail to bob would go to
If you prefer not to use the default LDAP schema for your aliases, you can
specify the map parameters when setting ALIAS_FILE. For example:
define(`ALIAS_FILE', `ldap:-k (&(objectClass=mailGroup)(mail=%0)) -v mgrpRFC822MailMember')
FEATURE()'s which take an optional map definition argument (
e.g., access,
mailertable, virtusertable, etc.) can instead take the special keyword
FEATURE(`access_db', `LDAP')
FEATURE(`virtusertable', `LDAP')
When this keyword is given, that map will use LDAP lookups consisting of
the objectClass sendmailMTAClassObject, the attribute sendmailMTAMapName
with the map name, a search attribute of sendmailMTAKey, and the value
attribute sendmailMTAMapValue.
The values for sendmailMTAMapName are:
FEATURE() sendmailMTAMapName
--------- ------------------
For example, FEATURE(`mailertable', `LDAP') would use the map definition:
Kmailertable ldap -k (&(objectClass=sendmailMTAMapObject)
(sendmailMTAMapName=mailer)
(|(sendmailMTACluster=${sendmailMTACluster})
-1 -v sendmailMTAMapValue,sendmailMTAMapSearch:FILTER:sendmailMTAMapObject,sendmailMTAMapURL:URL:sendmailMTAMapObject
An example LDAP LDIF entry using this map might be:
dn: sendmailMTAMapName=mailer, dc=sendmail, dc=org
objectClass: sendmailMTAMap
sendmailMTACluster: Servers
sendmailMTAMapName: mailer
dn: sendmailMTAKey=
example.com, sendmailMTAMapName=mailer, dc=sendmail, dc=org
objectClass: sendmailMTAMap
objectClass: sendmailMTAMapObject
sendmailMTAMapName: mailer
sendmailMTACluster: Servers
CAUTION: If your LDAP database contains the record above and *ALSO* a host
dn: sendmailMTAKey=
example.com@etrn, sendmailMTAMapName=mailer, dc=sendmail, dc=org
objectClass: sendmailMTAMap
objectClass: sendmailMTAMapObject
sendmailMTAMapName: mailer
then these entries will give unexpected results. When the lookup is done
Servers cluster, LDAP would return two answers for the
example.com map key
in which case sendmail would treat this as no match at all.
If you prefer not to use the default LDAP schema for your maps, you can
specify the map parameters when using the FEATURE(). For example:
FEATURE(`access_db', `ldap:-1 -k (&(objectClass=mapDatabase)(key=%0)) -v value')
Normally, classes can be filled via files or programs. As of 8.12, they
can also be filled via map lookups using a new syntax:
F{ClassName}mapkey@mapclass:mapspec
mapkey is optional and if not provided the map key will be empty. This can
be used with LDAP to read classes from LDAP. Note that the lookup is only
done when sendmail is initially started. Use the special value `@LDAP' to
use the default LDAP schema. For example:
RELAY_DOMAIN_FILE(`@LDAP')
would put all of the attribute sendmailMTAClassValue values of LDAP records
with objectClass sendmailMTAClass and an attribute sendmailMTAClassName of
'R' into class $={R}. In other words, it is equivalent to the LDAP map
F{R}@ldap:-k (&(objectClass=sendmailMTAClass)
(|(sendmailMTACluster=${sendmailMTACluster})
-v sendmailMTAClassValue,sendmailMTAClassSearch:FILTER:sendmailMTAClass,sendmailMTAClassURL:URL:sendmailMTAClass
NOTE: The macros shown above ${sendmailMTACluster} and $j are not actually
used when the binary expands the `@LDAP' token as class declarations are
not actually macro-expanded when read from the
sendmail.cf file.
This can be used with class related commands such as RELAY_DOMAIN_FILE(),
MASQUERADE_DOMAIN_FILE(), etc:
Command sendmailMTAClassName
------- --------------------
CANONIFY_DOMAIN_FILE() Canonify
LDAPROUTE_DOMAIN_FILE() LDAPRoute
LDAPROUTE_EQUIVALENT_FILE() LDAPRouteEquiv
MASQUERADE_DOMAIN_FILE() M
MASQUERADE_EXCEPTION_FILE() N
VIRTUSER_DOMAIN_FILE() VirtHost
You can also add your own as any 'F'ile class of the form:
will use "ClassName" for the sendmailMTAClassName.
An example LDAP LDIF entry would look like:
dn: sendmailMTAClassName=R, dc=sendmail, dc=org
objectClass: sendmailMTAClass
sendmailMTACluster: Servers
sendmailMTAClassValue: 10.56.23
CAUTION: If your LDAP database contains the record above and *ALSO* a host
dn: sendmailMTAClassName=R@etrn.sendmail.org, dc=sendmail, dc=org
objectClass: sendmailMTAClass
the result will be similar to the aliases caution above. When the lookup
both the cluster match and the host match). In other words, the effective
If you prefer not to use the default LDAP schema for your classes, you can
specify the map parameters when using the class command. For example:
VIRTUSER_DOMAIN_FILE(`@ldap:-k (&(objectClass=virtHosts)(host=*)) -v host')
Remember, macros can not be used in a class declaration as the binary does
FEATURE(`ldap_routing') can be used to implement the IETF Internet Draft
LDAP Schema for Intranet Mail Routing
(draft-lachman-laser-ldap-mail-routing-01). This feature enables
LDAP-based rerouting of a particular address to either a different host
or a different address. The LDAP lookup is first attempted on the full
address (
e.g., user@example.com) and then on the domain portion
LDAPROUTE_DOMAIN(),
e.g.:
Additionally, you can specify equivalent domains for LDAP routing using
LDAPROUTE_EQUIVALENT() and LDAPROUTE_EQUIVALENT_FILE(). 'Equivalent'
hostnames are mapped to $M (the masqueraded hostname for the server) before
the LDAP query. For example, if the mail is addressed to
user@host1.example.com, normally the LDAP lookup would only be done for
done on 'user@example.com' and '@
example.com' after attempting the
By default, the feature will use the schemas as specified in the draft
and will not reject addresses not found by the LDAP lookup. However,
this behavior can be changed by giving additional arguments to the FEATURE()
FEATURE(`ldap_routing', <mailHost>, <mailRoutingAddress>, <bounce>,
<detail>, <nodomain>, <tempfail>)
where <mailHost> is a map definition describing how to lookup an alternative
mail host for a particular address; <mailRoutingAddress> is a map definition
describing how to lookup an alternative address for a particular address;
the <bounce> argument, if present and not the word "passthru", dictates
that mail should be bounced if neither a mailHost nor mailRoutingAddress
is found, if set to "sendertoo", the sender will be rejected if not
found in LDAP; and <detail> indicates what actions to take if the address
contains +detail information -- `strip' tries the lookup with the +detail
and if no matches are found, strips the +detail and tries the lookup again;
`preserve', does the same as `strip' but if a mailRoutingAddress match is
found, the +detail information is copied to the new address; the <nodomain>
argument, if present, will prevent the @domain lookup if the full
address is not found in LDAP; the <tempfail> argument, if set to
"tempfail", instructs the rules to give an SMTP 4XX temporary
error if the LDAP server gives the MTA a temporary failure, or if set to
"queue" (the default), the MTA will locally queue the mail.
The default <mailHost> map definition is:
ldap -1 -T<TMPF> -v mailHost -k (&(objectClass=inetLocalMailRecipient)
The default <mailRoutingAddress> map definition is:
ldap -1 -T<TMPF> -v mailRoutingAddress
-k (&(objectClass=inetLocalMailRecipient)
Note that neither includes the LDAP server hostname (-h server) or base DN
(-b o=org,c=COUNTRY), both necessary for LDAP queries. It is presumed that
your .mc file contains a setting for the confLDAP_DEFAULT_SPEC option with
these settings. If this is not the case, the map definitions should be
changed as described above. The "-T<TMPF>" is required in any user
specified map definition to catch temporary errors.
The following possibilities exist as a result of an LDAP lookup on an
mailHost is mailRoutingAddress is Results in
----------- --------------------- ----------
set to a set mail delivered to
"local" host mailRoutingAddress
set to a not set delivered to
"local" host original address
set to a set mailRoutingAddress
remote host relayed to mailHost
set to a not set original address
remote host relayed to mailHost
not set set mail delivered to
not set not set delivered to
The term "local" host above means the host specified is in class {w}. If
the result would mean sending the mail to a different host, that host is
looked up in the mailertable before delivery.
Note that the last case depends on whether the third argument is given
to the FEATURE() command. The default is to deliver the message to the
The LDAP entries should be set up with an objectClass of
inetLocalMailRecipient and the address be listed in a mailLocalAddress
attribute. If present, there must be only one mailHost attribute and it
must contain a fully qualified host name as its value. Similarly, if
present, there must be only one mailRoutingAddress attribute and it must
contain an RFC 822 compliant address. Some example LDAP records (in LDIF
objectClass: inetLocalMailRecipient
mailLocalAddress: tom@example.com
mailRoutingAddress: thomas@mailhost.example.com
This would deliver mail for tom@example.com to thomas@mailhost.example.com.
objectClass: inetLocalMailRecipient
mailLocalAddress: dick@example.com
This would relay mail for dick@example.com to the same address but redirect
objectClass: inetLocalMailRecipient
mailLocalAddress: harry@example.com
mailRoutingAddress: harry@mkt.example.com
This would relay mail for harry@example.com to the MX records listed for
when talking to that host.
objectClass: inetLocalMailRecipient
mailRoutingAddress: virtual@example.com
virtual@example.com on that relay machine.
+---------------------------------+
| ANTI-SPAM CONFIGURATION CONTROL |
+---------------------------------+
The primary anti-spam features available in sendmail are:
* Relaying is denied by default.
* Better checking on sender information.
Relaying (transmission of messages from a site outside your host (class
{w}) to another site except yours) is denied by default. Note that this
changed in sendmail 8.9; previous versions allowed relaying by default.
If you really want to revert to the old behaviour, you will need to use
FEATURE(`promiscuous_relay'). You can allow certain domains to relay
through your server by adding their domain name or IP address to class
{R} using RELAY_DOMAIN() and RELAY_DOMAIN_FILE() or via the access database
(described below). Note that IPv6 addresses must be prefaced with "IPv6:".
The file consists (like any other file based class) of entries listed on
IPv6:2002:c0a8:51d2::23f4
Notice: the last entry allows relaying for connections via a UNIX
socket to the
MTA/MSP. This might be necessary if your configuration
doesn't allow relaying by other means in that case,
e.g., by having
localhost.$m in class {R} (make sure $m is not just a top level
FEATURE(`relay_entire_domain')
then any host in any of your local domains (that is, class {m})
will be relayed (that is, you will accept mail either to or from any
You can also allow relaying based on the MX records of the host
portion of an incoming recipient address by using
FEATURE(`relay_based_on_MX')
For example, if your server receives a recipient of user@domain.com
and
domain.com lists your server in its MX records, the mail will be
accepted for relay to
domain.com. This feature may cause problems
if MX lookups for the recipient domain are slow or time out. In that
case, mail will be temporarily rejected. It is usually better to
maintain a list of
hosts/domains for which the server acts as relay.
Note also that this feature will stop spammers from using your host
to relay spam but it will not stop outsiders from using your server
as a relay for their site (that is, they set up an MX record pointing
to your mail server, and you will relay mail addressed to them
without any prior arrangement). Along the same lines,
FEATURE(`relay_local_from')
will allow relaying if the sender specifies a return path (
i.e.MAIL FROM:<user@domain>) domain which is a local domain. This is a
dangerous feature as it will allow spammers to spam using your mail
server by simply specifying a return address of user@your.domain.com.
It should not be used unless absolutely necessary.
A slightly better solution is
FEATURE(`relay_mail_from')
which allows relaying if the mail sender is listed as RELAY in the
access map. If an optional argument `domain' (this is the literal
word `domain', not a placeholder) is given, the domain portion of
the mail sender is also checked to allowing relaying. This option
only works together with the tag From: for the LHS of the access
map entries. This feature allows spammers to abuse your mail server
by specifying a return address that you enabled in your access file.
This may be harder to figure out for spammers, but it should not
be used unless necessary. Instead use STARTTLS to
allow relaying for roaming users.
If source routing is used in the recipient address (
e.g.,
RCPT TO:<user%site.com@othersite.com>), sendmail will check
user@site.com for relaying if
othersite.com is an allowed relay host
in either class {R}, class {m} if FEATURE(`relay_entire_domain') is used,
or the access database if FEATURE(`access_db') is used. To prevent
the address from being stripped down, use:
FEATURE(`loose_relay_check')
If you think you need to use this feature, you probably do not. This
should only be used for sites which have no control over the addresses
that they provide a gateway for. Use this FEATURE with caution as it
can allow spammers to relay through your server if not setup properly.
NOTICE: It is possible to relay mail through a system which the anti-relay
rules do not prevent: the case of a system that does use FEATURE(`nouucp',
`nospecial') (system A) and relays local messages to a mail hub (
e.g., via
LOCAL_RELAY or LUSER_RELAY) (system B). If system B doesn't use
FEATURE(`nouucp') at all, addresses of the form
<
example.net!user@local.host> would be relayed to <user@example.net>.
System A doesn't recognize `!' as an address separator and therefore
forwards it to the mail hub which in turns relays it because it came from
a trusted local host. So if a mailserver allows UUCP (bang-format)
addresses, all systems from which it allows relaying should do the same
or reject those addresses.
As of 8.9, sendmail will refuse mail if the MAIL FROM: parameter has
an unresolvable domain (
i.e., one that DNS, your local name service,
or special case rules in ruleset 3 cannot locate). This also applies
to addresses that use domain literals,
e.g., <user@[1.2.3.4]>, if the
IP address can't be mapped to a host name. If you want to continue
to accept such domains,
e.g., because you are inside a firewall that
has only a limited view of the Internet host name space (note that you
will not be able to return mail to them unless you have some "smart
FEATURE(`accept_unresolvable_domains')
Alternatively, you can allow specific addresses by adding them to
Notice: domains which are temporarily unresolvable are (temporarily)
rejected with a 451 reply code. If those domains should be accepted
(which is discouraged) then you can use
sendmail will also refuse mail if the MAIL FROM: parameter is not
fully qualified (
i.e., contains a domain as well as a user). If you
want to continue to accept such senders, use
FEATURE(`accept_unqualified_senders')
Setting the DaemonPortOptions modifier 'u' overrides the default behavior,
i.e., unqualified addresses are accepted even without this FEATURE. If
this FEATURE is not used, the DaemonPortOptions modifier 'f' can be used
to enforce fully qualified domain names.
An ``access'' database can be created to accept or reject mail from
selected domains. For example, you may choose to reject all mail
originating from known spammers. To enable such a database, use
Notice: the access database is applied to the envelope addresses
and the connection information, not to the header.
The FEATURE macro can accept as second parameter the key file
definition for the database; for example
Notice: If a second argument is specified it must contain the option
`-T<TMPF>' as shown above. The optional parameters may be
`skip' enables SKIP as value part (see below).
`lookupdotdomain' another way to enable the feature of the
`relaytofulladdress' enable entries of the form
To:user@example.com RELAY
to allow relaying to just a specific
e-mail address instead of an entire domain.
file as described below, you must use makemap to create the database
The table itself uses e-mail addresses, domain names, and network
numbers as keys. Note that IPv6 addresses must be prefaced with "IPv6:".
From:spammer@aol.com REJECT
Connect:192.168.212 REJECT
Connect:IPv6:2002:c0a8:02c7 RELAY
Connect:IPv6:2002:c0a8:51d2::23f4 REJECT
top level domain TLD, 192.168.212.* network, and the IPv6 address
2002:c0a8:51d2::23f4. It would allow relay for the IPv6 network
Entries in the access map should be tagged according to their type.
Three tags are available:
Connect: connection information (${client_addr}, ${client_name})
Notice: untagged entries are deprecated.
If the required item is looked up in a map, it will be tried first
with the corresponding tag in front, then (as fallback to enable
backward compatibility) without any tag, unless the specific feature
requires a tag. For example,
From:spammer@some.dom REJECT
This would deny mails from spammer@some.dom but you could still
send mail to that address even if FEATURE(`blacklist_recipients')
is enabled. Your system will allow relaying to
friend.domain, but
not from it (unless enabled by other means). Connections from that
domain will be allowed even if it ends up in one of the DNS based
rejection lists. Relaying is enabled from
from.domain but not to
it (since relaying is based on the connection information for
outgoing relaying, the tag Connect: must be used; for incoming
relaying, which is based on the recipient address, To: must be
used). The last two entries allow mails from good@another.dom but
reject mail from all other addresses with
another.dom as domain
The value part of the map can contain:
OK Accept mail even if other rules in the running
ruleset would reject it, for example, if the domain
name is unresolvable. "Accept" does not mean
"relay", but at most acceptance for local
recipients. That is, OK allows less than RELAY.
RELAY Accept mail addressed to the indicated domain
(or address if `relaytofulladdress' is set) or
received from the indicated domain for relaying
through your SMTP server. RELAY also serves as
an implicit OK for the other checks.
REJECT Reject the sender or recipient with a general
DISCARD Discard the message completely using the
$#discard mailer. If it is used in check_compat,
it affects only the designated recipient, not
the whole message as it does in all other cases.
This should only be used if really necessary.
search for this entry without accepting or rejecting
it but causing the default action.
### any text where ### is an RFC 821 compliant error code and
"any text" is a message to return for the command.
The entire string should be quoted to avoid
Otherwise sendmail formats the text as email
addresses,
e.g., it may remove spaces.
This type is deprecated, use one of the two
ERROR: entries below instead.
as above, but useful to mark error messages as such.
If quotes need to be used to avoid modifications
(see above), they should be placed like this:
where
D.S.N is an RFC 1893 compliant error code
and the rest as above. If quotes need to be used
to avoid modifications, they should be placed
ERROR:
D.S.N:"### any text"
Quarantine the message using the given text as the
Connect:IPv6:1:2:3:4:5:6:7 RELAY
Connect:[IPv6:1:2:3:4:5:6:7:8] OK
It would allow relaying mail from and to any hosts in the
sendmail.orgdomain, and allow relaying from the IPv6 1:2:3:4:5:6:7:* network
and from the 128.32.*.* network except for the 128.32.2.* network,
last two entries are for checks against ${client_name} if the IP
address doesn't resolve to a hostname (or is considered as "may be
forged"). That is, using square brackets means these are host
names, not network numbers.
Warning: if you change the RFC 821 compliant error code from the default
value of 550, then you should probably also change the RFC 1893 compliant
error code to match it. For example, if you use
To:user@example.com ERROR:450 mailbox full
the error returned would be "450 5.0.0 mailbox full" which is wrong.
Use "ERROR:4.2.2:450 mailbox full" instead.
Note, UUCP users may need to add
hostname.UUCP to the access database
FEATURE(`relay_hosts_only')
then the above example will allow relaying for
sendmail.org, but not
hosts within the
sendmail.org domain. Note that this will also require
hosts listed in class {R} to be fully qualified host names.
You can also use the access database to block sender addresses based on
the username portion of the address. For example:
Note that you must include the @ after the username to signify that
this database entry is for checking only the username portion of the
FEATURE(`blacklist_recipients')
then you can add entries to the map for local users, hosts in your
domains, or addresses in your domain which should not receive mail:
To:badlocaluser@ ERROR:550 Mailbox disabled for badlocaluser
To:
host.my.TLD ERROR:550 That host does not accept mail
To:user@other.my.TLD ERROR:550 Mailbox disabled for this recipient
This would prevent a recipient of badlocaluser in any of the local
domains (class {w}), any user at
host.my.TLD, and the single address
user@other.my.TLD from receiving mail. Please note: a local username
must be now tagged with an @ (this is consistent with the check of
the sender address, and hence it is possible to distinguish between
hostnames and usernames). Enabling this feature will keep you from
sending mails to all addresses that have an error message or REJECT
as value part in the access map. Taking the example from above:
That's why tagged entries should be used.
There are several DNS based blacklists which can be found by
querying a search engine. These are databases of spammers
maintained in DNS. To use such a database, specify
This will cause sendmail to reject mail from any site listed in the
DNS based blacklist. You must select a DNS based blacklist domain
to check by specifying an argument to the FEATURE. The default
Rejected: IP-ADDRESS listed at SERVER
where IP-ADDRESS and SERVER are replaced by the appropriate
information. A second argument can be used to specify a different
text or action. For example,
would quarantine the message if the client IP address is listed
By default, temporary lookup failures are ignored
and hence cause the connection not to be rejected by the DNS based
rejection list. This behavior can be changed by specifying a third
argument, which must be either `t' or a full error message. For
If `t' is used, the error message is:
451 Temporary lookup failure of IP-ADDRESS at SERVER
where IP-ADDRESS and SERVER are replaced by the appropriate
This FEATURE can be included several times to query different
DNS based rejection lists.
Notice: to avoid checking your own local domains against those
blacklists, use the access_db feature and add:
to the access map, where 10.1 is your local network. You may
want to use "RELAY" instead of "OK" to allow also relaying
instead of just disabling the DNS lookups in the blacklists.
The features described above make use of the check_relay, check_mail,
and check_rcpt rulesets. Note that check_relay checks the SMTP
client hostname and IP address when the connection is made to your
server. It does not check if a mail message is being relayed to
another server. That check is done in check_rcpt. If you wish to
include your own checks, you can put your checks in the rulesets
Local_check_relay, Local_check_mail, and Local_check_rcpt. For
example if you wanted to block senders with all numeric usernames
(
i.e. 2312343@bigisp.com), you would use Local_check_mail and the
Kallnumbers regex -a@MATCH ^[0-9]+$
# check address against various regex checks
R@MATCH $#error $: 553 Header Error
These rules are called with the original arguments of the corresponding
check_* ruleset. If the local ruleset returns $#OK, no further checking
is done by the features described above and the mail is accepted. If
the local ruleset resolves to a mailer (such as $#error or $#discard),
the appropriate action is taken. Other results starting with $# are
interpreted by sendmail and may lead to unspecified behavior. Note: do
NOT create a mailer with the name OK. Return values that do not start
with $# are ignored,
i.e., normal processing continues.
By using FEATURE(`delay_checks') the rulesets check_mail and check_relay
will not be called when a client connects or issues a MAIL command,
respectively. Instead, those rulesets will be called by the check_rcpt
ruleset; they will be skipped if a sender has been authenticated using
a "trusted" mechanism,
i.e., one that is defined via TRUST_AUTH_MECH().
If check_mail returns an error then the RCPT TO command will be rejected
with that error. If it returns some other result starting with $# then
check_relay will be skipped. If the sender address (or a part of it) is
listed in the access map and it has a RHS of OK or RELAY, then check_relay
will be skipped. This has an interesting side effect: if your domain is
in the access map, then any e-mail with a sender address of
<user@my.domain> will not be rejected by check_relay even though
it would match the hostname or IP address. This allows spammers
to get around DNS based blacklist by faking the sender address. To
avoid this problem you have to use tagged entries:
if you need those entries at all (class {R} may take care of them).
FEATURE(`delay_checks') can take an optional argument:
FEATURE(`delay_checks', `friend')
FEATURE(`delay_checks', `hater')
If such an argument is given, the recipient will be looked up in the
access map (using the tag Spam:). If the argument is `friend', then
the default behavior is to apply the other rulesets and make a SPAM
friend the exception. The rulesets check_mail and check_relay will be
skipped only if the recipient address is found and has RHS FRIEND. If
the argument is `hater', then the default behavior is to skip the rulesets
check_mail and check_relay and make a SPAM hater the exception. The
other two rulesets will be applied only if the recipient address is
This allows for simple exceptions from the tests,
e.g., by activating
the friend option and having
in the access map, mail to abuse@localdomain will get through (where
"localdomain" is any domain in class {w}). It is also possible to
specify a full address or an address with +detail:
Spam:abuse@my.domain FRIEND
Note: The required tag has been changed in 8.12 from To: to Spam:.
This change is incompatible to previous versions. However, you can
(for now) simply add the new entries to the access map, the old
ones will be ignored. As soon as you removed the old entries from
the access map, specify a third parameter (`n') to this feature and
the backward compatibility rules will not be in the generated .cf
You can also reject mail on the basis of the contents of headers.
This is done by adding a ruleset call to the 'H' header definition command
in
sendmail.cf. For example, this can be used to check the validity of
HMessage-Id: $>CheckMessageId
R$* $#error $: 553 Header Error
HSubject: $>+CheckSubject
that is, $>+ instead of $>, gives the full Subject: header including
comments to the ruleset (comments in parentheses () are stripped
A default ruleset for headers which don't have a specific ruleset
defined for them can be given by:
1. All rules act on tokens as explained in
doc/op/op.{me,ps,txt}.
That may cause problems with simple header checks due to the
tokenization. It might be simpler to use a regex map and apply it
2. There are no default rulesets coming with this distribution of
sendmail. You can write your own or search the WWW for examples.
3. When using a default ruleset for headers, the name of the header
currently being checked can be found in the $&{hdr_name} macro.
After all of the headers are read, the check_eoh ruleset will be called for
any final header-related checks. The ruleset is called with the number of
headers and the size of all of the headers in bytes separated by $|. One
example usage is to reject messages which do not have a Message-Id:
header. However, the Message-Id: header is *NOT* a required header and is
not a guaranteed spam indicator. This ruleset is an example and should
probably not be used in production.
HMessage-Id: $>CheckMessageId
# Record the presence of the header
R$* $: $(storage {MessageIdCheck} $@ OK $) $1
R$* $#error $: 553 Header Error
R$* $: < $&{MessageIdCheck} >
# Clear the macro for the next message
R$* $: $(storage {MessageIdCheck} $) $1
# Has a Message-Id: header
# Allow missing Message-Id: from local mail
R$* $: < $&{client_name} >
# Otherwise, reject the mail
R$* $#error $: 553 Header Error
The features ratecontrol and conncontrol allow to establish connection
limits per client IP address or net. These features can limit the
rate of connections (connections per time unit) or the number of
incoming SMTP connections, respectively. If enabled, appropriate
rulesets are called at the end of check_relay,
i.e., after DNS
blacklists and generic access_db operations. The features require
FEATURE(`access_db') to be listed earlier in the mc file.
Note: FEATURE(`delay_checks') delays those connection control checks
after a recipient address has been received, hence making these
connection control features less useful. To run the checks as early
as possible, specify the parameter `nodelay',
e.g.,
FEATURE(`ratecontrol', `nodelay')
In that case, FEATURE(`delay_checks') has no effect on connection
control (and it must be specified earlier in the mc file).
An optional second argument `terminate' specifies whether the
rulesets should return the error code 421 which will cause
sendmail to terminate the session with that error if it is
returned from check_relay,
i.e., not delayed as explained in
the previous paragraph. Example:
FEATURE(`ratecontrol', `nodelay', `terminate')
In this text, cert will be used as an abbreviation for X.509 certificate,
DN (CN) is the distinguished (common) name of a cert, and CA is a
certification authority, which signs (issues) certs.
For STARTTLS to be offered by sendmail you need to set at least
these variables (the file names and paths are just examples):
On systems which do not have the compile flag HASURANDOM set (see
See
doc/op/op.{me,ps,txt} for more information about these options,
especially the sections ``Certificates for STARTTLS'' and ``PRNG for
Macros related to STARTTLS are:
${cert_issuer} holds the DN of the CA (the cert issuer).
${cert_subject} holds the DN of the cert (called the cert subject).
${cn_issuer} holds the CN of the CA (the cert issuer).
${cn_subject} holds the CN of the cert (called the cert subject).
${tls_version} the
TLS/SSL version used for the connection,
e.g., TLSv1,
${cipher} the cipher used for the connection,
e.g., EDH-DSS-DES-CBC3-SHA,
EDH-RSA-DES-CBC-SHA, DES-CBC-MD5, DES-CBC3-SHA.
${cipher_bits} the keylength (in bits) of the symmetric encryption algorithm
${verify} holds the result of the verification of the presented cert.
OK verification succeeded.
FAIL cert presented but could not be verified,
e.g., the cert of the signing CA is missing.
NONE STARTTLS has not been performed.
TEMP temporary error occurred.
PROTOCOL protocol error occurred (SMTP level).
SOFTWARE STARTTLS handshake failed.
${server_name} the name of the server of the current outgoing SMTP
${server_addr} the address of the server of the current outgoing SMTP
SMTP STARTTLS can allow relaying for remote SMTP clients which have
successfully authenticated themselves. If the verification of the cert
failed (${verify} != OK), relaying is subject to the usual rules.
Otherwise the DN of the issuer is looked up in the access map using the
tag CERTISSUER. If the resulting value is RELAY, relaying is allowed.
If it is SUBJECT, the DN of the cert subject is looked up next in the
access map using the tag CERTSUBJECT. If the value is RELAY, relaying
To make things a bit more flexible (or complicated), the values for
${cert_issuer} and ${cert_subject} can be optionally modified by regular
expressions defined in the m4 variables _CERT_REGEX_ISSUER_ and
_CERT_REGEX_SUBJECT_, respectively. To avoid problems with those macros in
rulesets and map lookups, they are modified as follows: each non-printable
character and the characters '<', '>', '(', ')', '"', '+', ' ' are replaced
by their HEX value with a leading '+'. For example:
Darth+20Mail+20+28Cert+29/Email=darth+2Bcert@endmail.org
(line breaks have been inserted for readability).
The macros which are subject to this encoding are ${cert_subject},
${cert_issuer}, ${cn_subject}, and ${cn_issuer}.
To allow relaying for everyone who can present a cert signed by
Darth+20Mail+20+28Cert+29/Email=darth+2Bcert@endmail.org
Darth+20Mail+20+28Cert+29/Email=darth+2Bcert@endmail.org RELAY
To allow relaying only for a subset of machines that have a cert signed by
Darth+20Mail+20+28Cert+29/Email=darth+2Bcert@endmail.org
Darth+20Mail+20+28Cert+29/Email=darth+2Bcert@endmail.org SUBJECT
- line breaks have been inserted after "CN=" for readability,
each tagged entry must be one (long) line in the access map.
- if OpenSSL 0.9.7 or newer is used then the "Email=" part of a DN
is replaced by "emailAddress=".
Of course it is also possible to write a simple ruleset that allows
relaying for everyone who can present a cert that can be verified,
e.g.,
The rulesets tls_server, tls_client, and tls_rcpt are used to decide whether
an SMTP connection is accepted (or should continue).
tls_server is called when sendmail acts as client after a STARTTLS command
(should) have been issued. The parameter is the value of ${verify}.
tls_client is called when sendmail acts as server, after a STARTTLS command
has been issued, and from check_mail. The parameter is the value of
${verify} and STARTTLS or MAIL, respectively.
Both rulesets behave the same. If no access map is in use, the connection
will be accepted unless ${verify} is SOFTWARE, in which case the connection
is looked up in the access map using the tag
TLS_Srv/TLS_Clt, which is done
with the ruleset LookUpDomain. If no entry is found, ${client_addr}
(${server_addr}) is looked up in the access map (same tag, ruleset
LookUpAddr). If this doesn't result in an entry either, just the tag is
looked up in the access map (included the trailing colon). Notice:
requiring that e-mail is sent to a server only encrypted,
e.g., via
doesn't necessarily mean that e-mail sent to that domain is encrypted.
If the domain has multiple MX servers,
e.g.,
tls_rcpt can be used to address this problem.
tls_rcpt is called before a RCPT TO: command is sent. The parameter is the
current recipient. This ruleset is only defined if FEATURE(`access_db')
is selected. A recipient address user@domain is looked up in the access
map in four formats: TLS_Rcpt:user@domain, TLS_Rcpt:user@, TLS_Rcpt:domain,
and TLS_Rcpt:; the first match is taken.
The result of the lookups is then used to call the ruleset TLS_connection,
which checks the requirement specified by the RHS in the access map against
the actual parameters of the current TLS connection, esp. ${verify} and
${cipher_bits}. Legal RHSs in the access map are:
VERIFY verification must have succeeded
VERIFY:bits verification must have succeeded and ${cipher_bits} must
be greater than or equal bits.
ENCR:bits ${cipher_bits} must be greater than or equal bits.
The RHS can optionally be prefixed by TEMP+ or PERM+ to select a temporary
or permanent error. The default is a temporary error code (403 4.7.0)
unless the macro TLS_PERM_ERR is set during generation of the .cf file.
If a certain level of encryption is required, then it might also be
possible that this level is provided by the security layer from a SASL
algorithm,
e.g., DIGEST-MD5.
Furthermore, there can be a list of extensions added. Such a list
starts with '+' and the items are separated by '++'. Allowed
CN:name name must match ${cn_subject}
CN ${client_name}/${server_name} must match ${cn_subject}
CS:name name must match ${cert_subject}
CI:name name must match ${cert_issuer}
should only be accepted if they have been authenticated. The host which
receives e-mail for darth@endmail.org must present a cert that uses the
Disabling STARTTLS And Setting SMTP Server Features
---------------------------------------------------
By default STARTTLS is used whenever possible. However, there are
some broken MTAs that don't properly implement STARTTLS. To be able
to send to (or receive from) those MTAs, the ruleset try_tls
(srv_features) can be used that work together with the access map.
Entries for the access map must be tagged with Try_TLS (Srv_Features)
and refer to the hostname or IP address of the connecting system.
A default case can be specified by using just the tag. For example,
the following entries in the access map:
will turn off STARTTLS when sending to
broken.server (or any host
in that domain), and request a client certificate during the TLS
handshake only for hosts in
my.domain. The valid entries on the RHS
for Srv_Features are listed in the Sendmail Installation and
The Received: header reveals whether STARTTLS has been used. It contains an
(version=${tls_version} cipher=${cipher} bits=${cipher_bits} verify=${verify})
+--------------------------------+
| ADDING NEW MAILERS OR RULESETS |
+--------------------------------+
Sometimes you may need to add entirely new mailers or rulesets. They
should be introduced with the constructs MAILER_DEFINITIONS and
LOCAL_RULESETS respectively. For example:
Local additions for the rulesets srv_features, try_tls, tls_rcpt,
tls_client, and tls_server can be made using LOCAL_SRV_FEATURES,
LOCAL_TRY_TLS, LOCAL_TLS_RCPT, LOCAL_TLS_CLIENT, and LOCAL_TLS_SERVER,
respectively. For example, to add a local ruleset that decides
whether to try STARTTLS in a sendmail client, use:
Note: you don't need to add a name for the ruleset, it is implicitly
defined by using the appropriate macro.
+-------------------------+
| ADDING NEW MAIL FILTERS |
+-------------------------+
Sendmail supports mail filters to filter incoming SMTP messages according
to the "Sendmail Mail Filter API" documentation. These filters can be
configured in your mc file using the two commands:
MAIL_FILTER(`name', `equates')
INPUT_MAIL_FILTER(`name', `equates')
The first command, MAIL_FILTER(), simply defines a filter with the given
name and equates. For example:
The INPUT_MAIL_FILTER() command performs the same actions as MAIL_FILTER
but also populates the m4 variable `confINPUT_MAIL_FILTERS' with the name
of the filter such that the filter will actually be called by sendmail.
For example, the two commands:
INPUT_MAIL_FILTER(`spamcheck', `S=inet:2525@localhost, F=T')
are equivalent to the three commands:
MAIL_FILTER(`spamcheck', `S=inet:2525@localhost, F=T')
define(`confINPUT_MAIL_FILTERS', `archive, spamcheck')
In general, INPUT_MAIL_FILTER() should be used unless you need to define
more filters than you want to use for `confINPUT_MAIL_FILTERS'.
Note that setting `confINPUT_MAIL_FILTERS' after any INPUT_MAIL_FILTER()
commands will clear the list created by the prior INPUT_MAIL_FILTER()
+-------------------------+
| QUEUE GROUP DEFINITIONS |
+-------------------------+
In addition to the queue directory (which is the default queue group
called "mqueue"), sendmail can deal with multiple queue groups, which
are collections of queue directories with the same behaviour. Queue
groups can be defined using the command:
QUEUE_GROUP(`name', `equates')
For details about queue groups, please see
doc/op/op.{me,ps,txt}.
+-------------------------------+
| NON-SMTP BASED CONFIGURATIONS |
+-------------------------------+
These configuration files are designed primarily for use by
SMTP-based sites. They may not be well tuned for UUCP-only or
UUCP-primarily nodes (the latter is defined as a small local net
connected to the rest of the world via UUCP). However, there is
one hook to handle some special cases.
You can define a ``smart host'' that understands a richer address syntax
define(`SMART_HOST', `mailer:hostname')
In this case, the ``mailer:'' defaults to "relay". Any messages that
can't be handled using the usual UUCP rules are passed to this host.
If you are on a local SMTP-based net that connects to the outside
world via UUCP, you can use LOCAL_NET_CONFIG to add appropriate rules.
define(`SMART_HOST', `uucp-new:uunet')
R$* < @ $* .$m. > $* $#smtp $@ $2.$m. $: $1 < @ $2.$m. > $3
This will cause all names that end in your domain name ($m) to be sent
via SMTP; anything else will be sent via uucp-new (smart UUCP) to uunet.
If you have FEATURE(`nocanonify'), you may need to omit the dots after
the $m. If you are running a local DNS inside your domain which is
not otherwise connected to the outside world, you probably want to
R$* < @ $* . > $* $#smtp $@ $2. $: $1 < @ $2. > $3
That is, send directly only to things you found in your DNS lookup;
anything else goes through SMART_HOST.
You may need to turn off the anti-spam rules in order to accept
UUCP mail with FEATURE(`promiscuous_relay') and
FEATURE(`accept_unresolvable_domains').
Normally, the $j macro is automatically defined to be your fully
qualified domain name (FQDN). Sendmail does this by getting your
host name using gethostname and then calling gethostbyname on the
result. For example, in some environments gethostname returns
only the root of the host name (such as "foo"); gethostbyname is
supposed to return the FQDN ("
foo.bar.com"). In some (fairly rare)
cases, gethostbyname may fail to return the FQDN. In this case
you MUST define confDOMAIN_NAME to be your fully qualified domain
name. This is usually done using:
define(`confDOMAIN_NAME', `$w.$m')dnl
+-----------------------------------+
| ACCEPTING MAIL FOR MULTIPLE NAMES |
+-----------------------------------+
If your host is known by several different names, you need to augment
class {w}. This is a list of names by which your host is known, and
anything sent to an address using a host name in this list will be
treated as local mail. You can do this in two ways: either create the
line), and use ``FEATURE(`use_cw_file')'' in the .mc file, or add
name of the host, rather than a short name.
If you want to have different address in different domains, take
a look at the virtusertable feature, which is also explained at
To use FEATURE(`mailertable'), you will have to create an external
database containing the routing information for various domains.
For example, a mailertable file in text format might be:
database version of the mailertable is built using:
The semantics are simple. Any LHS entry that does not begin with
a dot matches the full host name indicated. LHS entries beginning
with a dot match anything ending with that domain name (including
the leading dot) -- that is, they can be thought of as having a
leading ".+" regular expression pattern for a non-empty sequence of
characters. Matching is done in order of most-to-least qualified
-- for example, even though "
.my.domain" is listed first in the
entry since it is more explicit. Note: e-mail to "user@my.domain"
does not match any entry in the above table. You need to have
The RHS should always be a "mailer:host" pair. The mailer is the
configuration name of a mailer (that is, an M line in the
sendmail.cf file). The "host" will be the hostname passed to
that mailer. In domain-based matches (that is, those with leading
dots) the "%1" may be used to interpolate the wildcarded part of
the host name. For example, the first line above sends everything
the (presumably experimental) xnet mailer.
In some cases you may want to temporarily turn off MX records,
particularly on gateways. For example, you may want to MX
everything in a domain to one machine that then forwards it
directly. To do this, you might use the DNS configuration:
The [square brackets] turn off MX records for this host only.
If you didn't do this, the mailertable would use the MX record
again, which would give you an MX loop. Note that the use of
wildcard MX records is almost always a bad idea. Please avoid
+--------------------------------+
| USING USERDB TO MAP FULL NAMES |
+--------------------------------+
The user database was not originally intended for mapping full names
it that way. (it is recommended that you set up aliases for this
purpose instead -- since you can specify multiple alias files, this
is fairly easy.) The intent was to locate the default maildrop at
a site, but allow you to override this by sending to a specific host.
If you decide to set up the user database in this fashion, it is
imperative that you not use FEATURE(`stickyhost') -- otherwise,
e-mail sent to Full.Name@local.host.name will be rejected.
To build the internal form of the user database, use:
As a general rule, it is an extremely bad idea to using full names
as e-mail addresses, since they are not in any sense unique. For
example, the UNIX software-development community has at least two
well-known Peter Deutsches, and at one time Bell Labs had two
Stephen R. Bournes with offices along the same hallway. Which one
will be forced to suffer the indignity of being Stephen_R_Bourne_2?
The less famous of the two, or the one that was hired later?
Finger should handle full names (and be fuzzy). Mail should use
handles, and not be fuzzy.
+--------------------------------+
| MISCELLANEOUS SPECIAL FEATURES |
+--------------------------------+
Sometimes it is convenient to merge configuration on a
centralized mail machine, for example, to forward all
root mail to a mail server. In this case it might be
useful to be able to treat the root addresses as a class
of addresses with subtle differences. You can do this
using plussed users. For example, a client might include
root: root+client1@server
On the server, this will match an alias for "root+client1".
If that is not found, the alias "root+*" will be tried,
A lot of sendmail security comes down to you. Sendmail 8 is much
more careful about checking for security problems than previous
versions, but there are some things that you still need to watch
* Make sure the aliases file is not writable except by trusted
system personnel. This includes both the text and database
* Make sure that other files that sendmail reads, such as the
mailertable, are only writable by trusted system personnel.
* The queue directory should not be world writable PARTICULARLY
if your system allows "file giveaways" (that is, if a non-root
user can chown any file they own to any other user).
* If your system allows file giveaways, DO NOT create a publically
writable directory for forward files. This will allow anyone
to steal anyone else's e-mail. Instead, create a script that
copies the .forward file from users' home directories once a
night (if you want the non-NFS-mounted forward directory).
* If your system allows file giveaways, you'll find that
sendmail is much less trusting of :include: files -- in
/etc/shells before they will be trusted (that is, before
files and programs listed in them will be honored).
In general, file giveaways are a mistake -- if you can turn them
+--------------------------------+
| TWEAKING CONFIGURATION OPTIONS |
+--------------------------------+
There are a large number of configuration options that don't normally
need to be changed. However, if you feel you need to tweak them,
you can define the following M4 variables. Note that some of these
variables require formats that are defined in RFC 2821 or RFC 2822.
Before changing them you need to make sure you do not violate those
(and other relevant) RFCs.
This list is shown in four columns: the name you define, the default
value for that definition, the option or macro that is affected
(either Ox for an option or Dx for a macro), and a brief description.
Some options are likely to be deprecated in future versions -- that is,
the option is only included to provide back-compatibility. These are
Remember that these options are M4 variables, and hence may need to
be quoted. In particular, arguments with commas will usually have to
be ``double quoted, like this phrase'' to avoid having the comma
confuse things. This is common for alias file definitions and for
M4 Variable Name Configuration [Default] & Description
================ ============= =======================
confMAILER_NAME $n macro [MAILER-DAEMON] The sender name used
for internally generated outgoing
confDOMAIN_NAME $j macro If defined, sets $j. This should
only be done if your system cannot
determine your local domain name,
and then it should be set to
confCF_VERSION $Z macro If defined, this is appended to the
configuration version name.
confLDAP_CLUSTER ${sendmailMTACluster} macro
If defined, this is the LDAP
cluster to use for LDAP searches
as described above in ``USING LDAP
FOR ALIASES, MAPS, AND CLASSES''.
confFROM_HEADER From: [$?x$x <$g>$|$g$.] The format of an
internally generated From: address.
confRECEIVED_HEADER Received:
[$?sfrom $s $.$?_($?s$|from $.$_)
$.$?{auth_type}(authenticated)
$.by $j ($v/$Z)$?r with $r$. id $i$?u
The format of the Received: header
in messages passed through this host.
It is unwise to try to change this.
confMESSAGEID_HEADER Message-Id: [<$t.$i@$j>] The format of an
internally generated Message-Id:
of file used to get the local
additions to class {w} (local host
file used to get the local additions
to class {t} (trusted users).
file used to get the local additions
to class {R} (hosts allowed to relay).
confTRUSTED_USERS Ct class [no default] Names of users to add to
the list of trusted users. This list
always includes root, uucp, and daemon.
See also FEATURE(`use_ct_file').
confTRUSTED_USER TrustedUser [no default] Trusted user for file
ownership and starting the daemon.
confTRUSTED_USERS (see above).
confSMTP_MAILER - [esmtp] The mailer name used when
SMTP connectivity is required.
confUUCP_MAILER - [uucp-old] The mailer to be used by
default for bang-format recipient
addresses. See also discussion of
class {U}, class {Y}, and class {Z}
in the MAILER(`uucp') section.
confLOCAL_MAILER - [local] The mailer name used when
local connectivity is required.
confRELAY_MAILER - [relay] The default mailer name used
for relaying any mail (
e.g., to a
BITNET_RELAY, a SMART_HOST, or
whatever). This can reasonably be
"uucp-new" if you are on a
confSEVEN_BIT_INPUT SevenBitInput [False] Force input to seven bits?
confEIGHT_BIT_HANDLING EightBitMode [pass8] 8-bit data handling
confALIAS_WAIT AliasWait [10m] Time to wait for alias file
rebuild until you get bored and
decide that the apparently pending
confMIN_FREE_BLOCKS MinFreeBlocks [100] Minimum number of free blocks on
queue filesystem to accept SMTP mail.
where minfree was the number of free
blocks and maxsize was the maximum
message size. Use confMAX_MESSAGE_SIZE
for the second value now.)
confMAX_MESSAGE_SIZE MaxMessageSize [infinite] The maximum size of messages
that will be accepted (in bytes).
confBLANK_SUB BlankSub [.] Blank (space) substitution
confCON_EXPENSIVE HoldExpensive [False] Avoid connecting immediately
to mailers marked expensive.
confCHECKPOINT_INTERVAL CheckpointInterval
[10] Checkpoint queue files every N
confDELIVERY_MODE DeliveryMode [background] Default delivery mode.
confERROR_MODE ErrorMode [print] Error message mode.
confERROR_MESSAGE ErrorHeader [undefined] Error message
header/file.
confSAVE_FROM_LINES SaveFromLine Save extra leading From_ lines.
confTEMP_FILE_MODE TempFileMode [0600] Temporary file mode.
confMATCH_GECOS MatchGECOS [False] Match GECOS field.
confMAX_HOP MaxHopCount [25] Maximum hop count.
confIGNORE_DOTS* IgnoreDots [False; always False in -bs or -bd
mode] Ignore dot as terminator for
confBIND_OPTS ResolverOptions [undefined] Default options for DNS
confMIME_FORMAT_ERRORS* SendMimeErrors [True] Send error messages as MIME-
encapsulated messages per RFC 1344.
confFORWARD_PATH ForwardPath [$z/.forward.$w:$z/.forward]
The colon-separated list of places to
search for .forward files.
N.B.: see
the Security Notes section.
confMCI_CACHE_SIZE ConnectionCacheSize
[2] Size of open connection cache.
confMCI_CACHE_TIMEOUT ConnectionCacheTimeout
[5m] Open connection cache timeout.
confHOST_STATUS_DIRECTORY HostStatusDirectory
[undefined] If set, host status is kept
on disk between sendmail runs in the
named directory tree. This need not be
a full pathname, in which case it is
interpreted relative to the queue
confSINGLE_THREAD_DELIVERY SingleThreadDelivery
[False] If this option and the
HostStatusDirectory option are both
set, single thread deliveries to other
hosts. That is, don't allow any two
sendmails on this host to connect
simultaneously to any other single
host. This can slow down delivery in
some cases, in particular since a
cached but otherwise idle connection
to a host will prevent other sendmails
from connecting to the other host.
confUSE_ERRORS_TO* UseErrorsTo [False] Use the Errors-To: header to
deliver error messages. This should
not be necessary because of general
confLOG_LEVEL LogLevel [9] Log level.
confME_TOO MeToo [True] Include sender in group
expansions. This option is
deprecated and will be removed from
confCHECK_ALIASES CheckAliases [False] Check RHS of aliases when
running newaliases. Since this does
DNS lookups on every address, it can
slow down the alias rebuild process
considerably on large alias files.
confOLD_STYLE_HEADERS* OldStyleHeaders [True] Assume that headers without
special chars are old style.
confPRIVACY_FLAGS PrivacyOptions [authwarnings] Privacy flags.
confCOPY_ERRORS_TO PostmasterCopy [undefined] Address for additional
copies of all error messages.
confQUEUE_FACTOR QueueFactor [600000] Slope of queue-only function.
confQUEUE_FILE_MODE QueueFileMode [undefined] Default permissions for
queue files (octal). If not set,
sendmail uses 0600 unless its real
and effective uid are different in
confDONT_PRUNE_ROUTES DontPruneRoutes [False] Don't prune down route-addr
syntax addresses to the minimum
confSAFE_QUEUE* SuperSafe [True] Commit all messages to disk
connect() to complete. This can only
shorten connection timeouts; the kernel
silently enforces an absolute maximum
(which varies depending on the system).
applies only to the very first attempt
to connect to a host in a message.
This allows a single very fast pass
followed by more careful delivery
[0] The overall timeout waiting for
all connection for a single delivery
attempt to succeed. If 0, no overall
confTO_HELO
Timeout.helo [5m] The timeout waiting for a response
to a HELO or EHLO command.
response to the MAIL command.
confTO_RCPT
Timeout.rcpt [1h] The timeout waiting for a response
[5m] The timeout waiting for a 354
response from the DATA command.
[1h] The timeout waiting for a block
[1h] The timeout waiting for a response
to the final "." that terminates a
confTO_RSET
Timeout.rset [5m] The timeout waiting for a response
confTO_QUIT
Timeout.quit [2m] The timeout waiting for a response
confTO_MISC
Timeout.misc [2m] The timeout waiting for a response
waiting for a command to be issued.
response to an IDENT query.
[60s] The timeout waiting for a file
(
e.g., :include: file) to be opened.
confTO_LHLO
Timeout.lhlo [2m] The timeout waiting for a response
[1h] The timeout waiting for a
response to an SMTP STARTTLS command.
[2m] The timeout for a complete
control socket transaction to complete.
[5d] The timeout before a message is
returned as undeliverable.
confTO_QUEUERETURN_NORMAL
[undefined] As above, for normal
confTO_QUEUERETURN_URGENT
[undefined] As above, for urgent
confTO_QUEUERETURN_NONURGENT
[undefined] As above, for non-urgent
[undefined] As above, for delivery
status notification messages.
[4h] The timeout before a warning
message is sent to the sender telling
them that the message has been
[undefined] As above, for normal
[undefined] As above, for urgent
confTO_QUEUEWARN_NONURGENT
[undefined] As above, for non-urgent
[undefined] As above, for delivery
status notification messages.
[30m] How long information about host
statuses will be maintained before it
is considered stale and the host should
be retried. This applies both within
a single queue run and to persistent
[varies] Sets the resolver's
retransmission time interval (in
[varies] Sets the resolver's
retransmission time interval (in
seconds) for the first attempt to
[varies] Sets the resolver's
retransmission time interval (in
seconds) for all resolver lookups
except the first delivery attempt.
[varies] Sets the number of times
to retransmit a resolver query.
[varies] Sets the number of times
to retransmit a resolver query for
the first attempt to deliver a
[varies] Sets the number of times
to retransmit a resolver query for
all resolver lookups except the
confTIME_ZONE TimeZoneSpec [USE_SYSTEM] Time zone info -- can be
USE_SYSTEM to use the system's idea,
USE_TZ to use the user's TZ envariable,
or something else to force that value.
confDEF_USER_ID DefaultUser [1:1] Default user id.
confUSERDB_SPEC UserDatabaseSpec
[undefined] User database
confFALLBACK_MX FallbackMXhost [undefined] Fallback MX host.
confFALLBACK_SMARTHOST FallbackSmartHost
[undefined] Fallback smart host.
confTRY_NULL_MX_LIST TryNullMXList [False] If this host is the best MX
for a host and other arrangements
haven't been made, try connecting
to the host directly; normally this
confQUEUE_LA QueueLA [varies] Load average at which
queue-only function kicks in.
Default values is (8 * numproc)
where numproc is the number of
processors online (if that can be
confREFUSE_LA RefuseLA [varies] Load average at which
incoming SMTP connections are
refused. Default values is (12 *
numproc) where numproc is the
number of processors online (if
confREJECT_LOG_INTERVAL RejectLogInterval [3h] Log interval when
refusing connections for this long.
confDELAY_LA DelayLA [0] Load average at which sendmail
will sleep for one second on most
SMTP commands and before accepting
connections. 0 means no limit.
confMAX_ALIAS_RECURSION MaxAliasRecursion
[10] Maximum depth of alias recursion.
confMAX_DAEMON_CHILDREN MaxDaemonChildren
[undefined] The maximum number of
children the daemon will permit. After
this number, connections will be
rejected. If not set or <= 0, there is
confMAX_HEADERS_LENGTH MaxHeadersLength
[32768] Maximum length of the sum
confMAX_MIME_HEADER_LENGTH MaxMimeHeaderLength
[undefined] Maximum length of
certain MIME header field values.
confCONNECTION_RATE_THROTTLE ConnectionRateThrottle
[undefined] The maximum number of
connections permitted per second per
daemon. After this many connections
are accepted, further connections
will be delayed. If not set or <= 0,
confCONNECTION_RATE_WINDOW_SIZE ConnectionRateWindowSize
[60s] Define the length of the
interval for which the number of
incoming connections is maintained.
confWORK_RECIPIENT_FACTOR
RecipientFactor [30000] Cost of each recipient.
confSEPARATE_PROC ForkEachJob [False] Run all deliveries in a
confWORK_CLASS_FACTOR ClassFactor [1800] Priority multiplier for class.
confWORK_TIME_FACTOR RetryFactor [90000] Cost of each delivery attempt.
confQUEUE_SORT_ORDER QueueSortOrder [Priority] Queue sort algorithm:
Priority, Host, Filename, Random,
confMIN_QUEUE_AGE MinQueueAge [0] The minimum amount of time a job
must sit in the queue between queue
runs. This allows you to set the
queue run interval low for better
responsiveness without trying all
confDEF_CHAR_SET DefaultCharSet [unknown-8bit] When converting
unlabeled 8 bit input to MIME, the
character set to use by default.
confSERVICE_SWITCH_FILE ServiceSwitchFile
to use for the service switch on
systems that do not have a
confHOSTS_FILE HostsFile [
/etc/hosts] The file to use when doing
"file" type access of hosts names.
confDIAL_DELAY DialDelay [0s] If a connection fails, wait this
long and try again. Zero means "don't
retry". This is to allow "dial on
demand" connections to have enough time
to complete a connection.
confNO_RCPT_ACTION NoRecipientAction
[none] What to do if there are no legal
recipient fields (To:, Cc: or Bcc:)
in the message. Legal values can
be "none" to just leave the
nonconforming message as is, "add-to"
to add a To: header with all the
known recipients (which may expose
blind recipients), "add-apparently-to"
to do the same but use Apparently-To:
instead of To: (strongly discouraged
in accordance with IETF standards),
"add-bcc" to add an empty Bcc:
header, or "add-to-undisclosed" to
``To: undisclosed-recipients:;''.
confSAFE_FILE_ENV SafeFileEnvironment
[undefined] If set, sendmail will do a
chroot() into this directory before
confCOLON_OK_IN_ADDR ColonOkInAddr [True unless Configuration Level > 6]
If set, colons are treated as a regular
character in addresses. If not set,
they are treated as the introducer to
the RFC 822 "group" syntax. Colons are
handled properly in route-addrs. This
option defaults on for V5 and lower
confMAX_QUEUE_RUN_SIZE MaxQueueRunSize [0] If set, limit the maximum size of
any given queue run to this number of
entries. Essentially, this will stop
reading each queue directory after this
number of entries are reached; it does
_not_ pick the highest priority jobs,
so this should be as large as your
system can tolerate. If not set, there
confMAX_QUEUE_CHILDREN MaxQueueChildren
[undefined] Limits the maximum number
of concurrent queue runners active.
This is to keep system resources used
within a reasonable limit. Relates to
Queue Groups and ForkEachJob.
confMAX_RUNNERS_PER_QUEUE MaxRunnersPerQueue
[1] Only active when MaxQueueChildren
defined. Controls the maximum number
of queue runners (aka queue children)
active at the same time in a work
group. See also MaxQueueChildren.
confDONT_EXPAND_CNAMES DontExpandCnames
[False] If set, $[ ... $] lookups that
do DNS based lookups do not expand
CNAME records. This currently violates
the published standards, but the IETF
seems to be moving toward legalizing
with this option set a lookup of
you may not see any effect until your
downstream neighbors stop doing CNAME
confFROM_LINE UnixFromLine [From $g $d] The From_ line used
when sending to files or programs.
confSINGLE_LINE_FROM_HEADER SingleLineFromHeader
[False] From: lines that have
embedded newlines are unwrapped
confALLOW_BOGUS_HELO AllowBogusHELO [False] Allow HELO SMTP command that
does not include a host name.
confMUST_QUOTE_CHARS MustQuoteChars [.'] Characters to be quoted in a full
name phrase (@,;:\()[] are automatic).
confOPERATORS OperatorChars [.:%@!^/[]+] Address operator
confSMTP_LOGIN_MSG SmtpGreetingMessage
The initial (spontaneous) SMTP
greeting message. The word "ESMTP"
will be inserted between the first and
second words to convince other
sendmails to try to speak ESMTP.
confDONT_INIT_GROUPS DontInitGroups [False] If set, the initgroups(3)
routine will never be invoked. You
might want to do this if you are
running NIS and you have a large group
map, since this call does a sequential
scan of the map; in a large site this
can cause your ypserv to run
essentially full time. If you set
this, agents run on behalf of users
will only have their primary
confUNSAFE_GROUP_WRITES UnsafeGroupWrites
[True] If set, group-writable
:include: and .forward files are
considered "unsafe", that is, programs
and files cannot be directly referenced
from such files. World-writable files
are always considered unsafe.
Notice: this option is deprecated and
will be removed in future versions;
Set GroupWritableForwardFileSafe
and GroupWritableIncludeFileSafe in
DontBlameSendmail if required.
confCONNECT_ONLY_TO ConnectOnlyTo [undefined] override connection
confCONTROL_SOCKET_NAME ControlSocketName
[undefined] Control socket for daemon
confDOUBLE_BOUNCE_ADDRESS DoubleBounceAddress
[postmaster] If an error occurs when
sending an error message, send that
"double bounce" error message to this
address. If it expands to an empty
string, double bounces are dropped.
confSOFT_BOUNCE SoftBounce [False] If set, issue temporary errors
(4xy) instead of permanent errors
(5xy). This can be useful during
testing of a new configuration to
avoid erroneous bouncing of mails.
confDEAD_LETTER_DROP DeadLetterDrop [undefined] Filename to save bounce
messages which could not be returned
to the user or sent to postmaster.
If not set, the queue file will
confRRT_IMPLIES_DSN RrtImpliesDsn [False] Return-Receipt-To: header
confRUN_AS_USER RunAsUser [undefined] If set, become this user
when reading and delivering mail.
Causes all file reads (
e.g., .forward
and :include: files) to be done as
this user. Also, all programs will
be run as this user, and all output
files will be written as this user.
confMAX_RCPTS_PER_MESSAGE MaxRecipientsPerMessage
[infinite] If set, allow no more than
the specified number of recipients in
an SMTP envelope. Further recipients
receive a 452 error code (
i.e., they
are deferred for the next delivery
confBAD_RCPT_THROTTLE BadRcptThrottle [infinite] If set and the specified
number of recipients in a single SMTP
transaction have been rejected, sleep
for one second after each subsequent
RCPT command in that transaction.
confDONT_PROBE_INTERFACES DontProbeInterfaces
[False] If set, sendmail will _not_
insert the names and addresses of any
local interfaces into class {w}
(list of known "equivalent" addresses).
If you set this, you must also include
some support for these addresses (
e.g.,
in a mailertable entry) -- otherwise,
mail to addresses in this list will
bounce with a configuration error.
If set to "loopback" (without
quotes), sendmail will skip
loopback interfaces (
e.g., "lo0").
confPID_FILE PidFile [system dependent] Location of pid
confPROCESS_TITLE_PREFIX ProcessTitlePrefix
[undefined] Prefix string for the
process title shown on 'ps' listings.
confDONT_BLAME_SENDMAIL DontBlameSendmail
[safe] Override sendmail's file
safety checks. This will definitely
compromise system security and should
not be used unless absolutely
confREJECT_MSG - [550 Access denied] The message
given if the access database contains
REJECT in the value portion.
confRELAY_MSG - [550 Relaying denied] The message
given if an unauthorized relaying
confDF_BUFFER_SIZE DataFileBufferSize
[4096] The maximum size of a
memory-buffered data (df) file
before a disk-based file is used.
confXF_BUFFER_SIZE XScriptFileBufferSize
[4096] The maximum size of a
memory-buffered transcript (xf)
file before a disk-based file is
confTLS_SRV_OPTIONS TLSSrvOptions If this option is 'V' no client
verification is performed,
i.e.,
the server doesn't ask for a
confLDAP_DEFAULT_SPEC LDAPDefaultSpec [undefined] Default map
specification for LDAP maps. The
value should only contain LDAP
specific settings such as "-h host
-p port -d bindDN", etc. The
settings will be used for all LDAP
maps unless they are specified in
the individual map specification
confCACERT_PATH CACertPath [undefined] Path to directory
confCACERT CACertFile [undefined] File containing one CA
confSERVER_CERT ServerCertFile [undefined] File containing the
cert of the server,
i.e., this cert
is used when sendmail acts as
confSERVER_KEY ServerKeyFile [undefined] File containing the
private key belonging to the server
confCLIENT_CERT ClientCertFile [undefined] File containing the
cert of the client,
i.e., this cert
is used when sendmail acts as
confCLIENT_KEY ClientKeyFile [undefined] File containing the
private key belonging to the client
confCRL CRLFile [undefined] File containing certificate
revocation status, useful for X.509v3
authentication. Note that CRL requires
at least OpenSSL version 0.9.7.
confDH_PARAMETERS DHParameters [undefined] File containing the
confRAND_FILE RandFile [undefined] File containing random
data (use prefix file:) or the
name of the UNIX socket if EGD is
used (use prefix egd:). STARTTLS
requires this option if the compile
flag HASURANDOM is not set (see
confNICE_QUEUE_RUN NiceQueueRun [undefined] If set, the priority of
queue runners is set the given value
confDIRECT_SUBMISSION_MODIFIERS DirectSubmissionModifiers
[undefined] Defines {daemon_flags}
confUSE_MSP UseMSP [undefined] Use as mail submission
confDELIVER_BY_MIN DeliverByMin [0] Minimum time for Deliver By
SMTP Service Extension (RFC 2852).
confREQUIRES_DIR_FSYNC RequiresDirfsync [true] RequiresDirfsync can
be used to turn off the compile time
flag REQUIRES_DIR_FSYNC at runtime.
confSHARED_MEMORY_KEY SharedMemoryKey [0] Key for shared memory.
confSHARED_MEMORY_KEY_FILE
[undefined] File where the
automatically selected key for
confFAST_SPLIT FastSplit [1] If set to a value greater than
zero, the initial MX lookups on
addresses is suppressed when they
are sorted which may result in
faster envelope splitting. If the
mail is submitted directly from the
command line, then the value also
limits the number of processes to
confMAILBOX_DATABASE MailboxDatabase [pw] Type of lookup to find
information about local mailboxes.
confDEQUOTE_OPTS - [empty] Additional options for the
confMAX_NOOP_COMMANDS MaxNOOPCommands [20] Maximum number of "useless"
commands before the SMTP server
will slow down responding.
confHELO_NAME HeloName If defined, use as name for
EHLO/HELOconfINPUT_MAIL_FILTERS InputMailFilters
A comma separated list of filters
which determines which filters and
the invocation sequence are
contacted for incoming SMTP
messages. If none are set, no
filters will be contacted.
actions, defaults to LogLevel.
[j, _, {daemon_name}, {if_name},
{if_addr}] Macros to transmit to
milters when a session connection
[{tls_version}, {cipher},
{cipher_bits}, {cert_subject},
{cert_issuer}] Macros to transmit to
[i, {auth_type}, {auth_authen},
{auth_ssf}, {auth_author},
{mail_mailer}, {mail_host},
{mail_addr}] Macros to transmit to
milters after MAIL FROM command.
[{rcpt_mailer}, {rcpt_host},
{rcpt_addr}] Macros to transmit to
milters after RCPT TO command.
[{msg_id}] Macros to transmit to
milters after the terminating
Macros to transmit to milters
after the end of headers.
Macros to transmit to milters
after DATA command is received.
See also the description of OSTYPE for some parameters that can be
tweaked (generally pathnames to mailers).
ClientPortOptions and DaemonPortOptions are special cases since multiple
CLIENT_OPTIONS(`field1=value1,field2=value2,...')
DAEMON_OPTIONS(`field1=value1,field2=value2,...')
Note that multiple CLIENT_OPTIONS() commands (and therefore multiple
ClientPortOptions settings) are allowed in order to give settings for each
protocol family (
e.g., one for Family=inet and one for Family=inet6). A
restriction placed on one family only affects outgoing connections on that
If DAEMON_OPTIONS is not used, then the default is
DAEMON_OPTIONS(`Port=smtp, Name=MTA')
DAEMON_OPTIONS(`Port=587, Name=MSA, M=E')
If you use one DAEMON_OPTIONS macro, it will alter the parameters
of the first of these. The second will still be defaulted; it
represents a "Message Submission Agent" (MSA) as defined by RFC
2476 (see below). To turn off the default definition for the MSA,
use FEATURE(`no_default_msa') (see also FEATURES). If you use
additional DAEMON_OPTIONS macros, they will add additional daemons.
Example 1: To change the port for the SMTP listener, while
still using the MSA default, use
DAEMON_OPTIONS(`Port=925, Name=MTA')
Example 2: To change the port for the MSA daemon, while still
using the default SMTP port, use
FEATURE(`no_default_msa')
DAEMON_OPTIONS(`Name=MTA')
DAEMON_OPTIONS(`Port=987, Name=MSA, M=E')
Note that if the first of those DAEMON_OPTIONS lines were omitted, then
there would be no listener on the standard SMTP port.
Example 3: To listen on both IPv4 and IPv6 interfaces, use
DAEMON_OPTIONS(`Name=MTA-v4, Family=inet')
DAEMON_OPTIONS(`Name=MTA-v6, Family=inet6')
A "Message Submission Agent" still uses all of the same rulesets for
processing the message (and therefore still allows message rejection via
the check_* rulesets). In accordance with the RFC, the MSA will ensure
that all domains in envelope addresses are fully qualified if the message
is relayed to another MTA. It will also enforce the normal address syntax
rules and log error messages. Additionally, by using the M=a modifier you
can require authentication before messages are accepted by the MSA.
Notice: Do NOT use the 'a' modifier on a public accessible MTA! Finally,
the M=E modifier shown above disables ETRN as required by RFC 2476.
Mail filters can be defined using the INPUT_MAIL_FILTER() and MAIL_FILTER()
MAIL_FILTER(`myfilter', `S=inet:3333@localhost')
The INPUT_MAIL_FILTER() command causes the filter(s) to be called in the
same order they were specified by also setting confINPUT_MAIL_FILTERS. A
filter can be defined without adding it to the input filter list by using
MAIL_FILTER() instead of INPUT_MAIL_FILTER() in your .mc file.
Alternatively, you can reset the list of filters and their order by setting
confINPUT_MAIL_FILTERS option after all INPUT_MAIL_FILTER() commands in
+----------------------------+
| MESSAGE SUBMISSION PROGRAM |
+----------------------------+
This section contains a list of caveats and
a few hints how for those who want to tweak the default configuration
absolutely sure you need them. Options you may want to change
- confTRUSTED_USERS, FEATURE(`use_ct_file'), and confCT_FILE for
avoiding X-Authentication warnings.
- confTIME_ZONE to change it from the default `USE_TZ'.
- confDELIVERY_MODE is set to interactive in
msp.m4 instead
of the default background mode.
- FEATURE(stickyhost) and LOCAL_RELAY to send unqualified addresses
to the LOCAL_RELAY instead of the default relay.
The MSP performs hostname canonicalization by default. Mail may end
up for various DNS related reasons in the MSP queue. This problem
can be minimized by using
FEATURE(`nocanonify', `canonify_hosts')
define(`confDIRECT_SUBMISSION_MODIFIERS', `C')
See the discussion about nocanonify for possible side effects.
Some things are not intended to work with the MSP. These include
features that influence the delivery process (
e.g., mailertable,
aliases), or those that are only important for a SMTP server (
e.g.,
virtusertable, DaemonPortOptions, multiple queues). Moreover,
relaxing certain restrictions (RestrictQueueRun, permissions on
queue directory) or adding features (
e.g., enabling
prog/file mailer)
can cause security problems.
Other things don't work well with the MSP and require tweaking or
The file and the map created by makemap should be owned by smmsp,
its group should be smmsp, and it should have mode 640.
those should not be changed at all. Some of the features and options
can be overridden if really necessary. It is a bit tricky to do
this, because it depends on the actual way the option is defined
the modified value must be defined after
If it is conditionally defined (
i.e., ifdef()) then the desired
value must be defined before the FEATURE line in the .mc file.
+--------------------------+
| FORMAT OF FILES AND MAPS |
+--------------------------+
Files that define classes,
i.e., F{classname}, consist of lines
each of which contains a single element of the class. For example,
Maps must be created using makemap(8) ,
e.g.,
In general, a text file from which a map is created contains lines
where 'key' and 'value' are also called LHS and RHS, respectively.
By default, the delimiter between LHS and RHS is a non-empty sequence
of white space characters.
Within this directory are several subdirectories, to wit:
m4 General support routines. These are typically
very important and should not be changed without
very careful consideration.
cf The configuration files themselves. They have
".mc" suffixes, and must be run through m4 to
become complete. The resulting output should
ostype Definitions describing a particular operating
system type. These should always be referenced
using the OSTYPE macro in the .mc file. Examples
include "bsd4.3", "bsd4.4", "sunos3.5", and
domain Definitions describing a particular domain, referenced
using the DOMAIN macro in the .mc file. These are
mailer Descriptions of mailers. These are referenced using
the MAILER macro in the .mc file.
sh Shell files used when building the .cf file from the
.mc file in the cf subdirectory.
feature These hold special orthogonal features that you might
want to include. They should be referenced using
hack Local hacks. These can be referenced using the HACK
macro. They shouldn't be of more than voyeuristic
siteconfig Site configuration --
e.g., tables of locally connected
+------------------------+
| ADMINISTRATIVE DETAILS |
+------------------------+
The following sections detail usage of certain internal parts of the
sendmail.cf file. Read them carefully if you are trying to modify
the current model. If you find the above descriptions adequate, these
should be {boring, confusing, tedious, ridiculous} (pick one or more).
RULESETS (* means built in to sendmail)
5 * Local address rewrite (after aliasing)
1x mailer rules (sender qualification)
2x mailer rules (recipient qualification)
3x mailer rules (sender header qualification)
4x mailer rules (recipient header qualification)
5x mailer subroutines (general)
6x mailer subroutines (general)
7x mailer subroutines (general)
90 Mailertable host stripping
96 Bottom half of Ruleset 3 (ruleset 6 in old sendmail)
97 Hook for recursive ruleset 0 call (ruleset 7 in old sendmail)
98 Local part of ruleset 0 (ruleset 8 in old sendmail)
0 local, prog local and program mailers
1 [e]smtp, relay SMTP channel
2 uucp-* UNIX-to-UNIX Copy Program
3 netnews Network News delivery
4 fax Sam Leffler's HylaFAX software
D The local domain -- usually not needed
E reserved for X.400 Relay
H mail Hub (for mail clusters)
M Masquerade (who you claim to be)
R Relay (for unqualified names)
U my UUCP name (if you have a UUCP connection)
V UUCP Relay (class {V} hosts)
W UUCP Relay (class {W} hosts)
X UUCP Relay (class {X} hosts)
Y UUCP Relay (all other hosts)
B domains that are candidates for bestmx lookup
E addresses that should not seem to come from $M
F hosts this system forward for
G domains that should be looked up in genericstable
L addresses that should not be forwarded to $R
M domains that should be mapped to $M
O operators that indicate network operations (cannot be in local names)
P top level pseudo-domains: BITNET, DECNET, FAX, UUCP, etc.
R domains this system is willing to relay (pass anti-spam filters)
U locally connected UUCP hosts
V UUCP hosts connected to relay $V
W UUCP hosts connected to relay $W
X UUCP hosts connected to relay $X
Y locally connected smart UUCP hosts
Z locally connected domain-ized UUCP hosts
. the class containing only a dot
[ the class containing only a left bracket
1 Local host detection and resolution
2 Local Ruleset 3 additions
3 Local Ruleset 0 additions
4 UUCP Ruleset 0 additions
5 locally interpreted names (overrides $R)
6 local configuration (at top of file)
9 special local rulesets (1 and 2)
$Revision: 8.727 $, Last updated $Date: 2009/05/07 23:46:17 $